Martin Pring, Pring On Price Patterns The Definitive Guide To Pattern A

User Manual:

Open the PDF directly: View PDF .
Page Count: 185

Download
Open PDF In Browser	View PDF

Martin Pring
on Price
Patterns
The Definitive Guide to
Price Pattern Analysis
and Interpretation

Martin J. Pring

McGraw-Hill
New York Chicago San Francisco
Lisbon London Madrid Mexico City Milan
New Delhi San Juan Seoul Singapore
Sydney Toronto

The McGraw-Hill Companies

Copyright © 2005 by Martin J. Pring. All rights reserved. Printed in the United
States of America. Except as permitted under the United States Copyright Act
of 1976, no part of this publication may be reproduced or distributed in any form
or by any means, or stored in a data base or retrieval system, without the prior
written permission of the publisher.
4 5 6 7 8 9 0

DOC/DOC

0 9 8

P/N 144194-8
PART OF
ISBN 0-07-144038-0
McGraw-Hill books are available at special quantity discounts to use as pre­
miums and sales promotions, or for use in corporate training programs. For
more information, please write to the Director of Special Sales, Professional
Publishing, McGraw-Hill, Two Penn Plaza, New York, NY 10121-2298. Or con­
tact your local bookstore.
This publication is designed to provide accurate and authoritative information
in regard to the subject matter covered. It is sold with the understanding t h a t
neither the author nor the publisher is engaged in rendering legal, accounting,
or other professional service. If legal advice or other expert assistance is
required, the services of a competent professional person should be sought.
—From a Declaration of principles jointly adopted by a Committee of the American Bar
Association and a Committee of Publishers.

Contents
Introduction
Acknowledgments

Part I.

Basic Building Blocks

Pring, Martin J.
Martin Pring on price patterns : the definitive guide to price p a t t e r n
analysis and interpretation / by Martin J. Pring.
p.
cm.
ISBN 0-07-144038-0 (hardcover : alk. paper)
1. Stock price forecasting. 2. Investment analysis. I. Title: on price
patterns. II. Title.
HG4637.P75 2004
332.63'222—dc22

2004004831

1

1. Market Psychology and Prices:
Why Patterns Work
2. Three Introductory Concepts

Library of C o n g r e s s Cataloging-in-Publication D a t a

у
vii

3
9

3. Support and Resistance Zones:
How to Identify Them
4. Trendlines

22
35

5. Volume Principles as They Apply
to Price Patterns

54

Part II.

Traditional Patterns

69

6. Using Rectangles, a Case Study
for All Patterns

71

7. Head and Shoulders

96

8. Double Tops, Double Bottoms,
and Triple Patterns

127

Contents

iv

9 . Triangles

149

10. Broadening Formations

169

1 1 . Miscellaneous Patterns

188

P a r t III.

Short-Term P a t t e r n s

209

12. Smaller Patterns and Gaps

211

1 3 . Outside Bars

241

14. Inside Bars

259

1 5 . Key Reversal, Exhaustion, and Pinocchio Bars

269

16. Two- and Three-Bar Reversals

289

P a r t IV.

315

Miscellaneous Issues

17. How to Assess Whether a Breakout
Will Be Valid or False

317

1 8 . How Do Price Patterns Test?

335

Appendix Individual Patterns Summarized
Index

347
354

Introduction
In 2002, McGraw-Hill published eight of my books a n d b o o k / C D - R O M tutorial combinations. As an author, I can tell you that that was a lot of work,
and o n e of my 2003 New Year's resolutions was that e n o u g h was e n o u g h a n d
I would n o t write a n o t h e r b o o k for many years. So m u c h for resolutions,
because 2003 has seen the birth of this book a n d the DVD presentation that
is enclosed in the back, a n d 2004 will see their publication.
I first got the idea of writing this b o o k after b u m p i n g into Rick Escher of
Recognia. Recognia is an Ottawa-based software c o m p a n y that is dedicated
to offering scanning techniques for investors a n d traders. Its principal vehicle for this was originally chart pattern recognition, although this has b e e n
and will be e x p a n d e d to include o t h e r technical a n d possibly fundamental
indicators. Accurate scanning software for chart pattern recognition has
been o n e of the dreams of technicians for years, a n d the opportunity to work
with Recognia on this project a n d the ability to offer it on o u r Web site at
pring.com got me excited e n o u g h to come up with this book.
For those interested, the DVD enclosed at the back of the b o o k offers a
o n e - h o u r presentation taken from my Live in London video series. T h e contents have b e e n selected to reinforce many of the topics covered in the
book.
Several classic books on technical analysis have covered the subject of
price patterns in d e p t h . In the 1930s, R. W. Shabacker wrote several books
on the stock market, of which Technical Analysis and Stock Market Profits is the
most relevant. H. M. Gartley included a large section on this subject in Profits
and the Stock Market. Perhaps the most notable has b e e n Edwards a n d
Magee, Technical Analysis of Stock Trends, originally published in 1951 a n d
now, u n d e r the new editorship of Charles Basatti, e x p a n d e d to include other
technical a n d portfolio m a n a g e m e n t subjects. T h e r e is therefore a raft of
information available on this subject, so why offer more? T h e answer prob-

vi

Introduction

ably lies in the statement, "There is m o r e than o n e way to skin a cat." In the
old days, when charts were plotted by h a n d , time horizons were m u c h
longer. Today, with the advent of intraday trading, m o r e emphasis is being
placed on the short term. While a substantial n u m b e r of the examples featured h e r e rely on daily a n d weekly charts, quite a few intraday situations
have also b e e n included.
T h e m o r e I study m a r k e t action, the m o r e I am impressed by the fact that
prices are d e t e r m i n e d by the attitudes of market participants toward the
e m e r g i n g fundamentals. Consequently, I have tried to e x p a n d on the discussions in o t h e r books c o n c e r n i n g the psychological rationale for many of
the patterns. If it's possible to u n d e r s t a n d the logic b e h i n d these patterns,
t h e r e is a greater probability that they will be m o r e accurately—and, h o p e fully, m o r e profitably—interpreted.
A whole section of the book has b e e n devoted to what I call one- a n d twobar price patterns. These formations typically indicate exhaustion a n d are
often followed by sharp a n d timely reversals in trend. They are especially
suited to the swing a n d day trader, who is forced by time constraints to act
quickly. Earlier books covered some of these patterns, b u t o n e of the objectives of this book is to e x p a n d on this coverage with some ideas of my own.
In addition, I have tried to include a few patterns that are n o t described
in the classic texts, along with a few personal variations. Also, there are some
patterns that are described in o t h e r books, b u t that you will n o t find h e r e .
T h e r e are two reasons for this. First, it may be that they do n o t a p p e a r in
the charts very often. If I have to h u n t t h r o u g h h u n d r e d s of years of daily
data a n d am hard-pressed to find an example of a specific pattern, that pattern is hardly of practical day-to-day use. Second, some patterns, such as
o r t h o d o x b r o a d e n i n g tops a n d bottoms, trigger signals so far away from the
reversal point that m u c h of the new trend's potential has already b e e n
achieved. Discussion of such formations has b e e n kept to a m i n i m u m or
eliminated altogether. So, too, have explanations of patterns where the
d e m a r c a t i o n b o u n d a r i e s c a n n o t easily a n d conveniently be drawn.
Diamonds a n d r o u n d i n g formations c o m e to mind.
No indicator used in technical analysis is perfect, including price patterns.
In this respect, C h a p t e r 18 summarizes some of the research that Pring
Research a n d Recognia have u n d e r t a k e n t h r o u g h the identification of
5,000 patterns between 1982 a n d 2003. T h e results indicate that the two
types of formation tested, head-and-shoulders a n d double tops a n d bottoms,
generally work when the signals develop in the direction of the primary
trend. This demonstrates that correct interpretation a n d application, when
c o m b i n e d with o t h e r indicators, will p u t t h e o d d s in your favor. I say odds
because technical analysis deals only in probabilities, never in certainties.
Because of this, it is of p a r a m o u n t i m p o r t a n c e for all m a r k e t participants

Introduction

vii

to first ask the question "What is my risk?" before asking the obvious "What
is my reward?" This involves mentally rehearsing where the price would n e e d
to go in o r d e r to indicate that a pattern h a d failed. Any good driver looks
t h r o u g h the rearview m i r r o r prior to overtaking the car ahead. Traders a n d
investors should do the same by identifying risk before assessing any potential reward

Acknowledgments
T h e r e are several people w h o m I would like to thank for their help a n d
e n c o u r a g e m e n t in writing this book. T h e idea originally came to me after
I b u m p e d into my new friends at Recognia, a Canadian software company
devoted to pattern recognition software. In particular, I would like to thank
the president of Recognia, Rick Escher, who has provided me with several
ideas a n d has m a d e possible the launching of a pattern recognition subscription service at o u r Web site, pring.com. My thanks go also to Bob
Pelltier at csidata.com for kindly providing the historical data used for the
research in C h a p t e r 18.
T h e DVD at the back of the book was shot as part of a Live in London video
series. Permission to include the excerpts featured in the DVD was generously given by my friend a n d t h e sponsor of the c o n f e r e n c e , Vince
Stanzione, atwww.commodities-trader.com. United K i n g d o m - b a s e d traders
looking for some quality instruction may well want to look h i m u p .
Finally, a n d as usual, exceptional thanks goes to my wife, Lisa, who steadfastly applied herself to re-creating all the illustrations featured in the book
from my miserable original specimens despite a house move a n d personal
sadness caused by a close family bereavement.

PART I
To Lisa, who never fails to surprise me on the upside

Basic Building
Blocks

1
Market Psychology
and Prices: Why
Patterns Work
The more I work with markets, the more it becomes apparent that prices
are determined by one thing and one thing only, and that is people's changing attitudes toward the emerging fundamentals. In other words, prices are
determined by psychology. The great technician of the 1940s, Garfield Drew,
once wrote, "Stocks don't sell for what they are worth, but for what people
think they are worth." If it were not for the fact that these changing attitudes move in trends and that trends tend to perpetuate, market prices
would be nothing more than a random event, which would mean that technicians would be out of business.

Changing Attitudes and Changing
Prices
A classic example of changing attitudes that affected prices developed in
the 1970s and early 1980s. In 1973, a group of stocks known as the "Nifty
Fifty" peaked after a phenomenal rise during the 1960s. These were known
in the trade as "one-decision" stocks, because their earnings went up every
year, as did their prices. People came to the conclusion that there was only
one decision to make where these stocks were concerned: just buy! These
stocks included such growth names of the time as Kodak, Xerox,
McDonald's, and IBM. During 1973 and 1974, they declined substantially
in price, along with the rest of the market. Over the course of the next nine

3

4

Part I: Basic Building Blocks

years or so, the earnings for the g r o u p as a whole c o n t i n u e d to rise, b u t
the i n d e x did n o t make a new post-1973 high until n i n e years later.
T h u s we arrive at a situation where prices bear no reality to the earnings
trend. Perhaps prices were too high in 1973 relative to the earnings; perhaps they were not, a n d they should have c o n t i n u e d rising t h r o u g h o u t the
1970s as earnings rose. W h o knows? W h o can tell? Technicians would say,
"Who cares?" Why? Because technical analysis assumes that the changing
attitudes toward these e m e r g i n g fundamentals are reflected in price action
as displayed in charts. It's n o t dissimilar to a medical technician looking at
a patient's chart. He doesn't have to know that the patient is g r o a n i n g with
pain to diagnose a p r o b l e m . It's all t h e r e in the chart. T h e chart tells h i m
that the patient's vital signs are deteriorating to the point where d a n g e r lies
ahead and that remedial action should be taken. In a similar way, to the technician, p o o r price action signifies a weak price t r e n d a n d the probability of
trouble a h e a d in the form of a serious price decline. T h e technician does
n o t have to know the reason why; he merely observes the condition a n d
takes the necessary action.
Chart 1-1 shows the 1990s price action for Key Corp., a money-center
bank. T h e bank's earnings are shown in the lower panel. Note that there
are two periods when the price came down for a p r o l o n g e d period, the first
in the 1980s a n d the second in the late 1990s. In b o t h cases the earnings
rose, demonstrating once again that it is the attitude of market participants
toward the e m e r g i n g fundamentals rather t h a n the fundamentals themselves that is important. This is n o t the same thing as saying that earnings
are n o t important; of course they are. If we h a d known that earnings were
Chart 1-1 Key Corp. 1990-2002 vs. earnings. (Source: Telescan.)

Market Psychology and Prices: Why Patterns Work

Chart 1-2 eBay 1998-2003 vs. earnings. (Source: Telescan.)

going to rise at the b e g i n n i n g of b o t h these periods, it would have b e e n reasonable for us to assume that the price would rally as well. Only a review of
the technical position could have h e l p e d us to conclude otherwise.
Chart 1-2 shows a n o t h e r example, featuring eBay. O n c e again we can see
that the earnings increased pretty dramatically t h r o u g h o u t the period covered by the chart. However, the price fell slightly, showing the futility of buying a n d selling stocks based purely u p o n accurate earnings estimates.
History repeats, b u t never exactly, a n d as prices a p p r o a c h a turning point,
people react in roughly the same way. It is this similarity of behavior that
shows up in identifiable price patterns or formations, a n d that is the subject
of this book. Later on we will classify these various formations, establish their
reliability, a n d explain how they can be used as a basis for trading.

Technical Analysis Defined
At the outset, it is very important to understand that technical analysis is an
art form. Indeed I define it as "the art of identifying a trend reversal at a relatively early stage and riding on that trend until the weight of the evidence shows
or proves that the trend has reversed." You have probably noticed that I have
emphasized the words weight of the evidence. This is because price patterns should
be looked u p o n as one indicator in the weight-of-fhe-evidence approach. In
other words, we should not look at price patterns in isolation, but consider

Part I: Basic Building Blocks

6

m e m in conjunction with several other indicators. Over the years, technicians
have developed literally thousands of indicators, so it is obviously impossible
to follow them all. By "weight of the evidence" I m e a n four or five indicators
that the user feels comfortable with. T h e world's great religions are all primarily
concerned with finding the truth, but each has its own way of getting there.
So, too, with technical analysis; what o n e person sees as a great indicator
another may discard as useless. It's important for you as an individual to decide
which indicators to adopt in your trading by testing them over a period of time.
If you do not have confidence in your choices, I can assure you that you will
make wrong trading decisions once the trend goes against you.
By this point you may be asking, "What does he m e a n by indicators?" Well,
I m e a n oscillators such as the RSI, stochastic, KST, a n d so on. O t h e r
approaches include Elliott, G a n n , or the Wykoff m e t h o d . Still others rely
on cycles, volume, or trend-following indicators, such as moving averages
a n d trendlines. Price patterns are therefore o n e indicator in this weight-ofthe-evidence approach. I strongly believe that they should n o t be used in
isolation, b u t rather should be used in conjunction with several of these
o t h e r indicators with which you feel comfortable. Price patterns should n o t
be used blindly; they should be i n t e r p r e t e d a n d applied with a full understanding of the underlying psychology that gives rise to their development.
If you u n d e r s t a n d roughly how a n d why they work, you will be in a better
position to interpret t h e m in difficult situations.

Price Patterns and Psychology
I have used the word trend several times, but what is a trend? In my view, a
trend is a period in which a price moves in an irregular but persistent direction. T h e r e
will be a lot said on the subject of trends in the next chapter, b u t for now all
we n e e d to know is that there are various classes d e p e n d i n g u p o n the time
frame u n d e r consideration. For example, a 60-minute bar chart will reflect
very short trends, and a monthly bar chart will reflect trends of m u c h greater
duration, lasting for years. However, the principles of interpretation are identical. T h e only difference is that reversals of trends on intraday charts have
nowhere near the significance of those on the monthly charts. It should be
assumed that the longer the time span, the m o r e reliable the signal. It is
important to understand that this last statement is a generalization, since some
short-term signals can be very reliable a n d some long-term signals less reliable. T h e reason why longer-term trends have a habit of being slightly m o r e
reliable is that they are less subject to r a n d o m noise a n d manipulation.
W h e n a trend is underway, it means that either buyers or sellers are in control. During an u p t r e n d , it is the buyers, and during a downtrend, the sellers.

Market Psychology and Prices: Why Patterns Work

7

I have often h e a r d people respond to the question, "Why is so a n d so going
up?" with the flippant answer, "Because there are m o r e buyers than sellers!"
Well, strictly speaking, this is not true, because every transaction must be
equally balanced. If I sell 1,000 shares, there must be o n e or m o r e buyers
who are willing to purchase that 1,000 shares. T h e r e can never be m o r e , and
there can never be less. What moves prices is the enthusiasm of buyers relative to that of sellers. If buyers are m o r e motivated, they will bid prices higher.
On the other h a n d , if sellers are m o r e motivated, then the savvy buyers will
wait for the sellers to come down to their bids, a n d prices will decline.
Technicians have n o t e d over the years that prices do n o t usually reverse
on a dime. T h e r e is usually a transitional period between those times when
buyers have the u p p e r h a n d a n d those when sellers are pressing prices lower.
During these transitional phases, prices experience trading ranges. This
ranging action often takes the form of clearly identifiable price patterns or
formations. If these transitional periods are classified as a horizontal trend,
it follows that there are three possible trends: u p , down, a n d sideways.
Occasionally prices will resolve these horizontal price movements in favor
of the previous prevailing trend. In this case, the temporary battle between
buyers a n d sellers turns out, in retrospect, to be a period of consolidation.
Such formations would then be t e r m e d consolidation or continuation patterns,
since the prevailing trend would continue after their completion. By the
same token, if a pattern separates an u p t r e n d from a d o w n t r e n d or a downtrend from an u p t r e n d , the formation would be called a reversal pattern.
It is a generally known fact that rising prices attract bullish sentiment a n d
vice versa. W h e n prices begin their ascent, most p e o p l e do not anticipate a
large sell-off. This is because the news background remains very positive a n d
people generally extrapolate the recent past. It is only after prices have b e e n
falling for some time that bad news becomes believable. This means that
when we spot a bearish-looking pattern after a previously bullish trend, it is
unlikely that we will believe its bearish o m e n . In fact, we could say that the
less believable the pattern, the greater the odds that it is going to work.
Let's look at it a n o t h e r way. Say the gold m a r k e t has b e e n rallying for
months and there are widespread media reports telling us that gold and gold
shares have o u t p e r f o r m e d the stock market. In this kind of environment,
analysts a n d o t h e r market participants typically expect m o r e of the same.
It's possible that there is also a scary geopolitical background; for example,
oil supplies may be t h r e a t e n e d . However, the gold price forms a reversal
price pattern. At the time it would be inconceivable that this pattern could
"work," b u t that is precisely the time when it is most likely to do so. T h e tipoff
might come if the news becomes exceptionally bullish as a result of some
destabilizing geopolitical event, but the price does not make a new high. That
will give the bearish technical case substantial credibility, for if u n e x p e c t e d

8

Part I: Basic Building Blocks

"good" news (for gold) c a n n o t send the price higher, what will? Nothing,
because this news is already factored into the price. Such action tells us that the
underlying technical position is n o t as strong as it appears on the surface.
It's the market's way of saying, forget the media hype, bullish sentiment, a n d
what you hope will h a p p e n . Instead, focus on what the market is actually
telling you a n d act on that. T h e p r o b l e m is that when everyone a r o u n d you
is convinced that a specific trend is going to extend, it is very difficult to
take a different stance. Only after taking a series of losses because you
believed the crowd r a t h e r than the m a r k e t action are you likely to learn the
lesson that the m a r k e t speaks the truth a n d crowds speak with forked
tongues.

2
Three Introductory
Concepts
Introduction
Before we p r o c e e d to a discussion of price patterns themselves, it is important for us to lay the groundwork by describing a few introductory concepts.
By doing so, it is possible to obtain a firmer foundation a n d a better understanding of how markets work, a n d we will t h e n be in a better position to
interpret a n d apply price patterns for m o r e profitable trading and investing. This chapter will describe the i m p o r t a n c e a n d implications of time
frames a n d trends. It will conclude with a discussion of peak-and-trough
analysis a n d the pros a n d cons of logarithmic versus arithmetic scaling.
Incidentally, I will be using the word security extensively. This term is a
generic o n e a n d avoids the constant use of stocks, commodities, currencies,
bonds, etc. Just think of a security as any freely traded entity a n d you will
be on the right track.

Time Frames
We have already established the link between psychology a n d prices. It is also
a fact that h u m a n nature (psychology) is m o r e or less constant. This means
that the principles of technical analysis can be applied to any time frame,
from one-minute bars to weekly a n d monthly charts. T h e interpretation is
identical. T h e only difference is that the battle between buyers a n d sellers is
m u c h larger on the monthly charts than on the intraday ones. This means

9

Part I: Basic Building Blocks

10

Three Introductory Concepts

11

that any trend-reversal signals are far m o r e significant on the longer charts.
As we proceed, it will be evident that this b o o k contains a h u g e variety of
examples featuring many different time frames. For the purpose of interpretation, the timeframe really doesn't matter; it's the character of the pattern that does.
For example, if you are a long-term trader a n d you see a particular example
featured on a 10-minute bar chart, the example is just as relevant as it would
be to an intraday trader. T h e long-term trader would never initiate a trade
based on a 10-minute chart, b u t that trader can a n d should take action when
that same pattern appears on a weekly or monthly one, a n d vice versa.

Trends
A trend is a period in which a price moves in an irregular but persistent direction
a n d is a time m e a s u r e m e n t of the direction in price levels. T h e r e are many
different classifications of trends in technical analysis. It is useful to examine the m o r e c o m m o n ones, since an u n d e r s t a n d i n g of t h e m will give us
perspective on the significance of specific price patterns. T h e t h r e e most
widely followed trends are primary, intermediate, a n d short-term trends.
Whenever we talk of any specific category of t r e n d as lasting for such a n d
such a time period, please r e m e m b e r that the description is offered as a
r o u g h guide, encompassing most, b u t n o t all, trends of that particular type.
Some trends will last longer, a n d others for less time.

Primary
T h e primary trend revolves a r o u n d the business cycle, which extends for
approximately 3.6 years from trough to trough. Rising a n d falling primary
trends (bull and bear markets) last for 1 to 2 years. Since building takes longer
than tearing down, bull markets generally last longer than bear markets.
T h e primary t r e n d is illustrated in Fig. 2-1 by the thickest line. In an
ideal situation, the m a g n i t u d e a n d d u r a t i o n of the primary u p t r e n d (bull
market) are identical to those of the primary d o w n t r e n d (bear m a r k e t ) ,
b u t in reality, they are usually very different. Price patterns that offer reversal signals for primary t r e n d s usually take l o n g e r t h a n t h r e e m o n t h s to
complete.

Intermediate
It is unusual for prices to move in straight lines, so primary up- a n d downtrends are almost always interrupted by countercyclical corrections along the

Figure 2-1 The market cycle model. (Source: pring.com adapted from Yelton Fiscal.)

way. These trends are called intermediate price movements, a n d they last anywhere from six weeks to as long as nine m o n t h s . Occasionally they last even
longer, a n d some writers classify some trends that take as little as three weeks
to complete as intermediate price movements. T h e intermediate t r e n d is
represented in Fig. 2-1 by the t h i n n e r solid line. Price patterns that signal
reversals in intermediate trends do not take as long to form as those reversing primary price movements. As a r o u g h guide, I would say three to six
weeks. A lot will d e p e n d on the m a g n i t u d e a n d duration of the intermediate trend leading into the formation.

Short-Term
As a r o u g h guide, short-term trends (the d a s h e d line in Fig. 2-1) typically
last three or four weeks, although they are sometimes shorter a n d often
longer. They interrupt the course of the intermediate trend, just as the intermediate trend interrupts primary price movements. Short-term trends are
usually influenced by r a n d o m news events a n d are far m o r e difficult to identify than their intermediate or primary counterparts. Price patterns in this
case would take o n e to two weeks to develop.
T h e formation time varies a great deal, so the estimates provided h e r e
should be used as a p p r o x i m a t e guides. Quite often almost the whole of
the t r e n d is taken up by the formation of the p a t t e r n . We will show some
examples later on when some of these formations have actually b e e n
defined.

12

Part I: Basic Building Blocks

Three Introductory Concepts

13

The Interaction of Trends
It is apparent by now that the price level of any security is influenced simultaneously by several different trends. Indeed, there are many more types of
trend, some longer and some shorter than the three we have just been
describing. These too have an influence on price. Whenever we are considering a specific price pattern, our first objective is to understand which
type of trend is being reversed. For example, if a reversal in a short-term
trend has just taken place, a much smaller price movement may be expected
than if the pattern was reversing a primary trend.
Short-term traders are principally concerned with smaller movements in
price, but they also need to know the direction of the intermediate and primary
trends. This is because these longer-term trends dominate near-term price
action. This means that any surprises will develop in the direction of the
primary trend. In a bull market, the surprises will be on the upside, and in
a bear market, they will be on the downside. Just think of it this way: Rising
short-term trends that develop in a bull market are likely to be much greater
in magnitude than short-term downtrends, and vice versa. Trading losses usually happen when the trader is positioned in a countercyclical position against the
main trend. The implication for price patterns is that if a false signal is to be
given, it will almost always develop in a manner countercyclical to the trend
above it. For example, a security may be in a primary bear market. If it then
traces out a bullish intermediate price formation, chances are that this
breakout will turn out to be false because it is countercyclical in nature. We
cannot say that all bullish patterns that develop in bearish trends will fail.
What we can say, though, is that if a failure is going to take place, it is most
likely to happen following a countercyclical breakout.

Figure 2-2

The intraday market cycle model.

The Secular Trend
The primary trend consists of several intermediate cycles, but the secular
or very-long-term, trend is constructed from a number of primary trends'
This "super cycle," or long wave, extends over a substantially greater
period, usually lasting well over 10 years and often as long as 25 years A
diagram of the interrelationship between a secular and a primary trend is
shown in Fig. 2-3. It is certainly very helpful to understand the direction

Intraday Trends
What is true for longer-term trends is also true for intraday data. In this
case, the short-term trend in the daily charts becomes the long-term trend
in the intraday charts. Figure 2-2 represents three rough time approximations for the short-term, intermediate, and long-term trends in intraday
charts. Patterns on these charts have two principal differences from those
appearing on the longer-term ones. First, their effect is of much shorter
duration. Second, extremely short-term price trends are much more influenced by instant reactions to news events than are longer-term ones.
Decisions, therefore, have a tendency to develop as emotional, knee-jerk
reactions. Also, intraday price action is more susceptible to manipulation.
As a consequence, price data used in very-short-term charts are much more
erratic and generally less reliable than those that appear in the longer-term
charts.

Figure 2-3

The secular versus the cyclical trend.

14

Part I: Basic Building Blocks

Three Introductory Concepts

15

of t h e secular t r e n d . J u s t as t h e primary t r e n d influences t h e m a g n i t u d e of
the intermediate-term rally relative to t h e countercyclical reaction, so t h e
secular t r e n d influences t h e m a g n i t u d e a n d duration of a primary-trend rally
or reaction. For example, in a rising secular t r e n d , primary bull markets will
be of greater m a g n i t u d e t h a n primary b e a r markets. In a secular downtrend,
b e a r markets will be m o r e powerful, a n d will take longer to unfold, t h a n
bull markets. Price p a t t e r n s that reverse secular trends are obviously m u c h
larger t h a n those t h a t separate a primary bull a n d b e a r market, often form­
ing over many years. By t h e same token, they are also m u c h rarer.

Peak-and-Trough Progression
Widespread use of c o m p u t e r s has led to t h e d e v e l o p m e n t of very sophisti­
cated trend-identification techniques in m a r k e t analysis. Some of these work
reasonably well, b u t most do not. In t h e rush to develop these m o r e com­
plicated a p p r o a c h e s , t h e simplest a n d most basic t e c h n i q u e s of technical
analysis are often overlooked. O n e of these is the peak-and-trough approach.
It is o n e piece of evidence in t h e weight-of-the-evidence a p p r o a c h described
earlier, b u t it is also the building block for several price patterns.
T h e concept is very simple. A rising trend typically consists of a series of ral­
lies a n d reactions. Each high is higher than its predecessor, as is each low. When
the series of rising peaks a n d troughs is interrupted, a trend reversal is signaled.
In Fig. 2-4, t h e price has b e e n advancing in a series of waves, with each
peak a n d each t r o u g h b e i n g h i g h e r t h a n its predecessor. T h e n , for t h e first
time, a rally fails to move to a new high, a n d t h e s u b s e q u e n t reaction pushes

Figure 2-5

Peak-and-trough reversal and test.

it below the previous trough. This occurs at p o i n t A a n d gives a signal that, as
far as the peak-and-trough indicator is c o n c e r n e d , t h e t r e n d has reversed.
Point В in Fig. 2-4 shows a similar situation, b u t this time t h e t r e n d reversal
is from a d o w n t r e n d to an u p t r e n d .
T h e significance of a peak-and-trough reversal is d e t e r m i n e d by t h e dura­
tion a n d m a g n i t u d e of t h e rallies a n d reactions in question.
For example, if it takes two to t h r e e weeks to c o m p l e t e each wave in a
series of rallies a n d reactions, the t r e n d reversal will be an i n t e r m e d i a t e o n e ,
since i n t e r m e d i a t e price m o v e m e n t s consist of a series of short-term (twoto three-week) fluctuations. Similarly, t h e i n t e r r u p t i o n of a series of falling
i n t e r m e d i a t e peaks a n d troughs by a rising o n e signals a reversal from a pri­
mary b e a r to a primary bull market.
In Fig. 2-4 t h e price falls back to t h e level of t h e initial recovery high, b u t
in Fig. 2-5 it d r o p s below it. This is still a bullish situation because t h e ris­
ing peaks a n d troughs r e m a i n intact.

A Peak-and-Trough D i l e m m a
Occasionally, peak-and-trough progression becomes m o r e complicated t h a n
the examples shown in Figs. 2-4 a n d 2-5. In Fig. 2-6, t h e m a r k e t has b e e n
advancing in a series of rising peaks a n d troughs, b u t following the highest
peak, t h e price declines at p o i n t X t o a level that is below the previous low.
At this j u n c t u r e , t h e series of rising troughs has b e e n b r o k e n , b u t not t h e
series of rising peaks. In o t h e r words, at point X, only half a signal has been
generated. T h e c o m p l e t e signal of a reversal of b o t h rising peaks a n d troughs

Part I: Basic Building Blocks

16

Figure 2-6

Peak-and-trough half signal at a top.

arises at point Y, when the price slips below the level previously r e a c h e d at
point X, having failed in its attempt to register a new high.
At point X, there is quite a dilemma because the trend should still be classified as positive, a n d yet the very fact that the series of rising troughs has
b e e n i n t e r r u p t e d indicates underlying technical weakness. On the o n e
h a n d , we are p r e s e n t e d with half a bearish signal, while on the other, waiting for point Y could m e a n giving up a substantial a m o u n t of the profits
e a r n e d d u r i n g the bull market. T h e p r o b l e m is that if we do n o t wait, the
price could well e x t e n d its u p t r e n d , as in Fig. 2-7.

Three Introductory Concepts

Figure 2-8

17

Peak-and-trough half signal at a bottom.

That's why the d i l e m m a is probably best dealt with by referring to the
weight of the evidence from o t h e r technical indicators, such as moving
averages (MAs), volume, m o m e n t u m , a n d so on. If these o t h e r techniques
overwhelmingly indicate a t r e n d reversal, it is probably safe to anticipate a
c h a n g e in trend, even t h o u g h peak-and-trough progression has not completely confirmed the reversal. It is still a wise policy, t h o u g h , to view this signal with some d e g r e e of skepticism until the reversal is confirmed by an
interruption in both the series of rising peaks a n d the series of rising
troughs.
Figure 2-8 shows this type of situation for a reversal from a bear to a bull
trend. T h e same principles of interpretation apply at point Xas in Fig. 2-6.

What Constitutes a Legitimate Peak and Trough?
Most of the time, the various rallies and reactions are self-evident, so it is easy
to determine that these turning points are legitimate peaks a n d troughs. It is
generally assumed that a reaction to the prevailing trend should retrace
between one-third a n d two-thirds of the previous trend. This means that if we
take the first rally from the trough low to the subsequent peak in Fig. 2-9 as
100 percent, the ensuing correction should be anywhere from one-third to
two-thirds of that move. In this case it appears to be just over one-half, or a
50 percent retracement of the previous move. Occasionally, the retracement
can reach 100 percent. Technical analysis is far from precise, but if a retracem e n t move is a good deal less than the m i n i m u m one-third, then the peak
or trough in question is held to be suspect.
Figure 2-7

Peak-and-trough continuation.

Part I: Basic Building Blocks

18

Three Introductory Concepts

19

same principles are applied to trading ranges a n d rallies t h a t develop
within d o w n t r e n d s .
It is also i m p o r t a n t to categorize what kind of t r e n d is b e i n g m o n i t o r e d .
Obviously a reversal derived from a series of rallies a n d reactions each last­
ing, say, two to t h r e e weeks would be an i n t e r m e d i a t e reversal. This is
because the swings would be short-term in n a t u r e . On the o t h e r h a n d , peakand-trough reversals that develop in intraday charts are likely to have sig­
nificance over a m u c h shorter period. How short would d e p e n d on w h e t h e r
the swings were a reflection of hourly or, say, five-minute bars.

Arithmetic or Logarithmic
Scaling?
Figure 2-9 Peak-and-trough determination (magnitude).

T h e r e is o n e exception to this rule. Occasionally a correction takes the
form of a sideways t r a d i n g r a n g e r a t h e r t h a n an advance or decline. Since
the trading range is still an i n t e r r u p t i o n of the d o m i n a n t t r e n d a n d there­
fore r e f l e c t a psychological correction, we can still use the o n e , h i r d to twothirds rule, b u t in this instance substituting time for m a g n i t u d e . This idea
Is shown in Fig. 2-10, where the correction should last between one-third
a n d two-thirds of the time taken to c o m p l e t e the rally leading up to it. T h e

Figure 2-10 Peak-and-trough determination (time).

T h e r e are two axes on any m a r k e t chart. T h e x axis, along t h e b o t t o m , reg­
isters the date (except in point-and-figure charting), a n d the у axis, the price.
T h e r e are two m e t h o d s of plotting the у axis, arithmetic a n d logarithmic.
And which o n e is chosen can have very i m p o r t a n t implications.
Arithmetic charts allocate a specific point or dollar a m o u n t to a given ver­
tical distance. Thus, in C h a r t 2-1, each arrow has the same vertical distance
a n d reflects approximately 200 points. T h a t will be true at any price level.

Chart 2-1

S&P Composite, 1900-2003, arithmetic scale.

20

Part I: Basic Building Blocks

Chart 2-2 S&P Composite, 1900-2003, logarithmic scale.

Three Introductory Concepts

21

move proportionately as well, so it makes perfect rational sense to use logarithmic scaling. Having said that, when price fluctuations are relatively small,
say over a three-month period, there is very little difference between the two
scaling m e t h o d s . As a purist, though, I still prefer the log scale at all times.
T h e r e is an even m o r e i m p o r t a n t advantage of the logarithmic scale,
which we shall learn when the concept of pattern price objectives is discussed later.

Summary
• T h e principles of technical analysis can be applied equally to any time
frame.
• T h e longer the time frame the greater the significance of any technical
signal.
• A n u m b e r of different trends simultaneously influence the price level of
any market; primary, intermediate, a n d short term are the most important.
A logarithmic scale, on the o t h e r h a n d , allocates a given percentage price
move to a specific vertical distance. In Chart 2-2, each arrow represents a
move of approximately 100 percent, w h e t h e r it is at lower prices or h i g h e r
prices. T h e r e is very little noticeable difference between the scaling methods w h e n charts are plotted over short periods of time, where price fluctuations are relatively subdued. However, with large price fluctuations, t h e r e
are considerable differences.
T h e arithmetic scale suppresses price fluctuations at low levels a n d exaggerates t h e m at high points. T h u s the 85 p e r c e n t 1929-1932 decline hardly
shows up at all in Chart 2-1, b u t the 40 p e r c e n t late 1990s-early 2000s retreat
(no small decline) is greatly exaggerated. Chart 2-2 shows that the logarithmic scaling brings back 1929 a n d does n o t exaggerate the turn-of-thecentury b e a r market. T h e m e d i a love to hype stories a n d news because that
is what sells. You will find that charts featuring financial markets or economic
n u m b e r s are almost always plotted on an arithmetic scale because this has
the effect of exaggerating the most r e c e n t changes. A n o t h e r hyping technique used by the m e d i a is to p r e s e n t the data for a short period using a
very limited scale. T h e reader is t h e n left with the sense of a dramatic move.
This would n o t be the case if the data were displayed over a m u c h l o n g e r
period using a wider price scale.
As you can appreciate, I am very m u c h in favor of using a logarithmic scale
because it displays price trends in a proportionate way. Psychology tends to

• Peak-and-trough progression is the most basic trend-identification technique a n d is a central building block of price pattern analysis.
• In o r d e r to qualify as a new legitimate peak or trough a good rule of
t h u m b is that the price should retrace between one-third a n d two-thirds
of the previous move.
• Arithmetic scaling suppresses price fluctuations at lower price levels a n d
exaggerates t h e m at h i g h e r levels.

Support and Resistance Zones: How to Identify Them

3

Support and
Resistance Zones:
How to Identify
Them
S u p p o r t a n d resistance are two m o r e building blocks of the technical arse­
nal used in price p a t t e r n analysis. A discussion of these two concepts will
therefore h e l p us greatly in o u r u n d e r s t a n d i n g of how price p a t t e r n s work.
A lot of p e o p l e use t h e t e r m support w h e n they really m e a n resistance a n d
use resistance w h e n they really m e a n s u p p o r t . It's no w o n d e r t h a t t h e r e is a
lot of confusion. Basically, these are points on a c h a r t where t h e probabili­
ties favor at least a t e m p o r a r y halt in t h e prevailing t r e n d .

Figure 3-1

Previous low is a good place to expect support.

they will continually increase their bids until their purchasing d e m a n d s have
been satisfied. On the o t h e r h a n d , if sellers are the m o r e anxious, t h e n they
will be willing to liquidate at lower prices a n d the general price level will fall.
If in doubt, think of s u p p o r t as a t e m p o r a r y floor for prices a n d resistance
as a ceiling.
At t h e b e g i n n i n g of Fig. 3-1, t h e price is declining. It finds a b o t t o m at A
a n d t h e n moves u p . T h e n e x t time it falls to A, it again rallies, so A may now
be said to be a s u p p o r t area. This establishes o u r first principle of sup­
p o r t / r e s i s t a n c e analysis: A previous high or low is a potential resistance/support
level. T h e third time t h e price slips to A, it goes t h r o u g h or, as we say, vio­
lates support. O n e of t h e first principles of identifying a potential s u p p o r t
level, t h e n , is to look for previous lows. In t h e case of potential resistance,
this would be in t h e area of a previous high.
Figure 3-2 shows a m o r e e x t e n d e d e x a m p l e . This time t h e price f o u n d
t e m p o r a r y s u p p o r t at В. С also proves to be a s u p p o r t p o i n t , b u t n o t e t h a t

Support and Resistance
In their classic b o o k Technical Analysis of Stock Trends, Edwards a n d Magee
defined support as "buying (actual or potential) sufficient in volume to halt
a d o w n t r e n d in prices for an appreciable period," a n d resistance as "selling
(actual or potential) sufficient in volume to satisfy all bids a n d h e n c e stop
prices from going h i g h e r for a time."
A support zone represents a concentration of demand, a n d a resistance zone
represents a concentration of supply. T h e word concentration is emphasized
because supply a n d d e m a n d are always in balance, b u t it is the relative enthu­
siasm of buyers as c o m p a r e d to sellers, or vice versa, that is i m p o r t a n t because
that is what d e t e r m i n e s trends. If buyers are m o r e enthusiastic t h a n sellers,

23

Figure 3-2

Support reverses its role to resistance on the way back up.

Part I: Basic Building Blocks

24

Support and Resistance Zones: How to Identify Them

25

Consequently, w h e n t h e price r e t u r n s to t h e old high, those who b o u g h t at
that level have great motivation to sell in o r d e r to break even, so they begin
to liquidate. Also, those who b o u g h t at lower prices have a t e n d e n c y to take
profits at t h e old high, since t h a t is t h e t o p of familiar g r o u n d . By t h e same
token, any prices above the old high look expensive to potential buyers; con­
sequently, t h e r e is less enthusiasm on their part, so they begin to pull away
from t h e market.

Figure 3-3

Resistance reverses its role to support on the way down.

t h e rallies are reversed at s u p p o r t level B. T h u s t h e s e c o n d principle is t h a t
support reverses its role to resistance on the way up. J u s t t h i n k of it this way: A
floor in a b u i l d i n g acts as a s u p p o r t z o n e , b u t w h e n you fall t h r o u g h it, t h e
floor now b e c o m e s resistance, called a ceiling. T h e reason why s u p p o r t a n d
resistance reverse their roles can be a p p r e c i a t e d with an e x p l a n a t i o n of
some e l e m e n t a r y psychology. No o n e likes to take a loss, a n d while some
p e o p l e o v e r c o m e this feeling by cutting their losses at an early stage, oth­
ers h o l d on until t h e price c o m e s back to where t h e security in question
was originally b o u g h t . At t h a t p o i n t they are able to b r e a k even a n d sell,
thereby creating a quantity of supply sufficient to temporarily halt t h e
advance.
Finally, in Fig. 3-3 we see the price rally t h r o u g h resistance at В a n d A (the
f o r m e r s u p p o r t level). T h e ensuing decline t h e n finds s u p p o r t at A again.
T h u s o u r third principle is t h a t resistance reverses its role to support on the way
down.

Rules for Determining Potential
Support/Resistance Points
1. Previous Highs and Lows
We have already established that previous highs a n d lows are potential sup­
p o r t or resistance levels. Highs are i m p o r t a n t because many m a r k e t partic­
ipants may have b o u g h t close to or at t h e actual high for a move. W h e n
prices decline, t h e n o r m a l h u m a n response is n o t to take a loss b u t to h o l d
on. T h a t way, it is felt, t h e r e will n o t be t h e pain of actually realizing a loss.

W h e n a price rallies a n d t h e n falls back to the previous low, these bar­
gain b a s e m e n t prices appeal to potential buyers. After all, they missed t h e
o p p o r t u n i t y t h e first time prices r e t r e a t e d to this level, a n d they are there­
fore thankful to have a n o t h e r c h a n c e . For t h e same reason, sellers are reluc­
tant to p a r t with their securities as prices a p p r o a c h the previous low, since
they saw t h e m b o u n c e before a n d naturally w o n d e r why t h e same process
should n o t b e r e p e a t e d .
C h a r t 3-1 shows t h e sugar price for a p e r i o d s p a n n i n g 2002-2003. N o t e
how previous highs a n d lows offer g o o d s u p p o r t / r e s i s t a n c e points for
future trading. Unfortunately, t h e r e is no way of knowing w h e t h e r a par­
ticular level will t u r n o u t to be s u p p o r t or resistance, or even w h e t h e r it will
be a pivotal p o i n t at all. T h a t ' s why these are merely intelligent places for
anticipating a t e m p o r a r y reversal. Resort to o t h e r indicators such as oscil­
lators is therefore r e q u i r e d .

Chart 3-1 Sugar, 2002-2003, daily.

26

Part I: Basic Building Blocks

Support and Resistance Zones: How to Identify Them

Chart 3-2

27

Hewlett-Packard, 2001-2002, daily.

2. At Round Numbers
Support and resistance zones have a habit of forming at round numbers.
This is probably because numbers such as 10, 50, or 100 represent easy psychological points upon which traders and investors often base their decisions. In the 1970s, for example, the Dow Jones Industrials had a great deal
of difficulty surpassing the 1000 level. For gold in the 1980s and mid-1990s,
the magic number was $400, and so forth. The guide for potential turning
points, then, is to look for round numbers.

3. Trendlines and Moving Averages Represent Dynamic
Levels of Support and Resistance
A good trendline should reflect the underlying trend. One of the rules for
assessing the significance of a line relates to how many times it has been
touched or approached. The more the merrier in this case. If a price falls
back to a specific low on several occasions, this makes that particular price
level a strong support zone. The same is true of trendlines and moving averages (MAs). Every time a price moves back to an up trendline or a rising
MA and bounces, it is reinforced as a dynamic level of support. The same
would be true in reverse for a declining trendline or moving average. It
therefore makes sense to buy as the price falls to an up trendline (or rising
MA) and to sell when it rises to a down trendline (or rising MA). A low-risk
stop may then be placed just beyond the line or MA in case the
support/ resistance zone is violated.
Chart 3-2, for Hewlett-Packard, shows a very good example of how a down
trendline acted as resistance. Note also that the interaction of a reliable MA,
such as the 200-day MA featured in this chart, acts as reinforcement of the
resistance zone. This works in the same way as if we were building a house
and doubled the thickness of the roof. The identical principle holds when
a moving average and a trendline are at the same level; they double the
strength of the resistance (or support in the case of an up trendline and
MA intersection).
4. Emotional Points on a Chart Represent Potential
Support/Resistance Levels
This concept will be covered in subsequent chapters when we consider gaps,
extreme points of Pinocchio bars, two-bar reversals, key reversals, and so
forth. For the moment, suffice it to say that most emotional points are those
at which prices following a strong and persistent trend experience a strong
extension of that trend. During the course of the bar's formation, they then
abruptly reverse their direction.

Gaps represent another example of emotional points. They are formed
when buyers or sellers respond so emotionally to news that a blank space,
or gap, is left on the chart. In Chart 3-3, probably because of unexpected
bad news, the sugar price experiences three downside gaps. Later on, when
emotions become more stable, the price rallies and tries to "close" each of
Chart 3-3

March, 2003, sugar.

28

C h a r t 3-4

Part I: Basic Building Blocks

Boeing, 1998-2003, weekly.

Support and Resistance Zones: How to Identify Them

29

first leg of the 1929-1932 bear market ended in October 1929 at 195, just
over half the September high. The halfway mark in an advance sometimes
represents the point of balance, often giving a clue to the ultimate extent
of the move in question or, alternatively, indicating an important juncture
point for the return move. Thus, between 1970 and 1973, the market
advanced from 628 to 1067. The halfway point in that rise was 848, or
approximately the same level at which the first stage of the 1973-1974 bear
market ended.
By the same token, rising markets often find resistance after doubling from
a low; the first rally from 40 to 81 in the 1932-1937 bull market was a double.
In effect, the 50 percent mark falls in the middle of the one-third to twothirds retracement described in Chapter 2 in the discussion of peak-andtrough progression. These one-third and two-thirds proportions can be
widely observed in all securities and also serve as support or resistance
zones.

the gaps. In the case of the gap on the left, resistance is found at the gap's
opening. In the other two examples, resistance forms at the lower part of
the gap. Gaps are one of the most reliable technical concepts from the point
of view of projecting potential support or resistance areas.
Chart 3-4, for Boeing, shows another emotional point. This time it's the
bottom of a very wide bar in early 2002. Note that this low developed at a
round number, $50. Normally this would have been a support level the next
time the price fell to $50, but in the fall of 2002 the price went right through
it. Even so, $50 did turn out to be a pivotal point the next time Boeing ral­
lied. It goes to show that even if a support/resistance zone is violated once,
it can still turn out to be a pivotal point in subsequent price action.

5. Proportionate Moves, Retracements, and So On
The law of motion states that for every action, there is a reaction. Price
trends established in financial markets are really the measurement of crowd
psychology in motion and are also subject to this law. These swings in sen­
timent often show up in proportionate price moves.
Perhaps the best-known principle of proportion is the 50 percent rule. For
instance, many bear markets, as measured by the DJIA, have cut prices by
half. As examples, the 1901-1903, 1907, 1919-1921, and 1937-1938 bear
markets recorded declines of 46, 49, 47, and 50 percent, respectively. The

Ratio-scale charts are helpful in determining such points, since moves of
identical proportion can easily be projected up and down. Moreover, these
swings occur with sufficient consistency to offer possible reversal points at both
peaks and troughs. Remember, technical analysis deals with probabilities, which
means that forecasts should not be made using this method in isolation.
In addition, when undertaking a projection based on the rules of pro­
portion, it is always a good idea to see whether the price objective corre­
sponds to a previous support or resistance point. If it does, the odds are
much higher that this zone will represent a reversal point, or at least a tem­
porary barrier. When a security price is reaching new all-time high ground,
another possibility is to try to extend up trendlines. The point at which the
line intersects with the projection using the rules of proportion may well
represent the time and place of an important reversal. Experimentation will
show that each security has a character of its own, with some lending them­
selves more readily to this approach and others not at all.
Chart 3-5, for Dollar General, shows an example using one-third, twothirds, and 50 percent retracements. In this instance, the decline from A to
В is 100 percent of the move. If we want to establish possible resistance
points for subsequent rebounds, then the intelligent places to monitor are
these one-third, two-thirds, and 50 percent retracements. As you can see,
the rally ending at С represents a 50 percent retracement and that ending
at D a 66 percent or two-thirds retracement.
Many technicians use a sequence of numbers discovered by Leonardo
Fibonacci, a thirteenth-century Italian mathematician. The sequence has
many properties, but a key one is that each new number is the sum of the
two previous numbers in the series. Thus 5 and 8 = 13, 8 and 13 = 21, and
so on. The significance of this sequence for our purposes is that it offers

Part I: Basic Building Blocks

30
Chart 3-5

Dollar General, 1999-2000, daily.

some guidelines for proportionate moves. For example, each number in the
sequence is 61.8 percent of the next number, 38.2 percent of the number
after that, and so forth.
In this respect, Charts 3-6 and 3-7, for Palladium, show some possibilities.
In the case of Chart 3-6, the huge 1997-1998 rally represents 100 percent of
Chart 3-6 Palladium, 1997-2000, daily.

Support and Resistance Zones: How to Identify Them

31

Chart 3-7 Palladium, 1996-2000, daily.

the move. ВС a n d BD, which t u r n e d out to be s u p p o r t levels, were really 61.8
p e r c e n t a n d 50 p e r c e n t retracements, respectively. Note how t h e 61.8 per­
cent level b e c a m e support for an e x t e n d e d period in 1998. Also, the 23.6 per­
cent (one Fibonacci n u m b e r divided by two later ones in t h e series) level was
pivotal in 1998 a n d 1999, as reflected by t h e thick black line b e g i n n i n g at E.
Finally, C h a r t 3-7 shows t h e same principle applied to upside projections.
O n c e again, AB represents 100 p e r c e n t of t h e decline a n d lines are drawn
at upside Fibonacci p r o p o r t i o n s . In this case, t h e n e x t h i g h e r n u m b e r
divided by t h e c u r r e n t n u m b e r is 1.61, t h e n 2.61, a n d so on. It is self-evi­
d e n t how t h e 161.8 p e r c e n t a n d 261.8 p e r c e n t p r o p o r t i o n s b e c o m e key piv­
otal points in future price action. O n c e again, these levels are n o t
g u a r a n t e e d to b e c o m e i m p o r t a n t pivotal points b u t are intelligent places
on t h e c h a r t to anticipate that possibility.

Rules for Determining the
Probable Significance of a
Potential Support or Resistance
Zone
At this p o i n t you are probably asking, "How do I know how i m p o r t a n t each
s u p p o r t a n d resistance level is likely to be?" Unfortunately, t h e r e is no h a r d
a n d fast answer, b u t t h e r e are some general rules that can act as guidelines.

Part I: Basic Building Blocks

Support and Resistance Zones: How to Identify Them

32
1. The Amount of a Security That Changed Hands in a Specific
Area—the Greater the Activity, the More Significant the Zone
This is fairly self-evident, for whenever you have a large number of people
buying or selling at a particular price, they have a tendency to remember
their own experiences. Buyers, as we have already established, like to break
even. Sellers, on the other hand, may have bought lower down and recall
that prices previously stalled at the resistance level. Their motivation for
profit taking becomes that much greater.
2. The Greater the Speed and Extent of the Previous Move, the
More Significant a Support or Resistance Zone Is Likely to Be
The attempt to climb through the resistance level here can be compared to
the efforts of a person who tries to crash through a door. If he attacks the
door from, say, 10 or 12 feet away, he can propel himself with lots of momentum, and the door will probably give way. On the other hand, if he begins
his attempt from 100 feet away, he will arrive at the door with less velocity
and will probably fail in his attempt. In both cases the door represented the
exact same resistance, but it was the resistance relative to the velocity of the
person that was important. The same principle can be applied to the market, in that a long, steep climb in price is similar to the 100-foot run, and the
resistance level resembles the door. Consequently, the more overextended
the previous price swing, the less the resistance or support that is required
to halt it.
In this respect, Chart 3-8, for the Bank Commerciale Index (an Italian
stock average), falls to a low in August and then rallies. However, the subsequent decline is extremely steep, so by the time the index reaches the support level at around 880, it is completely exhausted and sellers are unable
to push prices below this level.
Chart 3-9 features the gold price. Note how it makes a high in January
1986 and then falls away. The subsequent rally is quite steep, then the price
runs out of steam and declines again. Finally, in late July, it works its way
higher in a fairly methodical manner. This time the same level of resistance
is easily overcome because buyers are not as exhausted as they were in their
March attempt.
3. Examine the Amount of Time Elapsed
The third rule for establishing the potency of a support or resistance zone
is to examine the amount of time that has elapsed between the formation

Chart 3-8 Bank Commerciale, daily.

33

Part I: Basic Building Blocks

34
of the original congestion a n d the n a t u r e of general market developments
during that period. A supply that is 6 m o n t h s old has greater potency than
o n e established 10 or 20 years previously. Even so, it is almost u n c a n n y how
support a n d resistance levels remain effective time a n d time again, even
w h e n separated by many years.

Summary
• S u p p o r t a n d resistance r e p r e s e n t a concentration of d e m a n d a n d supply

4

Trendlines

sufficient to halt a price move at least temporarily.
• They are n o t signals to buy or sell b u t rather are intelligent places for
anticipating a reversal. They should always be used in conjunction with
other indicators.
• Potential s u p p o r t / r e s i s t a n c e zones develop at previous highs a n d lows,
r o u n d n u m b e r s , trendlines a n d MA's, emotional points on charts, a n d
r e t r a c e m e n t points such as Fibonacci proportions.
• T h e significance of a s u p p o r t or resistance zone d e p e n d s u p o n the
a m o u n t of an asset that previously c h a n g e d h a n d s in that area, the speed
a n d extent of the previous price move, a n d the period of time that has
elapsed since the zone was last e n c o u n t e r e d .

Trendlines are perhaps the simplest of the tools in our technical arsenal a n d
are arguably o n e of the most effective. Since the construction of nearly all price
patterns requires the use of trendlines, this concept is a fundamental building block
of pattern identification and interpretation. In this chapter we will describe the
characteristics of trendlines a n d explain how the significance of individual
lines can be d e t e r m i n e d .
A trendline is a straight line connecting a series of ascending bottoms in
a rising market or the tops of a descending series of rally peaks. Those trendlines j o i n i n g the lows are called up trendlines a n d those c o n n e c t i n g the tops
are referred to as down trendlines. Typically a down trendline is constructed
by j o i n i n g the final peak with the top of the first rally, as in Fig. 4-1. W h e n
the price breaks above the trendline, a t r e n d change signal is given. T h e
opposite is true for an up trendline (see Fig. 4-4).

How Should Trendlines Be
Drawn?
In o r d e r to be a true trendline, a line must c o n n e c t two or m o r e peaks or
troughs. Otherwise it will be drawn in space a n d will have no significance.
You will often see p e o p l e constructing lines that touch only o n e point, as
in Fig. 4-2, or even no points at all, as in Fig. 4-3. Such lines have no meaning whatsoever, a n d are really worse than drawing n o t h i n g at all. This is
because by simply a p p e a r i n g on the charts, such lines give the observer the
impression that they actually have some significance. This is a fundamentally i m p o r t a n t point because a true trendline is a graphic way of representing
the underlying trend. Consequently, if a line touches only o n e point, it cannot be a true trendline.

35

Part I: Basic Building Blocks

36

Figure 4-3
Figure 4-1

Trendline not connected.

Down trendline.

Ideally, an up trendline is constructed by connecting the final low with the
first bottom following the initial rally, as is done by line AD in Fig. 4-4. This
is called the primary trendline. In the case of a primary trend, this would be
the bear market low and the first intermediate bottom. The example shown
here offers a fairly shallow angle of ascent. Unfortunately, the price rallies
sharply, which means that the violation develops well after the final peak.

In such situations it is better to redraw the line as the price moves up. In
Fig. 4-4 this new line is line ВС, which is obviously a better reflection of the
underlying trend. This is called a secondary trendline. Down trendlines are
constructed using the same principles, but in reverse.
Since trends can be sideways, it follows that trendlines can also be drawn
horizontally. This is often the case when we construct price patterns such

Figure 4-4
Figure 4-2

37

Trendlines

Trendline connected once.

Primary and secondary up trendlines.

Part I: Basic Building Blocks

Trendlines

39

38
as the neckline of a horizontal head-and-shoulders (H&S) pattern or the
u p p e r or lower b o u n d a r i e s of rectangles (described in later chapters). In
the case of price patterns, the p e n e t r a t i o n of these lines usually warns of a
change in trend, as does the violation of rising or falling trendlines.
It's i m p o r t a n t to u n d e r s t a n d at this p o i n t that drawing trendlines is m o r e
a matter of c o m m o n sense t h a n of following a set of h a r d a n d fast rules.

Bar versus Line or Close-Only
Charts
Some charts are plotted with bars a n d others are line charts. T h e question
naturally arises, "Which form should be used for the purposes of trendline
analysis?" In most cases, bar charts offer m o r e timely signals, w h e t h e r the
signal is a peak-and-trough progression, price pattern completion, or trendline violation. In technical analysis, timeliness comes at a price, and the price
in this case is m o r e whipsaws. With traditional daily or weekly charts, the
closing price is very i m p o r t a n t because it sorts out the m e n (i.e., those who
are willing to take h o m e a position overnight or over a weekend) from the
boys (i.e., those who are n o t ) . This has b e c o m e a less i m p o r t a n t factor in
some markets, as they trade for 24 h o u r s Sunday t h r o u g h Friday. (However,
since all markets are closed over the weekend, Friday closes c o n t i n u e to
maintain their importance.) Even so, closing prices are, for the most part, more
important chart points than highs or lows. Also, since t h e r e is m u c h excitement
during the day as u n e x p e c t e d news breaks, highs a n d lows often represent
r a n d o m points on the chart. For this reason, it is often a better idea to construct trendlines using closing data. I am n o t going to say that this is always
the case because some bar trendlines have greater significance t h a n closeonly ones, based on the rules for significance described later in this chapter. T h u s , it is always crucial to apply c o m m o n sense as m u c h as strict
technical rules. T h e question you should be constantly asking is, "Which line
better reflects the underlying trend?"

Figure 4-5 Reversal up trendline break.

a rising price trend. In this case, the trendline j o i n i n g the series of troughs
is eventually p e n e t r a t e d on the downside. T h e fourth peak r e p r e s e n t e d the
highest point in the bull trend, so the downward violation of the trendline
signals that a b e a r move is u n d e r way.
T h e upward price trend a n d trendline penetration in Fig. 4-6 are identical to those in Fig. 4-5, but the action following this warning signal is entirely

Trendline Breaks Can Signal
Reversals or Consolidation
T h e completion of a price pattern can signify either (1) a reversal in the
previous trend, which is known as a reversal pattern, or (2) a resumption of
the previous trend, which is called a consolidation or continuation pattern.
Similarly, the penetration of a trendline will result in either a reversal of that
trend or its continuation. Figure 4-5 illustrates this from the perspective of

Figure 4-6 Consolidation up trendline break.

Part I: Basic Building Blocks

Trendlines

41

40

Extended Trendlines

Figure 4-7 Consolidation turns into a reversal break.

different. This is because the trendline violation results in the advance continuing, b u t at a greatly r e d u c e d pace. T h e third alternative is that the price
consolidates in a sideways trading range prior to reversing. This is shown in
Fig. 4-7. T h u s , whenever a trendline is violated, the odds strongly favor a
change in trend. T h a t change can be either an actual reversal or a (sideways)
trading range following an up- or downtrend.
In most instances, t h e r e is, unfortunately, no way of telling at the time of
the violation which possibility will prove to be the o u t c o m e . O n e clue may
be provided by the trendline's angle of ascent or descent. Since sharp-angled
trendlines are less sustainable, their penetration has a tendency to be followed by a consolidation rather than a reversal.
Valuable clues can be gleaned by evaluating the state of health of the mar-

Most p e o p l e observe the violation of a trendline, assume that the trend has
changed, a n d eventually forget about the line. This is a mistake, because an
e x t e n d e d line can b e c o m e j u s t as i m p o r t a n t as the violated line itself.
If an up trendline is violated, for instance, the price often returns to the
e x t e n d e d line. This is known as a throwback move. Figure 4-8 shows a trendline reversing its previous role as support when the throwback move turns
it into an area of resistance. Figure 4-9 shows the same situation for a declining market.
Chart 4-1, for instance, shows an up trendline break for the U.S. governm e n t 20-year b o n d yield. T h e penetration of this relatively steep line was
followed by a small decline. However, the cyclical rally peak in 1984 was
t u r n e d back by the e x t e n d e d line, which reversed its role a n d proved to be
strong resistance. Chart 4-2 shows the same idea, b u t for a down trendline
for the Eurodollar yield. In this instance it was violated in 1987. Later the
e x t e n d e d line proved to be support for the 1993 decline.

Logarithmic (Ratio) versus
Arithmetic Scales
The importance of plotting charts on a logarithmic as opposed to an arithmetic
scale was discussed in Chapter 2. T h e choice of scale is even more critical for
a timely a n d accurate use of trendline analysis. This is because prices tend to

ket's overall technical structure. Also, a trendline p e n e t r a t i o n may occur at
the time of, or just before, the successful completion of a reversal price pattern.
If a series of ascending peaks and troughs is accompanied by progressively
lower volume, this is a sign that the advance is r u n n i n g out of steam (since
volume is no longer going with the t r e n d ) . In this situation, a trendline violation is likely to be of greater significance than if volume h a d continued
to e x p a n d with each successive rally. It is n o t necessary for a downside penetration to be a c c o m p a n i e d by high volume, b u t a violation that occurs as
activity expands emphasizes the bearish u n d e r t o n e because of the obvious
switch in the d e m a n d / s u p p l y balance in favor of sellers.

F'Sure 4-8 Extended up trendline becomes resistance.

42

Part 1: Basic Building Blocks

Trendlines

43

Chart 4-2 Cash Eurodollars, 1980-1994, weekly.

Figure

accelerate in the direction of the prevailing trend at the e n d of a major movement; i.e., they rise faster at the end of a rising trend and decline more sharply
at the termination of a bear market. In a bull market, prices rise slowly after
an initial burst, then advance at a steeper a n d steeper angle as they approach

the ultimate peak, looking rather like the left-hand cross section of a mountain.
Chart 4-3 shows an up trendline break for Intel based on a logarithmic scale.
Note that the downside penetration develops in mid-December. Chart 4-4
shows exactly the same period, but this time the scaling is arithmetic. T h e
Chart 4-3 Intel, 2001-2002, daily (logarithmic).

Chart 4-1 U.S. government 20-year bond yield, 1975-1986, weekly.

Part I: Basic Building Blocks

45

Trendlines

44
Chart 4-6
Chart 4-4

IBM, 2001-2002, daily (logarithmic).

Intel, 2001-2002, daily (arithmetic).

trendline break is entirely different, since it initially comes as a whipsaw in
late December, followed by a valid break in mid-January. T h e downwardpointing arrow on the left m a r k e d the logarithmic break. T h u s it is apparent that up trendlines are violated m o r e quickly on a logarithmic t h a n on
an arithmetic scale.

Conversely, down trendlines are violated sooner on an arithmetic scale.
This can be seen from a comparison of Charts 4-5 a n d 4-6, for IBM.
Generally speaking, penetration of a logarithmically drawn trendline is
more accurate in reflecting trend reversals than is penetration of an arithmetically drawn trendline, although in both these examples the final penetrations came at m o r e advantageous prices for the arithmetically scaled chart.

Chart 4-5 IBM, 2001-2002, daily (arithmetic).

Significance of Trendlines
It has b e e n established that a break in t r e n d caused by p e n e t r a t i o n of a
trendline results in either an actual trend reversal or a slowing in the pace
of the trend. Although it may n o t always be possible to d e t e r m i n e which of
these alternatives will develop, it is still i m p o r t a n t to u n d e r s t a n d the significance of a trendline penetration; the guidelines described next should help
m this evaluation. Essentially, this evaluation d e p e n d s on t h r e e factors: the
length of the line, the n u m b e r of times it has b e e n touched, a n d the angle
of ascent or descent. Let's consider each of t h e m in turn.

1 . L e n g t h o f t h e Line
A trendline measures a trend, so the longer the line, the longer the trend it
is monitoring a n d the m o r e significant the trendline. If a series of ascending

Part I: Basic Building Blocks

Trendlines

46
bottoms occurs over a three- to four-week span, the resulting trendline is only
of minor importance. If the trend extends over a period of o n e to three years,
however, its violation marks a significant j u n c t u r e point. Just remember, big
trends result in big signals, small trends in small ones.

2. Number of Times the Trendline Has Been Touched
or Approached
A trendline derives its authority from the number of times it has been
touched or approached; i.e., the larger the number, the greater the trendline's significance. This is because a trendline represents a dynamic area of
support or resistance. Each successive test of the line contributes to the
importance of this support or resistance role, and thus the authority of the
line is a true reflection of the underlying trend. Just remember that a move
close to the line (an approach) is almost as important as an actual touching because it reflects the line's importance as a support or resistance area.
Also, if a line gains significance from the fact that it has been touched or
approached on numerous occasions, the extended line will become equally
as important, but from a reverse point of view. This is because extended lines
reverse their support/resistance role. For example, if a good up trendline
has been violated, the price is now below it. Any rally will therefore find resistance at this line, which had, of course, previously been support.

Figure 4-10 Steep angles of ascent.

trendline. An example for a rising trend is shown in Fig. 4-11. This distance
is then projected in the direction of the new trend from the point of penetration.
T h e t e r m price objective is p e r h a p s misleading. Objectives are usually
reached when a trendline violation turns o u t to be a reversal, but because
they are m o r e often exceeded (as we shall learn with price patterns), the

3. Angle of Ascent or Descent
A very sharp trend, as shown by the dashed line (AB) in Fig. 4-10, is difficult to sustain and is liable to be broken rather easily, even by a short sideways movement. It is then necessary to draw a line with a smaller angle of
ascent or descent (AC).
All trends are eventually violated, but the steeper ones are likely to be ruptured more quickly. The violation of a particularly steep trend is not as significant as the violation of a more gradual one. Penetration of a steep line usually
results in a short corrective movement, following which the trend resumes, but
at a greatly reduced and more sustainable pace. Usually, the penetration of a
steep trendline represents a continuation rather than a reversal break.

Measuring Implications
Trendlines have measuring implications when they are broken. T h e meas u r e m e n t is the m a x i m u m vertical distance between the price a n d the

Figure 4-11

Down trendline violation measuring objective.

Part I: Basic Building Blocks

Trendlines

49

48

Figure 4 - 1 2

Up trendline violation measuring objective.

Figure 4 - 1 4

Down trend channel.

objective becomes m o r e of a m i n i m u m expectation. An objective for a downside reversal is illustrated in Fig. 4-12.

Trend Channels
So far only the possibilities of drawing trendlines j o i n i n g bottoms in rising
markets a n d top's in declining ones have b e e n examined. It is also useful to
draw lines that are parallel to those basic trendlines, as shown m Fig. « 3 . In

a rising market, the parallel line known as a return trendline joins the tops of
rallies, a n d d u r i n g declines the return line joins the series of bottoms (see
Fig. 4-14). T h e area between these trend extremities is known as a trend channel.
T h e r e t u r n line is useful from two points of view. First, it represents an
area of support or resistance, d e p e n d i n g on the direction of the trend.
Second, a n d p e r h a p s m o r e important, p e n e t r a t i o n of the r e t u r n trendline
represents a signal that either the trend will accelerate or a reversal in the
basic t r e n d of at least a temporary p r o p o r t i o n is a b o u t to take place.
In Fig. 4-15, the violation of the return line signifies that the price advance
has begun to accelerate. In effect, the channel in Fig. 4-15 represents a rising
trading range, and the trendline violation is a breakout from it.

Exhaustion
On the o t h e r h a n d , if the angle of the trend c h a n n e l is m u c h steeper, as in
Figs. 4-16 a n d 4-17, the violation of the r e t u r n line represents an exhaustion move. T h e failure of the price to hold above (below) the r e t u r n line
then signals an important reversal in trend. This is often the case if the break
t h r o u g h the return line is a c c o m p a n i e d by high volume.
Consider a situation in which a person is sawing a thick piece of wood.
At first the sawing strokes are slow b u t deliberate, but gradually the person
realizes that this task is going to take some time, becomes frustrated, a n d
slowly increases the speed of the strokes. Finally the p e r s o n bursts into a
Figure 4 - 1 3

Up trend channel.

Part I: Basic Building Blocks

50

Figure 4-15 Up trend channel breakout.

frantic effort a n d is forced to give up the task at least temporarily because
of complete exhaustion. T h e same principles h o l d true in a declining marke. In this case, the e x p a n d i n g volume at the low represents a selling climax As a general rule the steeper the channel, the m o r e likely it is that
the b r e a k o u t will turn o u t to be an exhaustion move.

Figure 4

Exhaustion break
16 Down trend channel exhaustion break.

Trendlines

Figure 4-17

Up trend channel exhaustion break.

Exhaustion also develops when a price rallies temporarily above a regular down trendline (or below an up t r e n d l i n e ) , then breaks back below
(above) it. In the case of a down trendline, the situation is akin to s o m e o n e
j u m p i n g up a n d temporarily p u s h i n g t h r o u g h the ceiling. He is able to pull
his h e a d t h r o u g h to the next floor for a few m o m e n t s , b u t he t h e n falls
sharply back to the floor below. At this point he has used up all his spare
energy in the attempt to move to the n e x t floor a n d is totally exhausted.
Before he can make a n o t h e r attempt he will n e e d some time to gain some
new energy. T h e same is true of the price, which makes an effort to rally
above the trendline b u t is unable to maintain the breakout.
This temporary break often indicates that the prevailing t r e n d has m u c h
further to r u n . It also raises a dilemma c o n c e r n i n g the way in which a trendline should be constructed. In Fig. 4-18, for instance, we see a false break
above trendline AB. Should AB now be a b a n d o n e d , or should the peak of
the exhaustion break be c o n n e c t e d to the rally high to form a new (dashed
AC) trendline? Again, it's a matter of c o m m o n sense. On the o n e h a n d , the
top of the whipsaw break is technically the correct place to draw the line,
but c o m m o n sense suggests that the original line is a better reflection of
the underlying trend. After all, at the time of the whipsaw, it h a d b e e n
t o u c h e d t h r e e times. If a new line is then drawn to reflect the break, that
line will have b e e n t o u c h e d only twice, o n c e at the outset a n d o n c e at the
whipsaw peak. In a sense, the whipsaw is a d d i n g further credibility to the
initial line because the price was unable to h o l d above it. If we h a d c o m e
u p o n this situation after the whipsaw break a n d tried to construct a line, it

52

Part I: Basic Building Blocks

Trendlines

53

a support/resistance area that the price is unable to hold the break. If it
were not such a significant barrier, the break would have held and the whipsaw would have been avoided. Consequently, when the price is able to experience a valid break, the signal is that much stronger.
Chart 4-7 shows an example of a whipsaw downside break for Microsoft
in 1998.

Summary
• Trendlines are an easy tool to understand, but they should be used with
a strong dose of common sense. Considerable experimentation and practice are required before the art of interpreting them can be successfully
mastered.
Figure 4-18 Down trendline construction.

would have b e e n even m o r e obvious that line AB was far superior to line
AC because it has b e e n t o u c h e d or a p p r o a c h e d on far m o r e occasions.
T h e same principles are true in reverse for an up trendline. W h e n you
think a b o u t it, a whipsaw break actually adds credibility to the trendline. This is
because the price is able to violate the line, b u t this line is so significant as

Chart 4-7

Microsoft, 1998-1999, daily.

• Trendline violations signal either a temporary interruption or a reversal
in the prevailing trend. It is necessary to refer to other pieces of technical evidence to determine which is being signaled.
• The significance of trendlines is a function of their length, the number
of times they have been touched or approached, and the steepness of the
angle of ascent or descent.
• A good trendline reflects the underlying trend and represents an important support and resistance zone.
• Extended trendlines reverse the former support/resistance role and
should not be overlooked.
• Exhaustion breaks are usually followed by sharp moves in the opposite
direction to the break. Exhaustion breaks enhance the significance of a
trendline.

Volume Principles as They Apply to Price Patterns

55

3. Occasionally, price action offers mild signs of an impending trend reversal, but volume can throw up characteristics of its own that literally shout
this message. In such cases, a study that was limited to price action would
fail to uncover a really good and obvious warning or opportunity.

5

Volume Principles
as They Apply to
Price Patterns
Almost everything that technicians use in plotting a specific security involves
either the price itself or a statistical variation on it. Volume can offer a new
dynamic in our interpretation of crowd psychology. Therefore, analyzing volume trends gives us a better understanding of how and why price patterns
work. In effect, the study of the characteristics of volume gives greater depth
to the weight-of-the-evidence approach described earlier. Volume not only
measures the enthusiasm of buyers and sellers but is a variable that is totally
independent of price. In this chapter we will discuss some general principles of volume interpretation. However, this is not the final word, because
I will have more to say on this subject as we expand the discussion to include
individual price patterns.

Principles of Volume
Interpretation
1. The first and most important principle is that volume typically goes with
the trend. It is normal for activity to expand in a rising market and to contract in a declining one (see Fig. 5-1). In this sense, volume is always
interpreted in relation to the recent past. Comparing twenty-first-century 1 billion-plus share days on the NYSE with early-twentieth-century
levels of 5 or 6 million is of little help. Such a comparison reflects institutional, not psychological, changes. Volume is higher today because of
more companies being listed, the advent of derivatives, lower commissions, and so forth. On the other hand, a 3 billion share day this week
compared to a recent 1.5 billion share day last month is relevant,
because it shows a significant change in activity over a period in which
institutional changes will be nonexistent.
We know that when prices move in trends, this does not occur in a
straight line. Instead, the price works its way up and down in a zigzag

Benefits of Volume Studies
Volume studies offer three major benefits.
1. When price and volume patterns are compared, it is important to see
whether they are in agreement. If they are, the probabilities favor an
extension of the trend.
2. If price and volume disagree, this tells us that the underlying trend is not
as strong as it looks on the surface. If a breakout from a price pattern
develops with such a disagreement, a warning of a potentially invalid signal is being given.

Figure 5-1

Volume goes with the trend.

Part I: Basic Building Blocks

56

Figure 5-2

Volume moves in trends.

fashion. Volume trends are similar. On the left side of Fig. 5-2, for instance,
the arrows indicate an expanding volume trend. It is apparent that the
level of activity does not expand in every period. T h e r e are quiet periods
and active ones, but the general thrust is up. T h e right-hand part of the
figure features a down trend in volume. It too is irregular. W h e n we talk
about volume rising or falling, we are usually referring to its trend. It is
normal for such trends to be interrupted by aberrations in volume levels.
Volume trends, like price trends, can be intraday, short, intermediate, or
long, d e p e n d i n g on the nature of the chart.
2. T h e a m o u n t of m o n e y flowing into a security must always equal the
a m o u n t of m o n e y flowing out. This is true regardless of the level of
volume.
Consequently, it is the level of enthusiasm of buyers or sellers that determines the course of prices.
If buyers are bullish, they will raise their bids until their d e m a n d s are satisfied. If sellers react to b a d news, they may panic, p u s h i n g prices down
sharply, b u t at all times the a m o u n t of a security being sold is equal to that
being purchased.
3. T h e combination of rising volume a n d rising price is normal. It indicates that things are in gear. Such a state of affairs has no forecasting
value, except to imply that it is likely that a negative divergence between
price a n d volume lies ahead.
4. Volume normally leads price during a bull move. A new high in price that
is n o t confirmed by volume should be regarded as a red flag, warning

Volume Principles as They Apply to Price Patterns

Figure 5-3

57

Volume leads price in uptrends.

that the prevailing trend may be a b o u t to reverse. In Fig. 5-3, the price
peaks at point C, yet the average volume reached its m a x i m u m a r o u n d
point A. Such action is normal; the declining volume peaks warn of
underlying technical weakness. Unfortunately, there are no h a r d a n d
fast rules about how many divergences precede a peak. Generally speaking, t h o u g h , the g r e a t e r the n u m b e r of negative divergences, the
weaker the underlying technical picture. Also, the lower the peaks relative to each other, the less enthusiasm is being generated, and the more
vulnerable the technical position becomes, once buying dries up or selling enthusiasm intensifies. A new high that is a c c o m p a n i e d by virtually
no volume is j u s t as bearish as a new price high with virtually no upside
momentum.
An example is shown in Chart 5-1, for the Mexico Fund, where you
can see that the volume clusters gradually b e c o m e smaller as the price
rallies. Eventually this negative technical characteristic is confirmed as
the price violates the M a r c h / M a y up trendline.
Rising prices a c c o m p a n i e d by a trend of falling volume (Fig. 5-4) is an
a b n o r m a l situation. It indicates a weak a n d suspect rally a n d is a bear
market characteristic. W h e n it is recognized, it can a n d should be used
as a piece of evidence pointing to a primary bear m a r k e t environment.
Volume measures the relative enthusiasm of buyers a n d sellers. W h e n
volume shrinks as prices rise, the advance occurs because of a lack of
selling r a t h e r than because of sponsorship from buyers. Sooner or later
the trend will reach a point where sellers b e c o m e m o r e motivated. After

Part I: Basic Building Blocks

58

Chart 5-1

Volume Principles as They Apply to Price Patterns

59

Mexico Fund. 1995-1996, daily.

Figure 5-5

Falling price and rising volume are bearish.

be that it picks up for two or three bars just after the peak. In fact, this
would be a more typical situation.
that, prices will start to pick up on the downside. One clue is provided
when volume increases noticeably as the price starts to decline. This is
shown in Fig. 5-5, where you can see that volume starts to pick up as the
price starts a sell-off. In such situations, it is not necessary for volume
to expand throughout the decline, as it does in this example. It could

Figure 5-4

Rising price and falling volume are bearish.

Figure 5-6 shows how the volume configurations change between a
bull market and a bear market. Chart 5-2, for Coors, shows the final rally
being accompanied by a trend of declining volume. When the lower
trendline is violated, volume picks up noticeably. In this instance, we

Figure 5-6

Volume characteristics change in bull and bear markets.

Part I: Basic Building Blocks

60

Volume Principles as They Apply to Price Patterns

61

Chart 5-2 Coors, 2000-2001, daily.

Figure 5-7 Volume and a parabolic blow-off.

have o n e bearish volume configuration that is instantly followed by
another. Chart 5-3, for Radio Shack, shows several b e a r m a r k e t rallies
in which the rising price t r e n d is a c c o m p a n i e d by declining volume.
6. Sometimes both price a n d volume expand slowly, gradually working into
an exponential rise with a final blow-offstage. Following this development,
Chart 5-3 Radio Shack, 2000-2001, daily.

both volume and price fall off equally sharply. This represents an exhaustion move and is characteristic of a trend reversal, especially when it is
supported by a one- or two-bar price pattern (discussed in subsequent
chapters). T h e significance of the reversal will d e p e n d u p o n the extent
of the previous advance a n d the degree of volume expansion. Obviously
an exhaustion move that takes four to six days to develop will be nowhere
near as significant as o n e that develops over a matter of weeks. This phen o m e n o n is t e r m e d a parabolic blow-off a n d is featured in Fig. 5-7.
Unfortunately, exhaustion, or blow-off, moves such as this are not easy to
define in the sense that it is possible to construct clearly definable trendlines or price patterns. For this reason, it is usually n o t possible to spot
the terminal phase until a period or so after volume a n d price have
reached their crescendos.
N e w m o n t Mining, in Chart 5-4, offers a classic example of an exponential increase in b o t h price a n d volume that ends in tears in the form
of an a b r u p t reversal in late S e p t e m b e r 1987.
A selling climax is the opposite of a parabolic blow-off. It occurs when
prices fall for a considerable time at an accelerating pace, accompanied
by e x p a n d i n g volume. Prices typically rise after a selling climax. T h e low
that is established at the time of the climax is unlikely to be violated for
a considerable time. A price rise from a selling climax is by definition
accompanied by declining volume. This is the only time when contracting volume a n d a rising price may be r e g a r d e d as normal. Even so,
it is i m p o r t a n t to make sure that volume expands on subsequent rallies,
as indicated in Fig. 5-8. T h e termination of a bear t r e n d is often, but

Part I: Basic Building Blocks

62
Chart 5-4 Newmont Mining, 1986-1987, daily.

not always, accompanied by a selling climax. Having said that, it is
important to note that in many instances the selling climax, after a rally,
is followed by new lows.
In Chart 5-5, for Dresser Industries, we see a selling climax develop in
the late summer. This is then followed by a rally, but in this instance the

Volume Principles as They Apply to Price Patterns

63

Chart 5-5 Dresser Industries, 1994-1995, daily.

climax is not the final bottom. At the peak of the rally, rising prices are
accompanied by declining volume, a typical bear market characteristic.
8. W h e n prices advance following a long decline a n d t h e n react to a level
at, slightly above, or marginally below the previous trough, this is a bullish sign if the volume on the second trough is significantly lower than
the volume on the first. T h e r e is an old saying on Wall Street, "Never
short a dull market." This saying applies very m u c h to this type of situation, in which a previous low is being tested with very low volume. Such
a situation indicates a complete lack of selling pressure (see Fig. 5-9).
9. A downside breakout from a price pattern, trendline, or moving average
(MA) that occurs on heavy volume is abnormal a n d is a bearish sign that
confirms the reversal in trend (Fig. 5-10). W h e n prices decline, it is usually because of a lack of bids, so volume contracts. This is normal activity
and does not give us m u c h information. However, when volume expands
on the downside, it is because sellers are m o r e motivated, so the decline,
other things being equal, is likely to be more severe.
10. W h e n the price has b e e n rising for many m o n t h s , an a n e m i c rally
(Fig. 5-11) a c c o m p a n i e d by high volume indicates churning action a n d
is a bearish factor.

Figure 5-8 A selling climax.

11. Following a decline, heavy volume with little price c h a n g e is indicative
of accumulation a n d is normally a bullish factor (Fig. 5-12).

Part I: Basic Building Blocks

64

Volume Principles as They Apply to Price Patterns

65

Figure 5-9 Look for low volume when testing lows.
Figure 5-11 Churning is bearish.
12. Record volume coming off a major low is usually a very reliable signal
that a significant b o t t o m has b e e n seen. This is because it indicates that
an underlying change in psychology has taken place. Such reversals in
sentiment are usually of a primary t r e n d m a g n i t u d e . Examples in the
U.S. stock market developed in March 1978, August 1982 a n d 1984, and
October 1998. A similar pattern also developed at the 1987 low in bonds
a n d eurodollars. This is n o t an infallible indicator, t h o u g h , because
record volume was achieved in J a n u a r y 2001 for b o t h the NYSE and
Nasdaq, yet this did n o t turn out to be the final low for the move.

Figure 5-10 Rising volume on a downside breakout is bearish.

13. W h e n volume a n d price e x p a n d at a sharp pace, b u t short of a parabolic blow-off, a n d t h e n contract slightly, this usually indicates a c h a n g e
in trend. Sometimes this is an actual reversal a n d at o t h e r times a consolidation. This p h e n o m e n o n is featured in Fig. 5-13 a n d represents a
temporary exhaustion of buying power. It is associated with several onea n d two-bar price patterns discussed in later chapters.
14. W h e n the price experiences a small r o u n d i n g top a n d volume a r o u n d ing bottom, this is a doubly a b n o r m a l situation, since price is rising a n d

Figure 5-12

Accumulation is bullish

Part I: Basic Building Blocks

Volume Principles as They Apply to Price Patterns

67

66
Chart 5-6 Microsoft, daily.

Figure 5-13 High volume after a rally indicates exhaustion.

volume falling as the peak is reached. After the peak, volume e x p a n d s
as the price declines, which is also a b n o r m a l a n d bearish. An example
is shown in Fig. 5-14. An example featuring Microsoft is featured in
Chart 5-6. Note how the letter n characterizes the price action, whereas
the volume configuration is closer to a r o u n d e d letter V.
N o n e of the indicators in the technical arsenal is guaranteed to work
every time. This is certainly true of volume characteristics. However, when
volume is used in combination with price characteristics in p a t t e r n

interpretation, it greatly enhances the probability that a specific formation will work. As we discuss specific formations, the basic volume principles described h e r e will be e x p a n d e d to suit individual cases.

Summary
• Volume is a totally i n d e p e n d e n t variable from price.
• It is n o r m a l for volume to go with the trend. W h e n these characteristics
are present, they have little forecasting value.
• W h e n volume trends are moving in a direction opposite to that of price,
this is a b n o r m a l a n d either warns of an i m p e n d i n g t r e n d reversal or
emphasizes the significance of any breakout.
• Volume trends experience exhaustion p h e n o m e n a . These are called parabolic blow-offs at tops a n d selling climaxes at lows.

Figure 5 - 1 4

Watch volume on rallies and reactions.

PART II
Traditional
Patterns

6

Using Rectangles,
a Case Study for
All Patterns
This chapter will explain the basic principles of price pattern interpretation
that apply to most formations by using one formation as an example. It will
then be possible to expand the discussion to include other patterns in subsequent chapters. Our chosen vehicle is the rectangle, so let's take a closer
look at it.

Basic Concepts
Price trends do not usually reverse on a dime. Instead, the up- and downor down- and uptrends are typically separated by a transitional period or
trading range, where buyers and sellers are equally matched.
The two possibilities are shown in Figs. 6-1 and 6-2. Figure 6-1 represents
a typical cycle consisting of three trends: up, sideways, and down. Then two
more trends develop as the price experiences another horizontal trading
range, followed by a renewed uptrend. Figure 6-2 shows a highly emotional
market that changes without warning. This is by far the exception, since
most trends are separated by some form of trading range. An oil tanker takes
a long time to slow down and then go into reverse. The same is normally
true of financial markets. Generally speaking, the longer the trend, the more
time spent in the reversal (turnaround) process.
This transitional or horizontal phase has great significance because it is
the demarcation between a rising and a falling trend (or vice versa). If
prices have been advancing, the enthusiasm of the buyers has outweighed
71

Part.II: Traditional Patterns

72

Using Rectangles, a Case Study for All Patterns

73

Figure 6-1 Top and bottom reversals.

the pessimism of the sellers up to this point, a n d buyers have optimistically
bid prices higher. During the transition phase, the battle between optimism
a n d pessimism b e c o m e s m o r e or less even, until finally, for o n e reason or
another, it is tipped in a new direction as the relative weight of selling
pushes t h e t r e n d (of prices) down. It is the b r e a k i n g to post-transitional
new low g r o u n d that alerts the trader to the fact that a reversal in t r e n d
has taken place. In o t h e r words, w h e n prices fall below the trading range,
a sell signal is given. At t h e t e r m i n a t i o n of a bear trend, the reverse process
occurs.
Over the years, technicians have n o t e d that such ranging action at b o t h
tops a n d bottoms has taken the form of clearly definable price patterns or
price formations.

The Transitional Turning Point
Figure 6-3 shows the price action at the e n d of a long rising trend. As soon
as the price rises above line B, it is in the transitional area, although this is
a p p a r e n t only after prices have b e g u n to fluctuate sideways.

Figure 6-2 Reversal on a dime.

Figure 6-3 Trading range reversal.

O n c e the price is in this trading zone, it rises to line A, which is a resistance area. T h e resistance is n o t initially identifiable, b u t after the price has
backed off from the initial high, it is evident that the price level at A represents a previous high. This makes it a potential candidate for a resistance
point. W h e n the next rally also fails at A, it is possible to construct a horizontal trendline. T h e same thing can be d o n e at B when the subsequent
reaction finds s u p p o r t at that level for a second time. T h e d e m a n d / s u p p l y
relationship comes into balance in favor of the sellers whenever the price
reaches A. This temporary reversal may occur because buyers refuse to pay
up for a security, because the higher price attracts m o r e sellers, or for a combination of b o t h reasons.
Just as the price level at A reverses the balance in favor of the sellers, so
does the s u p p o r t level B alter it once again. This time, the trend moves in
an upward direction, for at B, prices b e c o m e relatively attractive to buyers
who missed the boat on the way u p . Also, sellers who feel that the price will
again reach A hold off. For a while, there is a standoff between the two parties within the confines of the area b o u n d e d by lines A a n d B. Finally, the
price falls below B, a n d a new (downward) trend is signaled.
T h e contest between buyers and sellers is like a battle fought by two armies
engaged in trench warfare. In Fig. 6-4, example A, armies A a n d B are facing off. Line AA represents army A's t r e n c h line of defense, a n d BB is army
B's. T h e arrows indicate the forays between the two lines as both armies fight
their way to the opposing trench, but are unable to p e n e t r a t e the line of
defense. In example B, army B finally pushes t h r o u g h army A's line of

Part II: Traditional Patterns

Using Rectangles, a Case Study for All Patterns

75

74

Figure 6-5 Downside breakout signal.
Figure 6-4 Trench warfare.
defense. Army A is then forced to retreat and make a stand at the second
trench (line A2). In the stock market, line A x represents resistance, and once
it is overcome, this signifies a change in the balance between buyers and
sellers in favor of the buyers, so that prices will advance quickly until new
resistance is encountered. The second line of defense, line A2, then represents resistance to a further advance.
On the other hand, army B might break through A 2 quite easily, but the
farther it advances without time to consolidate its gains, the more likely it
is to become overextended, and the greater is the probability of its suffering a serious setback from a counterattack when it reaches the next line of
defense. At some point, therefore, it makes more sense for this successful
force to wait and consolidate its gains.
If prices in financial markets extend too far without having time to digest
their gains, they too are more likely to face a sharp and seemingly unexpected reversal. This is another application of the second rule for identifying the significance of a support or resistance level described in Chapter 3.

Introducing the Reversal
Rectangle and the Psychological
Conditions Contributing to Its
Formation
T h e trading range separating rising a n d falling price trends discussed h e r e
is a p a t t e r n known as a rectangle. T h e rectangle in Fig. 6-5, which marks the
t u r n i n g p o i n t between the bull a n d b e a r phases, is t e r m e d a reversal pattern.

T h e actual reversal signal is given when the price breaks decisively t h r o u g h
the lower horizontal trendline. Reversal patterns at market tops are known
as distribution because the security is said to be "distributed" from strong,
informed participants to weak, u n i n f o r m e d ones. T h e underlying psychology is that during the formation of the rectangle, the news is good, and uninformed m a r k e t participants, h e a r i n g it for the first time, are e n c o u r a g e d to
buy. Forecasts are also quite rosy at this time, so little risk is perceived from
the long side. Every transaction has to have two parties, a buyer a n d a seller.
For their part, the sellers, or distributors, have a different idea. They h a d
been carefully buying on the way up in anticipation of the good news. Now
that the news is materializing, it is time to sell, a n d who better to sell to than
those u n i n f o r m e d participants who are hearing the g o o d news for the first
time? This kind of activity has given rise to the well-known expression "buy
on the rumor, sell on the fact."
Price patterns, including rectangles, that develop at m a r k e t bottoms are
called accumulation formations (where the security passes from weak, uninformed participants to strong, informed ones; see Fig. 6-6). In this situation,
the underlying psychology is reversed: T h e sellers, having seen the price of
their security move lower, decide to sell when the bad news a n d forecasts of
m o r e to c o m e start to filter t h r o u g h the investment community. By the same
token, potential buyers h a d previously held off because they wanted to get
the bad news out of the way. T h e emergence of discouraging reports is therefore o n e of the signals that these so-called strong buyers h a d b e e n waiting
for. A n o t h e r might be a sign that things are going to improve down the road.
This process is called accumulation because informed buyers are said to be
accumulating the security in anticipation of better times ahead. These are

Part II: Traditional Patterns

76

Figure 6-6 Upside breakout signal.

said to be strong buyers because they are n o t easily p u t off by discouraging
news. T h e i r focus is on the future, a n d only w h e n they see signs of p o o r e r
t h a n expected developments down the r o a d are they likely to liquidate.
W h e n you think of it in very simple terms, the top reversal rectangle is
n o t h i n g m o r e than a sophisticated variation of a signal that a series of rising peaks a n d troughs has b e e n reversed. A b o t t o m formation is simply the
opposite.
We can also appreciate that these patterns incorporate all the basic building blocks that have already b e e n covered: peak-and-trough analysis, support a n d resistance, a n d trendline analysis. It is also important to u n d e r s t a n d
that a reversal pattern must have s o m e t h i n g to reverse. In o t h e r words, all
reversal patterns must be p r e c e d e d by a prior t r e n d in the opposite direction from the reversal signal. Tops must be p r e c e d e d by u p t r e n d s a n d bottoms by downtrends.

Consolidation Rectangles
If the rectangle following an u p t r e n d is completed with a victory for the buyers as the price pushes t h r o u g h the u p p e r line AA (see Fig. 6-7), a reversal
does n o t develop because the b r e a k o u t above AA reaffirms the underlying
trend. In this case, the corrective phase associated with the formation of the
rectangle temporarily interrupts the bull m a r k e t a n d becomes a consolidation pattern. Such formations are also referred to as continuation patterns.

Using Rectangles, a Case Study for All Patterns

Figure 6-7

77

Upside continuation breakout signal.

Some price patterns show up only in a continuation m o d e ; these are discussed in C h a p t e r 12.
During the period in which the pattern is being formed, there is no way
of knowing in advance which way the price will ultimately break. It should
always be assumed, then, that the prevailing trend is in existence until it is proved
to have been reversed. If we are faithfully applying the weight-of-the-evidence
approach, it may be possible to anticipate a possible reversal if the other indicators are pointing in that direction. However, the pattern itself can never be
categorized as a reversal type until the reversal is actually signaled.
A c o n t i n u a t i o n or consolidation rectangle in a d o w n t r e n d is shown in
Fig. 6-8. An example is featured in Chart 6-1, for the c o p p e r price.

What Constitutes a Rectangle?
In the example in Fig. 6-5, the u p p e r a n d lower b o u n d a r i e s of the rectangle were each t o u c h e d on at least three occasions by the two horizontal
trendlines. However, it was also possible to construct the rectangle using the
first two points. Would that have constituted a valid pattern? T h e answer is
yes. For the rationale for this, we n e e d to examine the psychology underlying the formation of the pattern. Remember, the transition period is nothing less t h a n a battle between buyers a n d sellers. If the battle is e x t e n d e d
by the two boundaries b e i n g t o u c h e d or a p p r o a c h e d on m o r e than two
occasions, this m e a n s that t h e conflict between the two parties is m o r e

78

Part II: Traditional Patterns

Using Rectangles, a Case Study for All Patterns

79

T h e longer the battle between the two trenches, the m o r e exhausted the
combatants will be. Thus the victory, w h e n it comes, will be that m u c h m o r e
decisive.
You may have noticed that I have used the word approached as well as
touched. This is important, because in real life you will often find that the
construction of the u p p e r a n d lower b o u n d a r i e s of a rectangle is n o t as precise as that in Fig. 6-5. I regard a close a p p r o a c h to any trendline, including the o u t e r boundaries of a rectangle, as being almost as valid as an actual
contact. After all, if the price comes close to a b o u n d a r y a n d t h e n reverses,
it definitely reinforces this level as a support or resistance zone.

Figure 6-8 Downside continuation breakout signal.

significant. It also implies that the bearish implications of the rectangle will
be stronger. We can even go so far as to say that the more times the outer boundaries of a rectangle have been touched or approached, the greater the significance of
the pattern. It goes back to the second rule described in C h a p t e r 4 concerning the significance of a trendline. T h e same holds true for the horizontal
trendlines that form a rectangle. Let's r e t u r n to o u r military analogy again.
Chart 6-1 Copper, daily.

The Significance of Any Given
Price Pattern
T h e principles of price pattern construction a n d interpretation can be
applied to any time frame, from one-minute bars all the way to monthly or
even a n n u a l charts. However, the significance of a price formation for a specific time frame is a direct function of the formation's size a n d d e p t h .
We have already established that a rectangle whose boundaries have b e e n
t o u c h e d many times is m o r e significant than o n e whose boundaries have
b e e n t o u c h e d only twice. We can extend this idea by saying that the longer
a pattern takes to complete, the greater the number of fluctuations within it; and the
deeper its trading range, the more substantial the following move is likely to be.
Let's consider these t h r e e factors in turn.

1. Time Frames
T h e longer the time frame, the m o r e significant the pattern. A p a t t e r n that
shows up on a m o n t h l y chart is likely to be far m o r e significant than o n e
on an intraday chart, a n d so forth. In addition, the longer a p a t t e r n takes
to develop in a particular time frame, the greater its significance within that
time frame. Let's say we are looking at a daily chart a n d we spot two formations. T h e first takes ten days to complete a n d the second four weeks.
Clearly the four-week battle between buyers a n d sellers is m u c h greater a n d
m o r e involved than the ten-day encounter. Consequently, when the outcome is resolved, t h e e n s u i n g price move is likely to be m u c h greater. I use
the word likely because this is a generality. Most of the time the larger pattern will be m o r e i m p o r t a n t , b u t n o t every time. In technical analysis, we are
dealing in probabilities, never certainties. This m e a n s that small patterns will
occasionally be followed by large moves, but normally it is the larger ones
that are.

80

Part II: Traditional Patterns

Let's c o m p a r e an accumulation p a t t e r n at a price low with the construction of a building. It is just as i m p o r t a n t to build a strong base from which
prices can rise as it is to build a large, strong, d e e p foundation u p o n which
to construct a skyscraper. In the case of security prices, the foundation is an
accumulation pattern that represents an area of indecisive combat between
buyers a n d sellers. D u r i n g an accumulation phase, m o r e sophisticated
investors a n d professionals are positioning or accumulating the security in
anticipation of improved conditions six to n i n e m o n t h s ahead. As m e n tioned earlier, ownership is b e i n g transferred from weak, u n i n f o r m e d
traders or investors to strong a n d knowledgeable h a n d s . Consequently, the
longer the p a t t e r n takes to complete a n d the greater the level of activity
within it, the m o r e significant the accumulation process a n d therefore the
stronger the technical position.
T h e reverse is true at m a r k e t tops, where a substantial a m o u n t of distribution inevitably results in a protracted period of price erosion.

Using Rectangles, a Case Study for All Patterns

81

Chart 6-2 CRB Composite, 1957-1982, daily.

2. The Significance of Pattern Fluctuations
T h e greater the n u m b e r of fluctuations within a pattern, the greater the significance of that pattern. W h e n the price action has b e e n at a stalemate for
a long time a n d investors a n d traders have b e c o m e used to buying at o n e
price a n d selling at the other, a move b e y o n d either limit represents a fund a m e n t a l change a n d has great psychological significance.
In 1972 a n d 1973 the commodity markets experienced a h u g e run-up.
Chart 6-2 shows the CRB Composite. You can see that the rally was p r e c e d e d
by a multiyear trading range. T h e r e was no way of knowing that the breakout from the range would be followed by such a spectacular advance. Even
so, the sheer size of the battle between buyers a n d sellers over many years
would have indicated that when a resolution finally did take place, it would
most likely signal an above-average price trend.

3. The Significance of Pattern Depth
T h e d e e p e r a pattern, the greater its significance. T h e d e p t h of a formation also determines its significance. Consider the t r e n c h war analogy once
m o r e . If the opposing trenches are very close together, say within 100 yards,
this m e a n s that the victorious assault, when it comes, will be less significant
than if they are separated by several miles. In that case, the battles will have
b e e n m u c h m o r e intense a n d the victory that m u c h greater. T h e same is
true in the financial markets. T h e b r e a c h i n g of a wide trading range generally has far greater psychological significance t h a n t h e b r e a c h i n g of a

narrow o n e . Psychology, as expressed in market prices, tends to move in proportion. T h e greater the (proportionate) swing within the pattern, the
greater the subsequent move is likely to be. If you get wide price swings during the formation of the pattern, you are also likely to get wide swings after
it has b e e n completed.
Having said that, it is also i m p o r t a n t to n o t e that whenever you see a very
tight a n d constrained trading range, this indicates that the battle between
buyers a n d sellers is very evenly balanced. This is especially true when the
level of activity shrinks to almost n o t h i n g . W h e n that balance is tipped o n e
way or the other, you will often find that prices move quite quickly a n d to
a greater d e g r e e than is suggested by the m e a s u r i n g implication. We see
an example in Fig. 6-24, w h e r e a pretty sharp decline follows the relatively
narrow rectangle. Perhaps the guiding light in this is to see how many times
the u p p e r a n d lower b o u n d a r i e s have b e e n t o u c h e d o r a p p r o a c h e d . T h e
greater the n u m b e r of times, the greater the significance of the line as a
s u p p o r t or resistance zone, a n d therefore the m o r e decisive the victory.

Measuring Implications
Technical analysis is best at identifying t r e n d changes at an early stage a n d
n o t so useful in forecasting how far a t r e n d will e x t e n d . Price p a t t e r n

82

Part II: Traditional Patterns

Figure 6-9 Rectangle top measuring objective.

interpretation is o n e exception, since the construction of these formations
offers some limited forecasting possibilities.
Nearly all price patterns obtain measuring objectives from their depth.
T h e rectangle is no exception. Figure 6-9 shows a rectangle that has formed
a n d completed a top (distribution). T h e measuring implication of this formation is the vertical distance between its outer boundaries, i.e., the distance
between lines AA a n d BB projected downward from line BB.
In many cases, the price t r e n d will e x t e n d b e y o n d the objective. In really
strong moves, it will achieve multiples of it. We can take the process a step
further by stating that the various multiples of t h e objective can b e c o m e
i m p o r t a n t s u p p o r t a n d resistance areas in their own right. T i m e a n d again,
these price objective areas t u r n o u t to be i m p o r t a n t s u p p o r t or resistance
points. Unfortunately, t h e r e is no way to d e t e r m i n e where t h e actual j u n c ture p o i n t will be for any rally or reaction. This emphasizes t h e principle
that in technical analysis, t h e r e is no known way of consistently determining the d u r a t i o n of a price m o v e m e n t . It is possible only to speculate on
the probability that a specific area will prove to be a s u p p o r t or resistance
zone.
Consequently, while this measuring formula offers a r o u g h guide, it is usually a minimum expectation. An example for a multiple-objective downside
break is shown in Fig. 6-10. H e r e we see the price sink by t h r e e times the
original objective. These multiples of the objective can be j u s t as important
in forecasting a p r o b a b l e pivotal p o i n t on the way back u p , as we can see in
the very right-hand part of the chart.

Using Rectangles, a Case Study for All Patterns

83

Figure 6-10 Rectangle top multiple measuring objective.

Arithmetic versus Logarithmic
Scaling
T h e choice of arithmetic or logarithmic scaling is really i m p o r t a n t when
measuring implications are being considered. This is because the m e t h o d
used d e t e r m i n e s the actual level of the objective. To recap, in arithmetically scaled charts, any given vertical space reflects the same dollar or p o i n t
a m o u n t . T h u s , we'll say that o n e inch represents $2, b o t h at the lower e n d
of the chart, in the $2 to $5 range, a n d at h i g h e r levels, such as the $100
to $110 range. All units of m e a s u r e are plotted using the same vertical distance. Prices plotted on a ratio or logarithmic scale show identical distances
for identical p e r c e n t a g e moves. T h u s , o n e inch could r e p r e s e n t a 20 percent move anywhere on the chart.
Fortunately, almost all c o m p u t e r software gives the user the o p t i o n of
choosing between arithmetic a n d logarithmic scales.
T h e i m p o r t a n c e of using logarithmic scales whenever possible is shown
in Fig. 6-1 la a n d b. In a, the price has traced out a n d b r o k e n down from a
rectangle. Projecting the vertical distance between 200 a n d 100 downward
gives an objective of 0, clearly a very unlikely possibility. On the o t h e r h a n d ,
Fig. 6-116 gives the same projection based on a logarithmic scale. In this case,
a m o r e realistic objective of 50 is obtained.
If a rectangle appears as a b o t t o m reversal pattern, the measuring rules
remain consistent with those given for the distribution formation. T h e only
difference is that we project the objective a n d multiples of the objective in
an upward, n o t a downward, direction. T h e exact same principles also apply

84

Part II: Traditional Patterns

Using Rectangles, a Case Study for All Patterns

Figure 6-12

Figure 6-11 Measuring implication, (a) Arithmetic, (b) Ratio.
to continuation rectangles. Figure 6-12 shows an upside breakout from a rectangle that forms d u r i n g a bullish trend. N o t e that in this case, t h e price
does n o t reach its upside objective immediately, b u t does so after a small
rally a n d reaction. This is why the objective is described by the term ultimate Most people buy the b r e a k o u t on the assumption that they will make
m o r e or less instant profits as the price moves straight to the objective, but
that is n o t necessarily the case.

85

Upside continuation rectangle measuring objective.

is known as a retracement move, a n d it offers an additional entry point, often
u n d e r substantially less emotional conditions. T h e r e t r a c e m e n t serves two
functions. First, it helps to correct the excessive emotion associated with the
breakout a n d bring people back to earth. From here, it is possible for the new
trend to extend on a m o r e sound basis. Second, it acts as a test of the breakout. A downside retracement will find support at the breakout point, a n d an
upside o n e will find resistance at the lower boundary of the pattern as these
two zones reverse their former roles.
Retracements, then, represent normal price behavior, and although they
can be frustrating, they are nothing to get concerned about. Indeed, the
breakout itself is often a volatile, illiquid affair as o n e side or the other heads
for the entrance or exit, d e p e n d i n g on the direction of the breakout. As a
result, orders are often executed with h o r r e n d o u s fills. Price activity during
the retracement process, on the other hand, is relatively quieter. This means
that buying or selling can be undertaken in a m u c h m o r e controlled environment. Figure 6-13 shows that it is often a good idea to wait for a retracement in a rising trend and buy as the price signals that the retracement is over.

Cancellations
Rctracement Moves
A great deal of the time, w h e n t h e price breaks o u t from a pattern, the initial thrust is followed by a corrective move back to the u p p e r or lower
reaches of the formation, d e p e n d i n g on the direction of the breakout, t h i s

If the minimal objective proves to be the ultimate extension of the new trend,
a substantial a m o u n t of accumulation or distribution, whichever is appropriate, will typically have to occur before prices can move in their previous
direction. Thus, a two-year rectangle might be completed a n d the upward

86

Part II: Traditional Patterns

Using Rectangles, a Case Study for All Patterns

87

T h e 3 p e r c e n t rule was developed in the first p a r t of the twentieth century, w h e n m a r k e t participants' h o l d i n g periods were m u c h longer a n d
weekly a n d monthly charts were m o r e popular. Today, with the use of intraday charts, 3 p e r c e n t could r e p r e s e n t t h e complete move a n d t h e n some!
I have no basic objection to the 3 p e r c e n t rule for longer-term price movements, where the fluctuations are m u c h greater. However, the best a p p r o a c h
is a c o m m o n s e n s e o n e based on experience a n d j u d g m e n t in each particular case. It would be very convenient to be able to say that anything over
a specific percentage a m o u n t represents a valid breakout, b u t unfortunately,
a lot d e p e n d s on t h e time frame b e i n g considered a n d the volatility of the
specific security in question.
Figure 6 - 1 3

Buy on the retracement breakout.

price objective reached. Even t h o u g h a further price rise does n o t take place,
it is still usually necessary for a top (distribution) of approximately the s a m e
size as t h e previous accumulation (in this case, two years) to be formed before
a valid d o w n t r e n d can take place. An example is shown in Fig. 6-14.

Confirmation of a Valid Breakout
So far, it h a s b e e n assumed that any move o u t of the price pattern, however small, constitutes a valid signal of a t r e n d reversal (or resumption, if
t h e p a t t e r n is o n e of consolidation). Q u i t e often, misleading moves k n o w n
as w h i p s a w s occur, so it is helpful to establish certain criteria to minimize
t h e possibility of misinterpretation. Conventional wisdom holds that y o u
s h o u l d wait for a 3 p e r c e n t penetration of the boundaries before c o n c l u d i n g
that t h e b r e a k o u t is valid. A rule of this n a t u r e filters o u t a substantial n u m b e r of m i s l e a d i n g moves, even t h o u g h t h e resulting signals are less timely.

Figure 6-14

The cancellation effect.

For example, electric utilities are very stable in their price action comp a r e d to mining stocks, where t h e volatility is far greater. Applying t h e same
p e r c e n t a g e b r e a k o u t rule to b o t h obviously doesn't make sense. W h a t constitutes a decisive breakout, with the chances of a whipsaw b e i n g considerably reduced, is thus very m u c h a matter of personal j u d g m e n t based on
experience, trial, a n d error. This j u d g m e n t should take into consideration
such factors as the type of t r e n d being m o n i t o r e d , the volatility of the security, volume, a n d m o m e n t u m characteristics.
Ironically, a false b r e a k o u t h e r e a n d there actually adds validity to a specific s u p p o r t or resistance zone. Say, for example, that after great effort I
am able to j u m p from the basement t h r o u g h the ceiling of the g r o u n d floor,
b u t t h e effort to do this leaves me with j u s t my h e a d above t h e floor a n d my
arms a n d h a n d s also resting on the floor. Technically, I suppose I am on the
g r o u n d floor, b u t it's still taking an awful lot of effort to maintain this position. Unless I am able to drag my feet up a n d actually stand on t h e g r o u n d
floor, I am likely to slip back t h r o u g h the ceiling a n d e n d up where I started.
In this situation, my false "breakout" above resistance (the ceiling) has
emphasized the i m p o r t a n c e of that resistance. It also means that I have
e x p e n d e d a great deal of effort trying to get to the g r o u n d floor, and, as a
result, I now n e e d an e x t e n d e d p e r i o d in which to gain e n o u g h strength to
make a n o t h e r attempt. Figure 6-15 shows an example of a rectangle containing a whipsaw breakout.
A n o t h e r factor that can h e l p in deciding early on w h e t h e r a b r e a k o u t is
valid is that a valid b r e a k o u t should h o l d for several periods. For example,
you may observe a decisive upside b r e a k o u t from a rectangle on a daily
chart, b u t if t h e price c a n n o t h o l d above the b r e a k o u t level for m o r e t h a n
o n e day, the signal is highly suspect. Often t h e technical position is worse
after such breakouts because those that c a n n o t h o l d indicate exhaustion,
a n d exhaustion moves are often followed by strong price trends in the direction opposite to that indicated by the (false) breakout. At the very least we would expect
an exhaustion b r e a k o u t to be followed by an e x t e n d e d p e r i o d of trading
within t h e body of t h e p a t t e r n until the technical structure can r e g r o u p .

88

Part II: Traditional Patterns

Using Rectangles, a Case Study for All Patterns

Figure 6-16

Figure 6-15 Identifying a whipsaw breakout.

O n c e the price moves away from the pattern, the odds favor a valid breakout. But what h a p p e n s if the price breaks back into the body of the pattern
again? Is the breakout still valid? If not, where is the p o i n t at which you pull
the plug on the whole thing?

Anticipating When Things Might
Go Wrong
O n e of the first things that should be d o n e u p o n entering any business venture is weighing the possible risk against the potential reward. T h e same is
true in the financial markets. Most p e o p l e , u p o n seeing a price break out
from a pattern, focus on potential profits as they calculate the p r o b a b l e
upside objective. Experienced professionals, on the o t h e r h a n d , always consider the risk as an equal, if n o t a m o r e important, part of the equation. If
the reward/risk ratio is n o t greater than 3:1, the trade or investment is probably n o t worth initiating.
This means that whenever you are planning on o p e n i n g a new position
based on a price pattern breakout, it is important that you decide ahead of time
what type of price action would cause you to conclude that the breakout was a whipsaw. In this exercise, we n e e d to r e m e m b e r that when an upside breakout
develops, the probabilities favor higher prices. This continues to be the case
during the retracement move. However, as soon as the price moves back into
the body of the pattern, the odds of higher prices begin to narrow. T h e question is, "When do the odds move below the 50 percent point?" Unfortunately,
there are no hard and fast rules that can be said to work on all occasions. Each
situation has to be judged on its own merits, a n d it's better to do the exercise
before you enter the position than while you are holding it. Otherwise emotion will creep into what will probably turn out to be an ad hoc decision.

89

Identifying a whipsaw breakout.

T h e first step is to bear in m i n d that o n c e the u p p e r b o u n d a r y of the
pattern has b e e n b r e a c h e d following an upside breakout, this is similar to
saying that support has b e e n violated. T h e same will be true, but in reverse,
for a downside breakout. This puts us on red alert until either the price
breaks back out of the pattern again or o t h e r support areas are violated. You
can see that in Fig. 6-15, the breaking of the u p p e r b o u n d a r y of the formation also signals a series of declining peaks a n d troughs. In my view, this
would be sufficient evidence to exit the position. Markets are no m o r e a n d
no less than an expression of people in action. Since individuals can and do
change their minds, so can markets. You will be far better off paying attention
to the market's message than to your own personal hopes a n d fears. W h e n
the situation is no longer flagging high probabilities of a price rise, it is better
to take a small loss than to let pride a n d stubbornness lead to a big o n e .
Let's say that the price fell straight back into the pattern without the benefit of a peak-and-trough reversal. What should we do then? It very m u c h
d e p e n d s on the chart. If t h e r e are no obvious s u p p o r t points, many traders
believe that a penetration of the 50 p e r c e n t mark is the place to exit. In this
case, the 50 p e r c e n t mark is the central point between the two horizontal
lines that make up the rectangle. An example is shown in Fig. 6-16. In this
case, the signal to sell would develop as the price crossed the 50 percent level.
Figure 6-17 shows a n o t h e r example in which the price breaks below a previous m i n o r low. In this instance, the break develops within the pattern.

Figure 6-17

Identifying a whipsaw breakout.

Part II: Traditional Patterns

90

Remember, the minor low is a potential support point. If the price breaks
below b o t h the u p p e r b o u n d a r y a n d the m i n o r low, then two support points
have b e e n violated. Such action is certainly n o t what was expected during
the original breakout. If you were considering a purchase, there would be
no g r o u n d s for buying at this point. Why, then, if you are long, should the
decision be any different? It shouldn't, of course, b u t it is often difficult to
take a loss. Invariably h o p e of a rally seeps into the psyche, a n d a rationale
for staying with the position develops. Unfortunately, markets are n o t as sentimental as their participants a n d often show no mercy. That's why it is important to take quick action as soon as the probabilities of an advance decrease.
A n o t h e r possibility is shown in Fig. 6-18, where it was possible to construct
a small up trendline a n d observe its violation. T h e p e n e t r a t i o n of that line
t h e n serves as a support violation a n d a sell signal.
In all these examples, it would be i m p o r t a n t to place a stop below the various s u p p o r t points: the 50 p e r c e n t mark, the previous low, the up trendline, a n d so on. As previously m e n t i o n e d , these stop points should be set
ahead of time. By doing this, you will have calculated the loss that you are
willing to accept a n d the p o i n t at which the original premise for the trade—
i.e., the breakout—is no longer operative. Failure to take such action ahead
of time m e a n s that w h e n things go wrong, the actual decision to sell is m o r e
likely to be based on emotional stress caused by a reaction to a news event
or s o m e t h i n g similar r a t h e r than on a logical, preset plan.
T h e examples we have dealt with h e r e relate to upside breakouts, since
that is the direction in which most m a r k e t participants look. However, in
the case of short positions initiated t h r o u g h downside breakouts, the principles r e m a i n the same, except that the direction is reversed. In this situation, resistance areas are substituted for support. An example is shown in
Fig. 6-19 where a stop would be placed above the last rally experienced by
the p a t t e r n prior to the downside breakout. A move above this p o i n t would
n o t necessarily invalidate the formation. However, it would certainly place

Using Rectangles, a Case Study for All Patterns

91

Figure 6-19 Placing a stop after a downside breakout.

the odds of its succeeding below 5 0 / 5 0 , whereas at the time of the downside b r e a k o u t they would have b e e n well above 5 0 / 5 0 . Consequently, if the
advance above the dashed trendline invalidates the reason for going short,
why c o n t i n u e to hold the trade?

Volume Considerations
So far we have considered only price in our analysis, but volume is an important i n d e p e n d e n t variable that can h e l p us obtain a m o r e accurate reflection of crowd psychology. To quickly recap, volume usually goes with the
trend, i.e., it expands with a rising trend of prices a n d contracts with a declining o n e . This is a n o r m a l relationship, a n d anything that diverges from it
should be considered a warning sign that the prevailing price trend may be
in the process of reversing. Volume is always m e a s u r e d in relation to the
recent past. Thus, heavy volume is related to volume 20 to 30 bars or so ago,
not to volume, say, 10 years ago, as institutional changes may have permanently increased the level of activity.
In the case of the rectangle, a n d with most o t h e r patterns, it is n o r m a l
for the t r e n d of volume to contract as the formation develops. Activity may
continue to fluctuate along with the price, b u t with the benefit of hindsight,
we would expect to see the various peaks a n d troughs of volume shrink as
the pattern develops, along the lines of Fig. 6-20. As the pattern nears completion, disinterest prevails a n d volume often dries u p .
T h e quality of an accumulation formation is certainly improved if volume
expands on the upside break. Sometimes it is even possible to draw a trendline j o i n i n g the lower-volume peaks, as shown in Fig. 6-20. It is the upward
surge in trading activity that confirms the validity of the b r e a k o u t because
it signals the enthusiasm of buyers. A similar move on declining volume

Figure 6 - 1 8

Identifying a whipsaw breakout.

Part II: Traditional Patterns

92

Figure 6 - 2 0

Volume trend shrinks as a rectangle is formed.

would be suspect and would indicate a failure of volume to move with the
trend. An example is shown in Fig. 6-21. In this instance, volume definitely
declines as the price is breaking out. Such action typically signals that prices
are advancing more because of a lack of sellers than because of strong,
enthusiastic buyers. As the price starts to slip, volume picks up noticeably,
suggesting that the price is slipping because of selling pressure. This is a definite sign that increases the possibility that the breakout is a whipsaw.

Figure 6-21

Shrinking volume on an upside breakout is bearish.

Using Rectangles, a Case Study for All Patterns

93

You will often see charts in which successful breakouts develop with no obvious change in volume, either on the upside or the downside. Unfortunately,
this is a fact of life. Thus a good expansion is a desirable, but not necessarily
a mandatory, condition for a valid breakout. It certainly increases the odds,
but other indicators, such as oscillators, could also tip the balance. If volume
actually declines on the breakout, as in Fig. 6-21, this is more than a missing
piece of positive evidence; it is an actual negative factor.
Figure 6-22 shows a downside breakout from a rectangle. The same
shrinking volume characteristics are present during the development of the
pattern as were present for the bullish variety. However, volume characteristics on a downside breakout are less critical. This is because it is normal
for volume to contract as prices decline. Thus, contracting volume on a
breakdown is perfectly normal. What is not typical, though, is for volume
to expand on a downside move. This in itself suggests that sellers are more
motivated and therefore adds an additional negative flavor to the pattern.
More often than not, prices will reverse and produce a small recovery or
retracement rally following the downside breakout (Fig. 6-23). This advance
is invariably accompanied by declining volume, which itself reinforces the
bearish indications. It is halted at the lower end of the rectangle, which now
becomes an area of resistance. The same idea of declining volume should
accompany a retracement move that follows an upside breakout.
Figure 6-24 shows an example in which both price volatility and volume
shrink dramatically. This combination indicates an extremely fine balance
between buyers and sellers that exists over an extended period. Normally a
price objective is determined by the depth of the formation. In this case,

Figure 6 - 2 2

Expanding volume on a downside breakout is bearish.

94

Part II: Traditional Patterns

Using Rectangles, a Case Study for All Patterns

95

Chart 6-3 St. Jude Medical, daily.

Figure 6-23 Volume should shrink on a retracement rally.
t h o u g h , the finely balanced s u p p l y / d e m a n d situation is usually followed by
a far greater a n d sharper move t h a n would be indicated by the n o r m a l measuring techniques. Figure 6-24 shows a sharp downside breakout, b u t the
principle of rapidly declining volume followed by a h u g e expansion applies
equally as well to an upside breakout. In this instance, volume typically
explodes as we move from a situation in which there is virtually no interest
by either party to o n e in which buyers c a n n o t get e n o u g h of the security at
any price. Such are the ingredients for the start of a dramatic rally. An

example is shown in Chart 6-3, for St. J u d e Medical, where a very narrow
rectangle developed with a dramatic d r o p in volume. W h e n the volume
e x p a n d e d , a short b u t sharp rally followed.

Summary
• Price characteristics: A trading range b o u n d e d by two parallel trendlines.
• Volume considerations: Volume shrinks as the p a t t e r n is formed. It is better
if it expands on an upside breakout; it is immaterial on a downside breakout.
• Measuring implications: T h e d e p t h of the pattern is projected in the direction of t h e b r e a k o u t from the b r e a k o u t point.
• Signs of false breakouts: Shrinking volume on an upside breakout.
• Benchmarks for upside failure: T h e price falls back into the p a t t e r n a n d violates a previous m i n o r low, violates an up trendline j o i n i n g previous m i n o r
lows, or crosses below the 50 p e r c e n t level of the pattern.
• Benchmarks for downside failure: T h e price rallies back into the p a t t e r n a n d
breaks above a previous m i n o r high, violates a down trendline j o i n i n g previous m i n o r highs, or crosses above the 50 p e r c e n t level of the pattern.

Figure 6-24 Narrow rectangle and nonexistent volume are often followed by a sharp move.

Head and Shoulders

97

7

Head and
Shoulders
T h e h e a d a n d shoulders is probably the most notorious of all patterns. It
forms at tops a n d bottoms as a reversal formation a n d also develops d u r i n g
an o n g o i n g t r e n d as a c o n t i n u a t i o n or consolidation p h e n o m e n o n .
C o m p a r e d to o t h e r patterns, such as triangles, the h e a d a n d shoulders has
the reputation for being o n e of the most reliable.

Figure 7-1

Classic head-and-shoulders top.

volume leads price. T h e real tip-off that an H&S pattern is developing comes
with the formation of the right shoulder, which is invariably accompanied by
distinctly lower volume than the head or the left shoulder. Ideally, the level
of volume contracts as the right-shoulder rally unfolds. In many market place
examples you will find that even though the volume characteristics differ from

Head and Shoulders as a Reversal
Pattern
Figure 7-1 shows a typical head-and-shoulders (H&S) distribution pattern.
It consists of a final rally (the head) separating two smaller, although not
necessarily identical, rallies (the left and right shoulders). If the two shoulders were trends of intermediate duration, the first shoulder would be the
penultimate advance in the bull market, a n d the second would be the first
b e a r market rally. T h e h e a d would, of course, represent the final intermediate rally in the bull market. T h e line j o i n i n g the bottoms of the two shoulders is called the neckline. W h e n the neckline is violated on the downside,
the pattern is completed a n d a reversal signal is given. To be valid, any reversal p a t t e r n must have s o m e t h i n g to reverse. A head-and-shoulders top must
therefore be p r e c e d e d by a strong i n b o u n d u p t r e n d .
Volume characteristics (see Fig. 7-2) are important in assessing the validity of these formations. Activity is normally heaviest d u r i n g the development
of the left shoulder a n d also tends to be quite heavy as prices approach the
peak. If the left shoulder rally is accompanied by a higher level of activity
than the head, this fits nicely into the principle, outlined in Chapter 5, that
96

Figure 7-2 Classic head-and-shoulders top with volume.

98

Part II: Traditional Patterns

those described here, the pattern still "works." Thus, the volume patterns are
a guide. If they are present, the formation is m o r e likely to work, b u t just
because they are n o t p r e s e n t doesn't m e a n that the pattern should be
ignored. Alternatively, if you can spot the price a n d volume action that are
representative of a head-and-shoulders top, this can be used as a guide in
o r d e r to anticipate the final completion of the pattern.
Figure 7-1 also illustrates the fact that the violation of the neckline represents a signal that the previous series of rising peaks a n d troughs has now
given way to at least o n e declining peak a n d trough. T h e right shoulder represents the first lower peak, a n d the b o t t o m of the move following the
breakdown, a lower t r o u g h .

Possible Psychology Causing These Formations
T h e psychology underlying a head-and-shoulders formation will d e p e n d very
m u c h on the time frame u n d e r consideration. If it develops on a weekly
chart following an extended rally lasting several m o n t h s or more, bullish sentiment is likely to be far m o r e e m b e d d e d in crowd psychology than if a headand-shoulders top is identified in, say, a chart of 10-minute bars. Not
surprisingly, the decline signaled by a h e a d a n d shoulders on a weekly chart
will be far greater, since the bullish psychological p e n d u l u m epitomized by
the p a t t e r n will take far longer to swing to the bearish e x t r e m e from which
an upside reversal could be expected. However, there are c o m m o n characteristics in the d e v e l o p m e n t of these patterns that apply to all time frames.
T h e difference is that smaller patterns on intraday charts are like a quick
slap on the wrist that is quickly forgotten, whereas multiyear patterns on the
monthly charts are m o r e akin to an amputation.
Let's consider how this m i g h t be applied to a daily chart. In Fig. 7-1, the
left shoulder was really the e n d of a pretty good rally. T h e psychology at this
p o i n t is pretty bullish, so w h e n a setback develops a n d prices take out the
left-shoulder high, m a r k e t participants expect the t r e n d to c o n t i n u e , which
it does. In most cases, the rally that forms the h e a d is quite large. Since rising prices p r o m o t e bullish sentiment, the crowd is pretty happy at the top
of the h e a d . Often t h e level of volume is smaller on the h e a d than on the
left-shoulder rally. This is a sign that u n d e r the surface t h e r e is less buying
power supporting the advance than was previously the case. Also, if you are
tracking an oscillator, it may form a negative divergence between the two
rallies, thereby pointing up o t h e r subsurface problems. [For a complete
description of oscillators, please see Martin Pring on Market Momentum
(McGraw-Hill, 2002) workbook/CD-ROM tutorial.] These volume a n d oscillator discrepancies are n o t signals to sell b u t m o r e of a warning that prices
are vulnerable in the case of a trend-reversal signal by the price itself.

Head and Shoulders

99

W h e n the h e a d rally is over, prices fall back to s u p p o r t a r o u n d the level of
the previous low. At that point, buyers who missed t h e advance are encouraged to enter the market. If prices rose from this level before, surely this represents a bargain and they will rise again. Owners of the security, for their part,
feel the same way a n d are less inclined to sell. Each rally up to this point has
taken prices progressively higher, so there is no reason to suspect that the next
advance will be any different. However, this o n e turns o u t to be quite weak.
Volume shrinks as it progresses, indicating that the rise is caused m o r e by a
lack of selling than by an influx of buying. W h e n selling does start to pick u p ,
the price reaches the top of the right shoulder and begins to slip. Participants
are still bullish at this point, expecting a small pullback before a rally to new
highs. However, to the technician, a decline in volume as prices rise on the
right shoulder is a bearish sign because it indicates far less enthusiasm than
was indicated by the higher volume on the left s h o u l d e r a n d head. Finally,
when the price breaks below the level of the two previous bottoms, the multitude of buyers who b o u g h t on the left shoulder and h e a d are locked in with
a loss. T h e r e is a tendency to sell, but where are the buyers? As it turns out,
they e x p e n d e d their potential on the way u p , when t h e news was good and
bullish sentiment contagious. In some instances bad news surfaces and the
price experiences a very sharp sell-off. Alternatively, we may see a series of
sharp rallies a n d reactions taking the form of a downward zigzag. H o p e alternates with despair as prices gradually work their way lower.
People can a n d do c h a n g e their minds, a n d events also change. For this
reason, quite a few head-and-shoulders tops are quickly cancelled out as a
base is builtjust below the neckline. T h e n the bearish psychology associated
with the head-and-shoulders breakdown is replaced with positive sentiment
emanating from the b r e a k o u t from the base. Prices are t h e n free to rise
again, b u t m o r e on that later.

Measuring Implications
T h e measuring formula for this price formation is t h e m a x i m u m d e p t h of
the pattern, which in the case of the h e a d a n d shoulders m e a n s the distance
between the top of the h e a d a n d the neckline. This objective is t h e n projected downward from the neckline at the breakdown p o i n t (Fig. 7-3). It follows that the d e e p e r the pattern, the greater its bearish significance once
it has b e e n completed. Sometimes a head-and-shoulders completion will be
followed by a fairly extensive downtrend; at o t h e r times the negative effect
of the pattern will be quickly cancelled by the c o m p l e t i o n of a base. This
means that you should n o t assume that all breakdowns lead to large profits (from the short side). Always k e e p an o p e n m i n d as the technical condition changes.

Part II: Traditional Patterns

100

Head and Shoulders

101

Chart 7-2 Commercial paper yield, 1910-2003, monthly (logarithmic scale).

Figure 7-3

Head-and-shoulders top measuring implications.

Arithmetic versus Logarithmic Scaling
For longer-term charts, the choice between arithmetic and logarithmic scaling can be of critical importance. Where large price movements are involved,
the difference in the price objectives can be considerable. Charts 7-1 and 7-2
offer an e x t r e m e example. Chart 7-1 shows the three-month commercial

Chart 7-1 Commercial paper yield, 1910-2003, monthly (arithmetic scale).

p a p e r yield plotted on an arithmetic scale. T h e head-and-shoulders projection calls for the rate to decline to a negative number, clearly a very unlikely
scenario. (I say unlikely because Switzerland h a d negative rates in the late
twentieth century to discourage foreigners from investing in Swiss francs.)
Chart 7-2, on the o t h e r h a n d , shows the same period in a logarithmic m o d e .
Note how the downside objective was far m o r e realistic, calling for the yield
to fall to s u p p o r t in the area of its 1940s low.
T h e objective is called a " m i n i m u m ultimate" target because prices often
move m u c h further than the objective. I n d e e d , they often progress in multiples of it. Chart 7-3, featuring the G e r m a n g o v e r n m e n t b o n d ( B u n d ) ,
shows that the price declined by t h r e e times the a m o u n t of the objective.
Notice that the horizontal line marking the initial objective b e c a m e a resistance point later on as the price rallied again. Therefore, when you are looking for potential s u p p o r t areas following a downside head-and-shoulders
breakout, it's a good idea to use multiples of the h e a d / n e c k l i n e measuring
objective as a guide. In this case it would have worked quite well. T h e problem, of course, is you do n o t know ahead of time which, if any, multiple of
the price objective will t u r n o u t to be support.

B a r C h a r t s v e r s u s C l o s e - O n l y o r Line C h a r t s
Bar charts have the advantage that they reveal all of t h e price action that
takes place in the specific period covered. O n e of the problems is that prices

102

Part II: Traditional Patterns

Chart 7-3 German Bund, 1998-2003, weekly.

are often subject to r u m o r a n d r a n d o m events that have little or n o t h i n g to
do with the trend. This m e a n s that bar charts are far m o r e susceptible to
false or misleading moves.
Line charts are not i m m u n e to whipsaw signals, but they certainly filter out
a lot of them. Often it is easier to spot a pattern using close-only charts, since
they show the underlying trend in a m o r e graphically pleasing manner.
Charts 7-4 and 7-5, featuring XL Capital, represent the same time period in
a bar a n d a line format. Notice that there were two whipsaws in the head-andshoulders formation in the bar chart, b u t these do not appear on the closeonly chart.
T h e neckline for the subsequent reverse head-and-shoulders was also easier to construct on the close-only chart, since it was a m o r e horizontal a n d
therefore significant line.
Often, traders observe the formation of a head-and-shoulders top and take
action in anticipation of a breakdown. This is an incorrect tactic because
based on this evidence alone it is n o t known until later w h e t h e r the prevailing
trend will continue, or w h e t h e r a reversal signal will be given by a decisive
break below the neckline. If a substantial n u m b e r of o t h e r indicators are
pointing in the direction of a trend change, that is a n o t h e r matter. Taken
on its own, though, an incomplete p a t t e r n c a n n o t signify a t r e n d reversal.
Over the years, I have seen many analysts who should know better forecast
a bearish trend based on an incomplete head-and-shoulders top. Remember,
in technical analysis, the prevailing trend is assumed to be in force until the iveight

Head and Shoulders

103

Chart 7-4 XL Capital, 1997-2002, weekly.

of the evidence proves otherwise. An incomplete h e a d a n d shoulders is n o t evid e n c e , j u s t a possible scenario.
H&S patterns can be formed in 10 to 15 minutes or take decades to
develop. Generally speaking, the longer the period, the greater the a m o u n t
of distribution that has taken place, a n d therefore the longer the ensuing

Chart 7-5 XL Capital, 1997-2002, weekly.

104

Part II: Traditional Patterns

Head and Shoulders

105

bear trend. T h e larger H&S formations are often very complex and comprise several smaller ones, as shown in Fig. 7-6.

Upward- and Downward-Sloping Patterns
T h e H&S p a t t e r n s illustrated in Figs. 7-1 a n d 7-2 h a d a h o r i z o n t a l neckline, b u t t h e r e are m a n y o t h e r varieties, all of which have t h e same bearish implications as t h e h o r i z o n t a l variety o n c e they have b e e n c o m p l e t e d .
W h e n the neckline of a horizontal p a t t e r n has b e e n violated, a series of
declining peaks a n d troughs is signaled. Not so with the upward-sloping variety shown in Fig. 7-4, because at the time of the breakout the price is still
above its previous low. By the same token, the series of declining peaks and
troughs has already b e e n set in motion when the neckline of a downward
h e a d a n d shoulders (Fig. 7-5) has b e e n violated. Chart 7-3, for the German
Bund, shows an upward-sloping pattern, a n d Chart 7-6, for U n i o n Planter's,
features a downward-sloping variety. In this instance, the price eventually sold
off to a four times multiple of the measured price objective. In terms of p u r e
price objective, the downward-sloping formations for any given depth are
m o r e bearish than their horizontal or upward-sloping counterparts. This is
because the breakdown from the neckline takes place at the lowest point of
the pattern. By the same token, we must be careful with a neckline that has
an extremely sharp angle of descent, as it is m o r e likely to be followed by a
false break. This is an extension of the rule relating to the steepness of the
angle of ascent or descent of a trendline, discussed in Chapter 4. After all,
a neckline is n o t h i n g m o r e than a special trendline.

Figure 7-4 Upward-sloping head and shoulders.

Figure 7-5 Downward-sloping head and shoulders.

Complex Patterns
Occasionally the battle between buyers a n d sellers is exceptionally complicated. This leads to the formation of what we call a c o m p l e x h e a d a n d
shoulders, in which t h e r e could be m o r e than o n e right or left shoulder.
Alternatively, t h e h e a d itself could be a head-and-shoulders top in its own

Chart 7-6 Union Planter's, 1986-2000, weekly.

106

Part II: Traditional Patterns

Head and Shoulders

107

shoulders, two on the left a n d two on the right. You can also see that the
dashed trendline is really the neckline of a small head-and-shoulders top
that forms the head. As a general rule, the m o r e complex the pattern, the
m o r e intense the battle between buyers a n d sellers, a n d the m o r e intense
the battle, the greater the implied significance of the new trend when it
begins.

Characteristics of the Formation

Figure 7-6 Complex head-and-shoulders top.

right. T h e possibilities are e n o r m o u s , so Fig. 7-6 is offered as a starting
point. Notice that the top shown in C h a r t 7-7 contains two right shoulders
a n d therefore qualifies as a c o m p l e x p a t t e r n . T h e r e is no theoretical limit
to the e x t e n t of complexity in a p a t t e r n . C h a r t 7-7, for Watson
Pharmaceutical, in fact shows a very intricate p a t t e r n consisting of four

A head-and-shoulders formation can take on many different characteristics.
Since the breakdown point is the critical one, the nature of the right shoulder deserves some additional attention. After all, this is the trapdoor through
which prices will d r o p . Figure 7-7 shows three possibilities that indicate a particularly interesting battle between buyers a n d sellers. T h e first displays what
I call "a right-shoulder shakeout." This develops when it looks as if the shoulder is developing in a fairly controlled way. T h e n the price falsely breaks above
resistance in the form of a previous trendline or high. This whipsaw action is
unnerving for bull a n d bear alike. As a result, this unwanted volatility results
in an unusually nasty decline. An example is shown in Chart 7-8, for Anheuser
Busch just prior to the 1987 crash. T h e false dashed trendline break on the
right shoulder is quickly followed by a decline. T h e n the trading action goes
very quiet; the price barely moves and volume shrinks to almost nothing, indicating a fine balance between buyers and sellers. However, it is the calm before

Chart 7-7 Watson Pharmaceutical, 1999-2002, daily.

Figure 7-7 Head-and-shoulders top variations.

108

Part II: Traditional Patterns

Chart 7-8 Anheuser Busch, daily.

the storm. Finally the pattern is completed a n d the price declines precipitously.
Example B in Fig. 7-7 also portrays a substantial a m o u n t of instability,
which again leads to a sharp price decline. T h e p r o b l e m with this type of
wide-ranging trading is that the breakdown is often followed by additional
choppy action. Thus, while the ultimate profit achievement from taking a
short position is usually well rewarded, the roller-coaster ride to get there
can often be extremely nerve racking. T h e figure shows a pretty straightforward decline. However, Chart 7-9, for Baker H u g h e s , shows a m o r e typical o u t c o m e , as the choppy early 1998 price action continually threw the
validity of the b r e a k o u t into doubt.
The third example in Fig. 7-7 offers a m u c h m o r e controlled right shoulder. In this instance, the price forms a triangle (see Chapter 9 for a full description of triangles). This offers the best of both worlds. First, you can see that
the rallies are getting progressively weaker. Second, the lower trendline offers
a great b e n c h m a r k against which to sell or go short. Finally, if the triangle fails,
there is a close point at which an intelligent stop can be placed, i.e. J u s t above
the u p p e r trendline. Also, the converging nature of the lines indicates a fine
balance between buyers a n d sellers. If this is resolved in favor of the sellers,
the bearish overtones of the whole formation indicate a potentially sharp
decline. An example featuring an upward-sloping head and shoulders is
shown in Chart 7-10, for J.P. Morgan. T h e right shoulder itself is fairly complex. T h e top of the shoulder is a small head and shoulders in its own right.

Head and Shoulders

109

Chart 7-9 Baker Hughes, 1997-1999, daily.

T h e shoulder itself is also a h e a d a n d shoulders, with the April rally being the
left shoulder a n d the triangle, contained within the two dashed lines, the right
shoulder. Generally speaking, when you get a complex pattern, it is a reflection of a significant battle between buyers a n d sellers. In this instance, the win
by the sellers resulted in a pretty dramatic decline.
Chart 7-10 J.P. Morgan, 2000-2002.

110

Part II: Traditional Patterns

Head and Shoulders

111

Reverse or Inverse Head
and Shoulders
Head-and-shoulders patterns also form at market bottoms. Figure 7-8 shows
an example. This is usually called an inverse H&S, a reverse H&S, or an H&S
bottom. It consists of the final decline separated by two smaller ones.
Normally, volume is relatively high at the bottom of the left shoulder
(Fig. 7-9) and during the formation of the head. The key volume characteristic is activity on the right shoulder, which should contract during the
decline to the trough and expand substantially on the breakout. The inverse
(accumulation) H&S, like the H&S distribution pattern, has a number of variations in such areas as trendline slope, number of shoulders, and so on. As
with tops, the more complex the formation, the greater its significance. Some
of these reverse head-and-shoulders variations are shown in Figs. 7-10 to 7-12.
Figure 7-9

Classic reverse head and shoulders with volume.

Marketplace Examples
Price objectives are based on the same principle as for head-and-shoulders
tops. The maximum depth is measured and then projected up from the
breakout point. Chart 7-11, featuring Aetna Insurance, offers a good example of a reverse head-and-shoulders pattern. This is really a complex pattern,
because if you look carefully, you can see that there are in fact two right shoulders and two left ones. Interestingly, the price objective was reached on the
first rally, though the price subsequently moved up from there. The small

Figure 7-8

Classic reverse head and shoulders.

upward-sloping dashed line, when combined with the neckline, indicates that
the right shoulder was really a symmetrical triangle (see Chapter 9).
Chart 7-12, for Alcan, shows another inverse head and shoulders; this time
the head, as contained within the two dashed lines, is actually a rectangle.
Another rectangle develops during the formation of the right shoulder. Note
how volume picks up during the rally from the final low. It contracts as the
price corrects, only to expand on the breakout. Volume does not cooperate as nicely as this all the time, But when it does, it offers a higher degree

Figure 7-10

Downward-sloping reverse head and shoulders.

Part II: Traditional Patterns

112

Head and Shoulders

113

Chart 7-11 Aetna, 1982-1986, daily.

Figure 7-11

Upward-sloping reverse head and shoulders.

of probability that the p a t t e r n is valid. In this instance, the upside objective
was slightly exceeded just prior to the 1987 crash.
Having said all that, it could be argued that this is n o t a h e a d a n d shoulders at all. Note the question mark against the left shoulder. This is because
the low of the shoulder is extremely close to that of the head—it's actually
just a bit above the right-hand part of the head, so technically this is a reverse
head-and-shoulders pattern. I b r i n g this up because I believe it's more important to apply common sense to price pattern construction than to apply strict rules.

After all, we are trying to identify battlegrounds between buyers a n d sellers from t h e p o i n t of view of forecasting t r e n d reversals. If trading action
does n o t quite m a t c h up to the prescribed rules, it really does n o t matter
as long as the formation works. T h e rules are actually guidelines for action,
Chart 7-12 Alcan, 1983-1987, weekly.

Figure 7-12 Complex reverse head and shoulders.

114

Part II: Traditional Patterns

Head and Shoulders

115

Chart 7-13 Mean, 1981-1982, weekly.

Chart 7-14 St. Jude Medical, 1983-1986, daily.

n o t cast in stone for every situation. Remember, it always comes down to
probabilities, never certainties. Following the rules exactly implies the kind
of perfection that does n o t exist in technical analysis.
T h e 1982 b o t t o m for Alcan appears in Chart 7-13, where t h e r e is a dramatic volume increase on the breakout. This is a real treat, since it indicates
extremely strong interest on the part of the buyers and signals a nice change
in psychology. T h e market itself b o t t o m e d in August, b u t this stock touched
its low in J u n e . By August the right shoulder h a d b e g u n to form, thereby
setting up a positive divergence with the overall market. T h e divergence was
confirmed with the completion of the reverse head-and-shoulders pattern
later that m o n t h . T h e upside objective was r e a c h e d at the initial rally peak.
Note the gap that developed on the second day of the breakout. Gaps are
potential s u p p o r t a n d resistance areas. See how the S e p t e m b e r correction
terminated at the u p p e r e n d of the gap. An attempt to close a subsequent
gap was m a d e right in the closing sessions of the chart.

o n c e the neckline was violated, the price h a d no difficulty in reaching its
m i n i m u m ultimate upside objective.
They say that there is m o r e than o n e way to skin a cat, a n d with price patterns t h e r e is often m o r e than o n e way in which they can be drawn. In this
spirit, C h a r t 7-15 features the reverse h e a d a n d shoulders for St. J u d e in a
different way, avoiding the weak right shoulder. Note that in this instance
the price m o r e than meets the objective, b u t this b e n c h m a r k nevertheless
turns o u t to be a good pivotal s u p p o r t level for a couple of subsequent
declines.
Sometimes the price action develops in such a way that it is quite difficult to decide w h e t h e r it represents o n e p a t t e r n or another. For example,
Chart 7-16, for Sysco, shows a head-and-shoulders b o t t o m with a nice
increase in volume on the breakout. Chart 7-17 shows exactly the same
period, b u t this time I have drawn in two parallel lines, indicating that the
pattern may really have b e e n a rectangle. T h e price even reached a two times
multiple of the indicated objective. It really doesn't matter what n a m e we
give to the price action. T h e essential point is that the price fell, there was
a battle between buyers a n d sellers (the trading r a n g e ) , a n d there was a subsequent b r e a k o u t to the upside on high volume. I bring this up because it
is extremely easy to get h u n g up on names a n d definitions. My feeling is
that you n e e d to interpret these formations in a c o m m o n s e n s e way, n o t
according to strict formulas. Always try to form an u n d e r s t a n d i n g of the

T h e question of the degree of price activity required to justify a right or
left shoulder often arises. Chart 7-14, for St. J u d e Medical, indicates a horizontal reverse h e a d a n d shoulders. T h e left shoulder is a definite rectangle a n d provides a good battle between buyers a n d sellers. T h e h e a d itself
is really a double-bottom formation (described in Chapter 8). However, the
right-shoulder decline may be pushing the envelope as far as the pattern
definition is c o n c e r n e d , since it did n o t involve m u c h price action. Even so,

116

Part II: Traditional Patterns

Chart 7-15 St. Jude Medical, 1983-1986, daily.

underlying psychology of any trading situation, because that's all that patterns are reflecting anyway. This m e a n s that if you can see a reversal of a
peak-and-trough progression at a time when the price violates a s u p p o r t or
resistance trendline, chances are that you have a reversal in trend. If volume

Head and Shoulders

117

Chart 7-17 Sysco, 1981-1983, daily.

is sympathetic to your interpretation, so m u c h the better. Remember, we give
price patterns names only so that we can m o r e easily recognize reversal p h e nomena.

Chart 7-16 Sysco, 1981-1983, daily.

Hcad-and-Shouldcrs Formations
as Continuation Patterns
H&S and reverse H&S formations often show up on the charts as continuation patterns. The measuring implications and volume characteristics
are the same as for the reversal type. The only difference is that these patterns develop during a. trend rather than at the end. Examples are shown
in Figs. 7-13 and 7-14.
Sprint PCS Group (Chart 7-18) offers a couple of good examples of consolidation head-and-shoulders tops. Notice how volume picks up noticeably
at the breakout points. In a bear trend, prices can fall of their own weight,
and this is normal. However, when activity picks up, it indicates that sellers
are definitely motivated, and thus the danger of a sharp price decline is at
its greatest. The pattern at the top of the rally, which was completed in the
fall of 2000, defies definition. It is not a head and shoulders because the
"right shoulder" reverses at the same level as the "head." Neither is it a true
rectangle because the late 1999 and early 2000 rallies fail to reach the upper

118

Part II: Traditional Patterns

Head and Shoulders

119

Chart 7-18 Sprint PCS Group, 1999-2002, daily.

Figure 7-13 Continuation reverse head and shoulders.

trendline. It is p e r h a p s best defined as a double top, a formation that is covered in the next chapter. As m e n t i o n e d earlier, how this pattern is described
is of no consequence, since an obvious trading range (battle between buyers a n d sellers) following a rally was violated on the downside. T h e day of
the lower trendline p e n e t r a t i o n also saw a h u g e expansion in volume.
Finally, South West Airlines, Chart 7-19, offers us a massive eight-year consolidation reverse head-and-shoulders pattern. Note how the first a n d seco n d multiples of the upside objective acted as support zones for subsequent

declines. T h e third multiple was barely r e a c h e d before a new bear t r e n d set
in. O n e sign of a potentially bullish stock market develops when many stocks
have b e e n consolidating the way South West was in 1991. In 1982, for example, the long-term chart books were stuffed full of multiyear consolidations.

Chart 7-19 South West Air, 1980-2003, monthly.

Figure 7-14 Continuation head-and-shoulders top.

120

Part II: Traditional Patterns
Head and Shoulders

W h e n a substantial n u m b e r of stocks a n d groups are in such a position, they
provide the foundation for a very long a n d sustainable bull market. This was
certainly true following major breakouts in late 1982 a n d early 1983. T h e
opposite would be true at a market top, where a p r e p o n d e r a n c e of large
head-and-shoulders tops in key stocks a n d industry groups would warn of
i m p e n d i n g trouble.

121

Chart 7-20 Albertson's, 1998-2003, weekly.

Head-and-Shoulders Failures
We have already established that prices are d e t e r m i n e d by crowd psychology. Individuals can a n d do c h a n g e their minds; so can crowds, a n d therefore markets. As a result, what m i g h t a p p e a r to be a perfectly valid
head-and-shoulders b r e a k o u t o n e day may well t u r n out to be a whipsaw the
next. This is generally n o t the case, b u t any trader or investor who does n o t
recognize the ability of markets to reverse otherwise perfectly legitimate signals is in a state of delusion.
T h e first step is to make sure that the p a t t e r n you are following is i n d e e d
a legitimate formation. For example, the price action may exhibit all the
characteristics of an H&S distribution pattern, b u t the price refuses to penetrate the neckline. We have already established that until the formation is
completed with a decisive break below the neckline, it is n o t a true pattern.
This is because the neckline represents a s u p p o r t area, a n d s u p p o r t has n o t
b e e n violated. In the case of a horizontal formation, failure to penetrate the
neckline also means that the series of rising peaks a n d troughs is still intact.
C h a r t 7-20 features a reverse h e a d a n d shoulders for Albertson's that did
n o t work. T h e price rallied up to the (solid) neckline for a final time in mid2002, b u t was unable to go t h r o u g h . T h e dashed trendline is there to indicate that the final part of the h e a d a n d the potential inverse right shoulder
actually formed a head-and-shoulders top. Often it is possible to spot these
technical situations where the glass is half full or half empty. In this case it
was half empty, a n d the price declined.
T h e 1992-1995 period shows two examples for Aflac in Chart 7-21. Note
that the volume on the right shoulder of the first pattern on the left was
particularly heavy. T h e second pattern's neckline was ever so slightly nicked
on t h e downside, b u t the formation never really worked. O n e indication of
the failure would have c o m e from the p e n e t r a t i o n of the trendline j o i n i n g
the peaks in the right-shoulder decline. Such moderately high-volume
upside breakouts are totally o u t of character with the n o r m a l low-volume
characteristic of a right-shoulder formation a n d could have offered an indication that the pattern was n o t going to work.
Alternatively, a failure can develop after t h e price actually p e n e t r a t e s
t h e neckline temporarily a n d t h e n reverses in the opposite direction. This

r e p r e s e n t s an H&S failure a n d is usually followed by an explosive rally or,
in t h e case of a failed inverse h e a d a n d s h o u l d e r s , a nasty d e c l i n e . C h a r t
7-22, for Albertson's, r e p r e s e n t s a g o o d e x a m p l e of what h a p p e n s after a
p a t t e r n has failed to work or simply failed. In the case of a failed top, this
Chart 7-21 Aflac, 1991-1995, weekly.

122

Part II: Traditional Patterns

Head and Shoulders

1 oo

Chart 7-22 Albertson's, 1996-1999, weekly.

Figure 7-15

is probably the result of misplaced pessimism. O n c e t h e real f u n d a m e n tals are perceived, n o t only do new buyers rush in, b u t also t r a d e r s holding s h o r t positions are forced to cover. Since fear is a s t r o n g e r motivator
t h a n g r e e d , these bears bid up t h e price very aggressively.
Failures used to be fairly rare, b u t they now a p p e a r to be m o r e c o m m o n ,
which indicates the necessity of waiting for a decisive breakout on the downside (or the upside in the case of a reverse h e a d a n d shoulders). They typically develop when the pattern suggests a break in the opposite direction to
the then-prevailing trend. Obviously, if this is the actual top or bottom, the
formation will be valid. However, when a head-and-shoulders top forms in a
bull market and does not experience a meaningful decline, this will tend to
be a countercyclical signal. In fact, the very failure of the pattern may be
interpreted as a sign that the prevailing (dominant) trend probably is still
in force.
T h e r e are several points in the chart where the probabilities of a valid signal sink below 50 p e r c e n t a n d those of an outright failure start to increase.
Figures 7-15 a n d 7-16 try to address these points. Point A in Fig. 7-15 represents the b o t t o m following the break below the neckline. T h e next rally,
which ends at B, is a perfectly typical development because retracements are
a n o r m a l , a n d i n d e e d healthy, p h e n o m e n o n . T h e price then falls to C and
something u n e x p e c t e d h a p p e n s : Instead of following t h r o u g h on the downside, as would be expected from a head-and-shoulders top, the price rises

Identifying a head-and-shoulders failure.

back to the neckline again. This is the first sign that things may not work out
as expected. When the price once again rallies back above the neckline (D),
the odds of a failure increase. The balance tips more to the bullish side when
the price moves above the down trendline joining the head with the right
shoulder (£). This is probably the time to cover all shorts, since the reason
for going short in the first place—i.e., the breakdown—no longer exists. The
nature of the trendline will have a great deal to do with the change in probabilities. For example, if the line is steep and has been touched only twice,
it will have nowhere near as much significance as it would if it were shallow

Figure 7-16 Identifying a reverse head-and-shoulders failure.

124

Part II: Traditional Patterns

a n d h a d b e e n t o u c h e d several times. A refresher on trendline interpretation in C h a p t e r 4 would be a good idea at this point.
T h e next line of defense is the right shoulder. If the price can rally above
this p o i n t (F), t h e n in some cases it will now be e x p e r i e n c i n g a series of
rising peaks a n d troughs. Finally, w h e n t h e price moves above t h e h e a d ,
the p a t t e r n is cancelled beyond a reasonable d o u b t .
If action on the long side is contemplated, it should be taken either when
the price breaks above the trendline joining the head and the right shoulder
(line E) or when it breaks above the right shoulder (F) on heavy volume.
Usually, such signals offer substantial profits in a very short period of time and
are well worth acting on. Again, some c o m m o n sense comes into play, for if
the trendline joining the h e a d and the right shoulder is unusually steep and
has been touched only twice, it will not have the authority of a more shallow
trendline that has b e e n touched or approached on n u m e r o u s occasions.
Inverse H&S patterns can also fail, as we see from Fig. 7-16. Again, the
failure is usually followed by a fairly lengthy decline as participants who
b o u g h t in anticipation of an upward b r e a k o u t are flushed out when the new
bearish fundamentals b e c o m e m o r e widely known. Note that the line joining the h e a d with the right shoulder is m o r e significant in this example than
that in Fig. 7-15. This is because the line is shallower a n d has b e e n t o u c h e d
on m o r e occasions. T h e j o i n t break with the neckline is also impressive a n d
would greatly increase the odds of a failed pattern.

Chart 7-23 Andrew Corp., daily.

Head and Shoulders

125

Chart 7-23 shows a failed head-and-shoulders top for A n d r e w Corp. This
o n e developed d u r i n g a very strong linear bull market. T h e first indication
of failure would have been given when the price broke back above the neckline after forming a small base. T h e clincher developed when the dashed
t r e n d l i n e j o i n i n g several rally peaks was b e t t e r e d on the upside. Failed patterns are often followed by dynamic moves in the opposite direction to that
indicated by the pattern. This m e a n s that these patterns should be viewed
n o t with fear, but as an o p p o r t u n i t y for profits. T h e d e g r e e of o p p o r t u n i t y
will d e p e n d on the strength of the signal a n d the closeness at which a realistic stop can be placed (the perceived risk). In this case, the t r e n d l i n e was
a very strong o n e a n d a stop could have b e e n placed j u s t below the low of
the b r e a k o u t day. Provided it was b o u g h t pretty close to the b r e a k o u t point,
this would have r e p r e s e n t e d a very-low-risk, potentially high-reward trade.

Summary
In summary, t h e r e are several clues we can look for that suggest that a pattern will fail.
• M o m e n t u m indicators at the time of the breakout that are extremely oversold in the case of a top or overbought in the case of a b o t t o m
• Heavy volume on the right shoulder of a potential head-and-shoulders
top. Light volume on a reverse head-and-shoulders breakout
• Failure of the price to follow t h r o u g h in the direction of the breakout following a r e t r a c e m e n t move
• Failure of the overall market a n d o t h e r stocks in the industry g r o u p to
act in sympathy
• T h e relative strength line failing to confirm the breakout

Head-and-Shoulders Top Review
• Price characteristics: A final rally separated by two smaller rallies at tops.
• Volume considerations: Very heavy volume on the left s h o u l d e r a n d sometimes the head. Low a n d shrinking volume on the right-shoulder rally.
Immaterial on breakdown, b u t heavy volume preferred.
• Measuring implications: T h e distance from the top of the h e a d to the neckline is projected down at the point of breakout.
• Signs of false breakout: T h e presence of a second r e t r a c e m e n t rally a n d failure to break the initial breakdown low. A rally above the trendline j o i n i n g

126

Part II: Traditional Patterns

the h e a d a n d the right shoulder, provided it is n o t unduly steep a n d / o r
above the right shoulder.

8

Head-and-Shoulders Bottom Review
• Price characteristics: A final low separated by two higher lows at bottoms.
• Volume considerations: Heavy volume on the left shoulder a n d sometimes
the head. Low a n d shrinking volume on the right-shoulder decline. Very
high volume accompanying the upside breakout.
• Measuring implications: T h e distance from the b o t t o m of the h e a d to the
neckline is projected up at the point of breakout.
• Signs of false breakout: Contracting volume on breakout. Price u n a b l e to
hold the b r e a k o u t for m o r e than two sessions.
• Places to unwind position in case ofwhipsaw breakout: Violation of up trendline j o i n i n g h e a d a n d right shoulder. Break below right shoulder, especially if a c c o m p a n i e d by e x p a n d i n g volume.

Double Tops,
Double Bottoms,
and Triple Patterns
Double Tops
A d o u b l e top consists of two peaks that are the culmination of a rally. They
are separated by a reaction or valley in prices. T h e formation is c o m p l e t e d
when the price breaks below the lower level of the valley. Figure 8-1 gives
an example. In Profits and the Stock Market, H. M. Gartley defines a double
top as "representing two unsuccessful attempts to p e n e t r a t e a supply area
with t h e resultant d i s a p p o i n t m e n t p r e c e d i n g a n d contributing to an important s u b s e q u e n t decline." T h e principal characteristic of this formation
arises from the fact that t h e second top is formed with distinctly less volume
than the first. It is n o r m a l for both peaks to form at the same price level,
but it is also possible for the second peak to slightly exceed the first or to
top o u t j u s t a little below it. T h e reaction from the first peak to the valley is
usually associated with a t r e n d of declining volume. Remember, this is n o t
an exact science, but a c o m m o n s e n s e interpretation of a battle between buyers a n d sellers.
In their book Technical Analysis of Stock Trends, Edwards a n d Magee point
out that double tops are referred to by n a m e by m o r e traders who have a small
knowledge of technical analysis than perhaps any other pattern. They go on
to point out that in most cases these neophyte technicians identify these patterns before they have b e e n completed, which is a definite no-no. According

127

128

Part II: Traditional Patterns

Double Tops, Double Bottoms, and Triple Patterns

129

idea that e n o u g h time has elapsed d u r i n g the valley phase that by the time
the price has b e g u n to a p p r o a c h the second peak, the rally is very believable, a n d most people are expecting a significant extension of the previous
u p t r e n d so the price will exceed the first peak by a wide margin.

Underlying Psychology

Figure 8-1 Double top.

to Edwards and Magee, double patterns are extremely rare. That certainly
appears to be the case for double tops, but n o t for double bottoms. In doing
research for this book, I ran through countless charts covering well over 10
years of data and was unable to find many examples of true double tops.
Edwards a n d Magee also point out that many patterns that a p p e a r to be
double tops eventually transmute into other patterns that turn out to be continuation formations rather t h a n reversals. They suggest that an important
differentiating characteristic is the time separating the two peaks. If they
develop on the daily charts, say at one-week intervals, this is m o r e likely to
result in a consolidation formation, such as a rectangle. Although they confess that there are no ironclad rules, their guideline for a double top is a
separation greater than a m o n t h with a "valley decline" of 15 to 20 percent
separating the two peaks. They admit that these rules are somewhat arbitrary, b u t they go on to p o i n t out that the time e l e m e n t is m o r e important
than the m a g n i t u d e of the valley decline. T h u s a two- or three-month separation between the two peaks requires less of a decline than, say, a onem o n t h interval. These guidelines are for daily charts. To apply t h e m to
monthly or hourly charts would not, of course, make m u c h sense because
it is the relative comparison of the time frame that is important. A possibility in this case is to take the n u m b e r of trading periods involved in the
"one m o n t h or greater" rule and apply it to the relevant bars. Thus o n e week
of daily trading is approximately 20 bars. Applying this to the weekly charts
would imply a separation of 20-plus weeks, a n d applying it to an hourly chart
would imply 20-plus hours. T h e key in all these time frames is to reflect the

T h e underlying psychology might work something like this: T h e first top
develops at the e n d of a substantial rally; volume is heavy, a n d sentiment is
very positive. T h e n prices "unexpectedly" decline on contracting volume
a n d d i s a p p o i n t m e n t sets in. This is then followed by a rally. Rising prices
attract m o r e bulls as the rally progresses, a n d the bullish arguments that
were associated with the first peak b e c o m e m o r e believable. However, this
second peak is accompanied by far lighter volume than the first. To the technician, this is a bearish factor because it indicates less enthusiasm by the buyers and suggests that prices are rallying m o r e because of a lack of sellers than
anything else. Prices subsequently decline again and break below the lower
level of the valley separating the two peaks. At that time the pattern is completed, a n d everyone who b o u g h t d u r i n g its formation is losing money a n d
is therefore a potential seller.

Other Considerations
I m e n t i o n e d earlier that the two peaks should be of roughly the same height.
Edwards a n d Magee use a 3 percent rule for this. This is again based on daily
charts. T h e key, t h o u g h , as they rightly point out, is that buying should n o t
push the second peak above the first by a decisive margin. This is because
the second peak really signifies failure—in this case, the failure of buyers
to m o u n t a strong second rally that takes the price convincingly above the
resistance indicated by the first peak. In o t h e r words, if the second top is decisively above the first, this indicates that the series of rising peaks and troughs is still
intact. Also, when the price breaks below the valley low, a second peak that
is close to the first will n o t cloud the interpretation that a break below the valley low is a signal that a new series of declining peaks and troughs is underway. What
is decisive is really a matter of experience a n d c o m m o n sense. Sometimes
valuable clues can be given by o t h e r indicators. For example, you may find
that there is a serious negative divergence between an oscillator a n d a price
at the second peak. Alternatively, a s m o o t h e d oscillator may be overbought
a n d reversing to the downside as the second peak is forming.
According to Edwards a n d Magee, the n o r m a l measuring objective cann o t be applied to the double patterns. However, I have n o t found anything

130

Part II: Traditional Patterns

wrong with the normal approach. After all, if we assume that prices are determ i n e d by psychology, a n d if psychology often moves in p r o p o r t i o n , why
should t h e distance between the highest t o p a n d t h e valley b o t t o m n o t be
projected down from the b r e a k o u t point? Thus, in my view, the m i n i m u m
downside measuring implications for d o u b l e tops should be calculated by
projecting the m a x i m u m distance between the h i g h e r of t h e two highs a n d
the valley low at the p o i n t of breakdown, as shown in Fig. 8-1. This suggested
a p p r o a c h is the same as that used with rectangles a n d head-and-shoulders
tops.

Marketplace Examples
Chart 8-1, for Jefferson Pilot, shows a double top in 1998. It's n o t a classic
pattern in t h e sense that volume dries up on the second top, b u t it definitely
has the price characteristics. T h e r e are two rallies separated by an approximately 15 p e r c e n t decline a n d three m o n t h s in time. Interestingly the
decline took the price down by a little m o r e than triple the measuring objective. If you look very carefully you can see that b o t h tops contained small
head-and-shoulders distribution patterns.
C h a r t 8-2 shows a n o t h e r double top, this time for t h e DJIA in the 1930s.
Notice that t h e July-August 1937 t o p was associated with considerably less
volume than the first peak. Prices declined substantially after the breakdown

Chart 8-1 Jefferson Pilot, 1997-1998, daily.

132

Part II: Traditional Patterns

took place. I n d e e d , this seems to be a characteristic of "double" formations:
They are either p r e c e d e d or followed, or sometimes b o t h , by a very sharp
a n d persistent price move.

Double Bottoms
T h e price action of double bottoms is exactly the opposite of that of double
tops. These formations are typically preceded by a very sharp price decline.
An example is shown in Fig. 8-2. An initial low is formed on pretty heavy volu m e , often a selling climax. A subsequent price bounce retraces some of the
ground that was previously lost. This advance then gives way to a decline that
tests the initial low. W h e n the price rallies above the "bounce" high, the pattern is completed. Typically, volume on the second low is extremely light, especially when compared to the initial bottom. It is a decided plus when the level
of activity expands on the upside breakout. Volume in double bottoms is as
easy as 1-2-3. In effect, we get high volume on the first low (1), lower volume
on the second low (2), a n d expanding activity on the breakout (3). T h e volu m e characteristics described here are not mandatory, since many patterns that
do not reflect such characteristics still appear to work. If the characteristics are
present, however, the probabilities of a valid formation will be greater.
Usually the second b o t t o m is formed above the first, b u t these formations
are equally valid whether or n o t the second reaction reaches (or even slightly
exceeds) t h e level of its predecessor.

Double Tops, Double Bottoms, and Triple Patterns

133

Underlying Psychology
Since most double-bottom formations are p r e c e d e d by a sharp decline, it's
likely that the initial b o u n c e rally, which is usually very sharp, is caused by
a lack of selling pressure a n d a panic move by the shorts to cover their positions. T h e greater the volume at the initial low, the less will be the overhang
from potential sellers. O n c e the short covering has b e e n achieved a n d bargain h u n t e r s have b e e n satisfied, prices once m o r e begin to slip. However,
since the bears have j u s t experienced a r o u g h ride, they are less inclined to
p u t out m o r e shorts. This means that selling pressure is n o t so intense.
A n o t h e r reason for a lack of selling pressure is that most of the pessimistic
holders will already have sold d u r i n g the initial decline, when they were
highly motivated. T h e second price d r o p develops m o r e from a lack of bids
than anything else. As the price a p p r o a c h e s the second bottom, volume
often shrinks to almost n o t h i n g . This is indicative of a sold out as opposed
to an oversold market. T h e r e is an old adage on Wall Street that says, "Never
short a dull market." It probably applies to the second low of a double-bottom formation. This d e a r t h of activity m e a n s that the balance between buyers a n d sellers is extremely closely matched, so the slightest event can have
a dramatic effect on the price. At this point all the bad news has b e e n disc o u n t e d a n d most of the selling is out of the way, so there is only o n e direction in which the price can move—up.
T h e final break above the b o u n c e high sets in m o t i o n a series of rising
peaks a n d troughs as it becomes evident that the initial rally was m o r e substantial than a dead cat b o u n c e .

Marketplace Examples

Figure 8-2 Double bottom.

Chart 8-3, for Lockheed Martin, shows a double bottom. Note that, as with
most of the examples shown h e r e , the pattern is p r e c e d e d by a sharp a n d
persistent decline. Also, volume at the first b o t t o m is well above that at the
second. I n d e e d , the second b o t t o m is really a small inverse h e a d a n d shoulders, as indicated by the dashed neckline. T h e contrast between the wild
price movements at the November low a n d the m o r e constrained a n d
r o u n d e d low established in February is also characteristic of a double bottom.
Chart 8-4, for Nvida, is also a double bottom, since it meets the price a n d
volume characteristics. It is p r e c e d e d by a sharp decline, then there is a successful test of the initial low on lighter volume, a n d finally there is a break
above the b o u n c e high. I n d e e d , volume expands on the day that the price
completes the pattern. T h e only missing ingredient is the time between the
two lows, which is relatively small.

134

Part II: Traditional Patterns

Chart 8-3 Lockheed Martin, 1999-2000, daily.

Double Tops, Double Bottoms, and Triple Patterns

135

Chart 8-5 Mercury Intract, 1998-1999, daily.

Mercury Intract, in Chart 8-5, shows a different type of double bottom.
In this case, the second low is well above the first. However, the three volume characteristics are present: high volume at the initial low, lower volume
on the second low, and rising volume on the breakout. Notice once again
how the price barely moves during the five days that the second bottom is

being formed. This once again indicates a close balance between buyers a n d
sellers, so that when volume expands, the price simply explodes to the
upside.

Chart 8-4 Nvida, 2000-2001, daily.

Chinese Double Bottom
If the price rallies sharply off the second bottom, buying on the b r e a k o u t
could involve a considerable price risk. This is because in many cases, the
only viable s u p p o r t point u n d e r which to place a stop is the second b o t t o m
itself. Such a situation is shown in Fig. 8-3. However, d u r i n g the formation
of many secondary bottoms, the price declines u n d e r the constraint of a
resistance trendline (Fig. 8-4). W h e n it breaks above the line, m o r e often
than n o t this signals that the pattern will be completed with a break above
the b o u n c e rally high. I call these Chinese double bottoms because the
r e t r a c e m e n t toward the secondary low can often be slow a n d very torturous
to those who are long. T h e great advantage of these Chinese double bottoms is that they provide a potentially high-reward but low-risk buying opportunity. Generally speaking, the longer the (Chinese) torture, the m o r e
bullish the situation when the breakout finally develops. Chart 8-6, featuring UST, offers a good example. We see the 1-2-3 volume pattern along with
a slow b u t steady decline held back by the dashed (torture) down trendline.
W h e n the price breaks above the trendline, a sharp rally is triggered. W h e n

136

Part II: Traditional Patterns

Double Tops, Double Bottoms, and Triple Patterns

137

Chart 8-6 UST, 1999-2001, daily.

Figure 8-3 Double bottom assessing risk.

a line is fairly steep, as this o n e is, a powerful rally often develops. Notice
the low risk, m e a s u r e d by the distance from an early b r e a k o u t to a point
just below the second b o t t o m .
C h a r t 8-7, for MBNA, indicates an even longer "torture" t r e n d l m e , b u t
the trendline develops at a shallower level than that in the previous chart.
Even so, a worthwhile rally with a relatively low risk develops. Generally

speaking, the most explosive breakouts seem to come from fairly steep torture trendlines like that in Chart 8-6. It is, however, i m p o r t a n t for t h e m to
have b e e n t o u c h e d or a p p r o a c h e d on n u m e r o u s occasions.
Finally, C h a r t 8-8, featuring Williams, indicates a n o t h e r Chinese double b o t t o m . This time t h e price rises quite a bit, b u t it does so in a m o r e
Chart 8-7 MBNA, 1987-1988, daily.

Figure 8-4 Double bottom assessing risk.

138

Part II: Traditional Patterns

Double Tops, Double Bottoms, and Triple Patterns

139

Chart 8-8 Williams, 2002-2003, weekly.

Figure 8-5 Double bottom platform.

controlled manner. W h e n the price finally rallies above the August b o u n c e
high, it has really completed a reverse h e a d a n d shoulders, with the left shoulder being formed in August, the "second bottom" being the head, a n d the
right shoulder developing in February of 2003. If a line j o i n i n g the two bottoms is extended, the whole thing could be interpreted as an ascending triangle (covered in Chapter 9). It really doesn't matter what the formation is
called. T h e important thing is that it was working as we moved into J u n e of
2003.

Platform Double B o t t o m
A platform d o u b l e b o t t o m is a variation on t h e Chinese d o u b l e b o t t o m .
In this case, t h e initial b o t t o m develops after a very s h a r p , panic-oriented
decline. T h e price t h e n rallies a n d e x p e r i e n c e s a t r a d i n g r a n g e , usually
a rectangle. T h e t r a d i n g r a n g e forms some way above t h e p a n i c low a n d
acts as a kind of platform. W h e n a b r e a k o u t above t h e platform takes
place, t h e p a t t e r n is c o m p l e t e d . An e x a m p l e of this c o n c e p t is shown in
Fig. 8-5. A m a r k e t p l a c e e x a m p l e a p p e a r s in C h a r t 8-9, featuring Sysco.
Normally t h e stop p o i n t would b e p l a c e d u n d e r s u p p o r t j u s t below the
b o t t o m of the platform. In this case, a less risky m i n o r low, set in J a n u a r y
of 2002, could have b e e n used. Alternatively, if a b r e a k o u t above the
d a s h e d t r e n d l i n e h a d b e e n used as an entry point, t h e stop could have

b e e n p l a c e d below t h e lower solid t r e n d l i n e m a r k i n g t h e b o t t o m of the
platform. A s e c o n d e x a m p l e , featuring Albertson's, is displayed in C h a r t
8-10. In this case, p a r t of t h e platform could be i n t e r p r e t e d as a consolid a t i o n reverse h e a d a n d s h o u l d e r s . T h e risk for this t r a d e would have
b e e n a b o u t 10 p e r c e n t .
Chart 8-9 Sysco, 2001-2002, daily.

1

.~

Part II: Traditional Patterns

Chart 8-10 Albertson's, 1987-1988, daily.

Double Tops, Double Bottoms, and Triple Patterns

Chart 8-11 Keycorp, 1991-1992, daily.

Double Bottoms as Consolidation Patterns
Occasionally a double-bottom formation will show up as a consolidation pattern. An example is shown in Chart 8-11, for Keycorp. In this situation, part
of the formation turned out to be a head-and-shoulders top (the dashed trendline on the left), which was quickly canceled as the second bottom was being
formed with a small reverse h e a d and shoulders (the second dashed trendline) .

D o u b l e - B o t t o m Failures
Like all p a t t e r n s , double b o t t o m s are occasionally subject to failed breakouts. Typically this will h a p p e n d u r i n g a b e a r market, w h e n t h e b r e a k o u t
is a contra-trend signal. An e x a m p l e is shown in C h a r t 8-12, for KB H o m e .
At t h e time of the breakout, this looked like a perfectly n o r m a l reversalp a t t e r n c o m p l e t i o n . T h e p r o b l e m was t h a t this f o r m a t i o n developed
toward t h e e n d of a bear market. In this case, a convenient failure indication was given when t h e price b r o k e below the d a s h e d s u p p o r t trendline. It is doubtful if this would have m a d e a g o o d r i s k / r e w a r d trade
anyway. This is because the distance between the b r e a k o u t p o i n t ($16) a n d
a place j u s t below the second low at $14 would have involved a risk of about
18 percent.

Chart 8-12 KB Home, daily.

141

142

Part II: Traditional Patterns

W h i p s a w D o u b l e B o t t o m ("Lucky S e v e n " D o u b l e B o t t o m )

Double Tops, Double Bottoms, and Triple Patterns

Chart 8-13 Electronic Data, 1990-1991, daily.

This is an unusual pattern that is measured in terms of waves a n d their relationships to previous highs and lows rather than being defined by trendlines.
An example is shown in Fig. 8-6. T h e pattern can be divided into seven waves,
h e n c e the "Lucky Seven" title. Four waves are associated with the first bottom
and three with the second. They are indicated in Fig. 8-6 by the dashed and
dotted lines. T h e idea is that prices rally off the first bottom in a robust manner as the initial three waves of the pattern form a rising peak a n d trough.
T h e n the fourth wave destroys this by breaking below the initial m i n o r bottom. This is the whipsaw part of the equation. However, it is n o t a decisive signal, for at this point the rising peaks are still intact. T h e second three waves
save the day because they result in the rising bottoms being reinstated. T h e
pattern is completed on the seventh (lucky) wave at X. A second, stronger
signal develops as the price rallies above the peak of the third wave at Y.
Often you see a pretty sharp or even explosive rally develop after the
breakout. Charts 8-13, for Electronic Data, a n d 8-14, for Intel, show a couple of examples of this formation. T h e Electronic Data formation turns out
to be a borderline reverse h e a d a n d shoulders because the horizontal
trendline almost connects with the late S e p t e m b e r high. It really doesn't
matter, for anyone who h a d b e e n following the Lucky Seven formula from
the long side would have d o n e quite well. T h e Intel chart is a m o r e clearcut example of a double b o t t o m .
Chart 8-14 Intel, 1995-1996, daily.

Figure 8-6 Whipsaw double bottom.

143

Part II: Traditional Patterns

Double Tops, Double Bottoms, and Triple Patterns

145

144
Chart 8-15 NCR, 1998-1999, weekly.
These patterns are n o t that c o m m o n , b u t when you can spot t h e m , they
usually offer a good risk/reward situation. T h e best place to go long, referring to Fig. 8-6, is at point X; provided wave 7 is n o t too long, it is t h e n possible to set a stop just below the b o t t o m of wave 6.

Triple Tops
Double patterns may e x t e n d to form triple tops or bottoms, or sometimes
even q u a d r u p l e or other complex formations. An example of a triple top is
shown in Fig. 8-7.
T h e measuring implication of all these patterns is d e t e r m i n e d by calculating the distance between the peak (trough) a n d the lower (upper) e n d
of the p a t t e r n a n d projecting this distance from the neckline. It is easy to
b e c o m e confused between triple tops, head-and-shoulders tops, a n d rectangle tops. Make sure that the highest rally is n o t the center one; if it is,
this is a h e a d a n d shoulders. Also, if t h e r e is n o t m u c h serious difference
between the three peaks, the pattern could be a rectangle. In actual fact,
all three patterns can r e p r e s e n t a m o r e or less horizontal trading range in
which buyers a n d sellers battle it out. W h e n the support line j o i n i n g any of
t h e m is violated, the formation is completed. Ideally, we would want to see
volume shrink on the third peak of a triple top. That's exactly what h a p p e n s
in Chart 8-15, for NCR. Notice also that volume expands noticeably on the
downside. This indicates selling pressure, as o p p o s e d to prices falling

because of a lack of bids. It therefore emphasizes the bearishness of the situation.

Triple Bottoms
Like triple tops, triple bottoms experience a series of three lows (see Fig. 8-8).
A line joins all the rally peaks, and the pattern is completed on the upside
breakout. Volume is normally greatest at the first two bottoms, declining noticeably on the third. Mellon Financial (Chart 8-16) provides us with an example
of a triple bottom. O n c e again volume is heavy on the August and September
lows but m u c h lighter at the October bottom.

Summary
Double-Top Review
• Price characteristics: Two highs following a worthwhile rally that form at
approximately the same level. They are separated by a decline a n d a
d e c e n t time interval.
• Pattern completed: With a decline below the valley low.
Figure 8-7 Triple top.

146

Part II: Traditional Patterns

Double Tops, Double Bottoms, and Triple Patterns

147

• Volume considerations: Heavy volume on the first top. Substantially lower volume on the second.
• Measuring implication: The distance between the higher of the two tops and
the valley low is projected down from the breakout point.
• Signs of false breakouts: Retracement of 50 percent or more of the distance
between the second top and the breakout point.
• Places to unwind the position in case of a whipsaw breakout: Retracement of
50 percent or more of the distance between the second top and the
breakout point.

Double-Bottom Review
Figure 8-8 Triple bottom.

• Price characteristics: Two lows following a serious decline that form at
approximately the same level. The two lows are separated by a rally.
• Pattern completed: With a move above the rally peak coming off the initial low.
• Volume consideration: Heavy volume on the first bottom; distinctly lower volume on the second. Very high volume should accompany the upside
breakout.

Chart 8-16 Mellon Financial, 1990-1991, daily.

• Measuring implications: The distance from the lower of the two bottoms is
projected up from the breakout point.
• Signs of false breakouts: Retracement of 50 percent or more of the distance
from the second low to the breakout point. Low volume accompanying
the breakout.
• Places to unwind the position in case of a whipsaw breakout: On a decline that
retraces more than 50 percent of the distance from the low to the breakout point.

Triple-Top Review
• Price characteristics: Following a worthwhile rally, three highs that form at
approximately the same level, where the second is not higher than the
other two.
• Pattern completed: With a decline below a more or less horizontal trendline
joining the three lows.
• Volume considerations: Heavier volume on the first two tops; substantially
lower volume on the second. These characteristics are less precise than
those of other patterns.

Part II: Traditional Patterns

148

• Measuring implications: The distance from the highest of the three tops is
projected down from the breakout point.
• Signs of false breakouts: Retracement of 50 percent or more of the distance
between the third top and the breakout point. Violation of a worthwhile
trendline joining the third top to any minor-rally highs.

9

• Places to unwind the position in case of a whipsaw breakout: Retracement of
50 percent or more of the distance between the third top and the breakout point. Violation of a worthwhile trendline joining the third top to any
minor-rally highs.

Triple-Bottom Review
• Price characteristics: Following a decline, three lows that form at approximately the same level, where the second is not lower than the other two.
• Pattern completed: With a move above an approximately horizontal trendline joining the rally peaks coming off the initial two lows.
• Volume considerations: Heavy volume on the first bottom; very high volume
accompanying the upside breakout.
• Measuring implication: The distance from the lower of the three bottoms
is projected up from the breakout point.
• Signs of false breakouts: Retracement of 50 percent or more of the rally from
the third low to the breakout point.
• Places to unwind the position in case of a whipsaw breakout: Retracement of
50 percent or more of the distance from the final low to the breakout
point. Breakout accompanied by low volume.
• Violation of any up trendline joining the final low and any higher minor
lows.

Triangles
Triangles are the most c o m m o n price pattern, but they are also o n e of the
least reliable. They develop as both consolidation and reversal formations, and
consist of an ever-narrowing trading range b o u n d e d by two converging trendlines. In o r d e r for a line to represent o n e of the boundaries, it needs to be
touched on at least two occasions. This means that a triangle consists of at
least four turning points, two for each line. In reality, though, such patterns
are usually m o r e reliable when o n e of t h e m has b e e n touched three or m o r e
times. I n d e e d , the m o r e contact or near-contact points, the better. In
Chapter 4 it was p o i n t e d out that trendlines are n o t h i n g m o r e than dynamic
levels of support a n d resistance. Thus, the m o r e times a triangle boundary
has b e e n touched or approached, other things being equal, the greater the
significance of that boundary as a support or resistance area, and therefore
the stronger the breakout signal when it comes. T h e r e are two types of triangle, the symmetrical a n d the right-angled, so let's consider them in turn.

Symmetrical Triangles
A symmetrical triangle is composed of a series of two or m o r e rallies and reactions in which each peak is lower than its predecessor and the bottom of each
reaction is higher than its predecessor (see Fig. 9-1). A triangle is the opposite of a b r o a d e n i n g formation (see Chapter 10), since the trendlines joining peaks a n d troughs converge; in b r o a d e n i n g formations, they diverge.
These patterns are also known as coils because the fluctuation in price a n d
volume diminishes as the pattern is completed. Finally, b o t h price a n d (usually) volume react sharply, as if a coil spring h a d b e e n w o u n d tighter a n d
tighter a n d t h e n s n a p p e d free as prices broke o u t of the triangle. Generally
speaking, triangles seem to work best when the breakout occurs somewhere
between one-half a n d two-thirds of the distance between the widest peak

149

Part II: Traditional Patterns

150

Figure 9-1

Symmetrical triangle.

a n d rally a n d the apex (as in Fig. 9-2). T h e volume rules used for o t h e r patterns are also appropriate for triangles. T h a t means that activity should gradually contract as the p a t t e r n is being formed. During upside breakouts, it is
i m p o r t a n t for volume to e x p a n d (see Fig. 9-3). In downside breakouts, it
doesn't m u c h matter w h e t h e r volume contracts or expands, t h o u g h an
expansion of activity indicates selling pressure a n d adds a few points to the
bearish case. In Fig. 9-1, for instance, volume expands on the downside

Triangles

151

Figure 9-3 Symmetrical triangle at a bottom.

breakout, having contracted as the p a t t e r n was forming. T h e point h e r e is
that the battle between buyers a n d sellers is b e c o m i n g m o r e finely balanced
as the pattern develops. T h e n , on the breakout, volume expands, thereby
signaling from b o t h a price a n d a volume p o i n t of view that control has
slipped decisively to o n e side or the other.
Most patterns give us a clue to the direction of the ultimate breakout in
the event that they are completed. T h e symmetrical triangle does not; it
keeps us guessing. Since it should be assumed that the prevailing trend is
intact until proven otherwise, it is a good idea to take it for granted that the
triangle will eventually break in the direction of the prevailing trend. An
example of a continuation triangle is shown in Fig. 9-4. Clues to the contrary—i.e., in favor of a reversal—would a p p e a r if the price was overext e n d e d in o n e direction or the other. Alternatively, it may be possible to
observe that many stocks have already b e g u n to break in a new direction.
A reversal would be m o r e likely if the pattern formed after the prevailing
trend h a d b e e n in place for a long time than if it h a d j u s t begun. However,
when taken in isolation with no o t h e r supporting evidence, the prevailing
t r e n d assumption should be applied.

The Underlying Psychology

Figure 9-2

Symmetrical triangle indicating ideal breakout range.

A triangle pattern is no m o r e than a gradually tightening battle between buyers a n d sellers. At t h e outset, the large price swings at the left-hand p a r t of
the formation indicate relative instability, showing that both sides are out
of control. Prices initially rally up at the b e g i n n i n g of the triangle until they

^n

Part II: Traditional Patterns

Triangles

153

Figure 9-5 Symmetrical triangle with traditional measuring objective.
Figure 9-4 Consolidation symmetrical triangle.

reach a point at which buyers become less enthusiastic and sellers quite motivated. T h e n the price slips back quite a ways, a n d those who missed the boat
earlier are m o r e inclined to buy at these lower prices. Selling is also less
intense. T h e price t h e n rises for a second time, b u t n o t to as great an extent
before supply once again overwhelms the buyers. This could be because
potential sellers, having missed o u t on the opportunity to sell at higher
prices on the first rally, t e m p e r their greed a n d are p r e p a r e d to settle for
less profit. In any event, prices decline a n d buyers come in again, b u t each
rally a n d reaction attracts fewer a n d fewer participants. T h e initial excitem e n t dies down, a n d m a r k e t participants await a resolution of the fine balance between them. As the pattern develops a n d the battle lines come closer
a n d closer, neither side is able to exert as m u c h upside or downside pressure. As a general rule, whenever price activity goes quiet, as it does close
to the apex of the triangle, the slightest tip in the balance between supply
a n d d e m a n d will result in a commensurately larger price move. Usually the
m o r e contact points, the m o r e persistent the price move following the breakout. A n o t h e r clue to the strength of the b r e a k o u t is the strength of the contrast between shrinking volume as the p a t t e r n is formed a n d e x p a n d i n g
activity on the breakout. T h e greater the contrast, the m o r e decisive the victory a n d the stronger the signal.

Measuring Objectives
Traditionally, measuring objectives for triangles at market tops are obtained
by drawing a line at the base of the triangle parallel to the u p p e r trendline.

This line (CD in Fig. 9-5) represents the objective that prices may be
expected to reach or exceed.
T h e reverse p r o c e d u r e at market bottoms is shown in Fig. 9-6. T h e same
technique is used to project prices when triangles are of the consolidation
variety.
In my own e x p e r i e n c e , I have n o t f o u n d this m e t h o d to be particularly
useful. This is because t h e actual price move is usually far m o r e t h a n t h e
p r i c e objective. I prefer, instead, to treat the triangle like any other pattern,

Figure 9-6 Symmetrical triangle with traditional measuring objective.

154

Part II: Traditional Patterns

Triangles

Chart 9-1 Yahoo, 2001-2002, daily.

Figure 9-7

Symmetrical triangle with alternative measuring objective.

calculating its maximum depth and then projecting this distance at the breakout.
Examples of this alternative m e t h o d for b o t h a top a n d a b o t t o m are shown
in Figs. 9-7 a n d 9-8. By a d o p t i n g this a p p r o a c h , the principle of proportionality is preserved. Charts 9-1 a n d 9-2, featuring Yahoo, show how the
two m e t h o d s might work in the m a r k e t p l a c e . T h e original t e c h n i q u e takes
a line parallel to the u p p e r down t r e n d l i n e of the p a t t e r n a n d anchors it

Figure 9-8 Symmetrical triangle with alternative measuring objective.

Chart 9-2 Yahoo, 2001-2002, daily.

155

156

Part II: Traditional Patterns

at the b o t t o m of the initial decline. In this instance, the line b e c o m e s o n e
of support, as it is able to reverse the July a n d August declines. In the seco n d example, the distance between the peak a n d the initial low is measured
a n d that distance is projected down from the b r e a k o u t point. In this
instance, t h e objective falls in line with t h e actual b o t t o m . Both m e t h o d s
result in useful b e n c h m a r k s , b u t the p r o p o r t i o n a t e or second t e c h n i q u e is
the o n e that calls the final low. I have to a d d that even t h o u g h this example was selected at r a n d o m , n o t all price-objective moves work o u t this accurately.
C h a r t 9-3 shows a n o t h e r e x a m p l e of a symmetrical triangle at a market t o p , for Intel. T h e original m e a s u r i n g objective is r e p r e s e n t e d by t h e
d e c l i n i n g d a s h e d line. This time t h e price immediately breaks below this
line, b u t t h e two s u b s e q u e n t rallies find resistance t h e r e . O n t h e o t h e r
h a n d , t h e u l t i m a t e low is m a d e at a r o u n d twice t h e objective called for
b y t h e p r o p o r t i o n a t e m e t h o d . This also r e p r e s e n t e d t h e a p p r o x i m a t e
level of t h e lows set in t h e previous J u n e / J u l y p e r i o d . It is also w o r t h noting t h a t while t h e initial price objective was easily e x c e e d e d , this same
level t u r n e d o u t to be resistance for t h e N o v e m b e r a n d D e c e m b e r rallies.
An e x a m p l e of a triangle b o t t o m is shown in C h a r t 9-4, for Alcoa. N o t e
how a line parallel to t h e lower p a r t of t h e rectangle a n d a n c h o r e d at the
secondary p e a k at A provided resistance t h r o u g h o u t t h e e n s u i n g advance.

Chart 9-3 Intel, daily.

Triangles

157

Chart 9-4 Alcoa, 1986-1989. weekly.

Weaving Symmetrical Triangles into Head-and-Shoulders Tops
Triangles are o n e of the least accurate of all the price patterns. O n e of the
reasons is that a formation often starts out looking like a triangle, b u t ends
up as s o m e t h i n g completely different. An example is shown in Fig. 9-9, in
which a b r e a k o u t develops above the dashed trendline m a r k i n g the top of

Figure 9-9 Failed symmetrical triangle.

158

Part II: Traditional Patterns

159

Triangles

a triangle. Later on, though, the price action unfolds into an upward sloping
head-and-shoulders top. T h e upside b r e a k o u t in the triangle therefore gave
a false indication of the direction of the next significant price move. This
is just o n e example of how a triangle can offer misleading signals. O n e way
to r e d u c e the probability of a whipsaw triangle b r e a k o u t is to consider only
formations where the price has h a d contact or close contact with the breakout line on m o r e than two occasions. In fact, the m o r e times the better, since
this would reinforce the line as a resistance or support area, thereby making its penetration m o r e likely to succeed. Bearing in m i n d some of the rules
for d e t e r m i n i n g the significance of a trendline established in Chapter 4,
a n o t h e r filtering a p p r o a c h would be to exclude triangles where the angle
of ascent or descent is particularly steep.
Figure 9-11

Bullish right-angled triangle with a retracement.

Right-Angled Triangles
Right-angled triangles are really a special form of the symmetrical type, in
that o n e of the two b o u n d a r i e s is formed at an angle of 90 degrees to the
vertical axis, i.e., is horizontal (Fig. 9-10). T h e symmetrical triangle does not
give an indication of the direction in which it is ultimately likely to break.
T h e right-angled triangle does, with its implied slanting level of s u p p o r t or
resistance.
Like most patterns, triangles often experience retracement moves following breakouts. An example is shown in Fig. 9-11 for a right-angled ascending

triangle. If an opportunity is missed on the breakout, the pullback or retracem e n t move offers a second chance, usually u n d e r far m o r e quiet conditions.
If it is possible to construct a retracement trendline, as in Fig. 9-11, the upside
penetration of the line is a good entry point, since it indicates a reassertion
of the original breakout. This is especially true if the retracement move is
accompanied by shrinking volume and the breakout above the retracement
trendline by slightly or significantly expanding activity, as in Fig. 9-11.

Spinning Right-Angled Triangles into Rectangles
O n e difficulty in interpreting these formations is that many rectangles begin
as right-angled triangles. Consequently, a great deal of caution should be used
when evaluating these elusive patterns. An example is shown in Fig. 9-12,
where a potential downward-sloping right-angled triangle develops into a
rectangle.

Marketplace Examples
An example of a right-angled triangle at a b o t t o m is shown in Chart 9-5, for
Intel. If you believe that the J u n e 1986 decline qualifies as a shoulder, it
could be argued that this p a t t e r n is a reverse h e a d a n d shoulders, with the
horizontal trendline representing the neckline. You could also say that the
"horizontal" line is n o t quite horizontal, a n d so the triangle would be m o r e
accurately called a symmetrical o n e . Frankly, I think we would be pushing
Figure 9-10 Bearish right-angled triangle.

Part II: Traditional Patterns

Triangles

161

160
Chart 9-6 Analog Devices, 1991-1994, weekly.

Figure 9-12

Failed bearish right-angled triangle.

the envelope on this o n e . It really doesn't matter what the pattern.is called;
Se fact is that it was a clear-cut battle between buyers a n d sellers h a t was
e l o v e d n favor of the buyers, a n d that's ultimately what c o u n o . A b o note
h s u b i n t i a , increase tn volume as the price breaks to the upside a n t h
fact that the lower trendline has b e e n t o u c h e d six times. If the standard

Chart 9-5 Intel. 1986-1987, daily.

triangles is a d o u b l e contact only, then this is definitely an impressive line
of support. O n c e again I applied the two price-objective techniques. T h e
parallel-line m e t h o d caused the price to find temporary resistance in the
m o n t h of January. W h e n it finally broke above the line, a very sharp rally
followed. T h e p r o p o r t i o n a t e a p p r o a c h was well exceeded, b u t it is interesting to n o t e that the objective level ultimately became s u p p o r t in the form
of a neckline for a failed head-and-shoulders top formation.
Chart 9-6, for Analog Devices, shows a n o t h e r ascending triangle. This time
it is formed above the low p o i n t for the move a n d is therefore a consolidation pattern. Note how volume shrinks as the p a t t e r n develops. It also
expands a little on the breakout, b u t n o t h i n g to get excited about.
N o r t h r o p G r u m m a n , in Chart 9-7, provides us with a fairly large descending triangle top. T h e original (dashed) descending line h a d to be redrawn
because of the two 1987 rallies. It is worth noting that when the price fell
t h r o u g h the e x t e n d e d dashed line for a second time, it actually experienced
a pretty wide gap j u s t prior to completing the pattern.

Right-Angled Failures
To t h e flexible t r a d e r or investor, a p a t t e r n failure offers great o p p o r t u nity. This is probably m o r e t r u e of right-angled triangles t h a n of any o t h e r
formation. We saw earlier how a right-angled triangle can transform itself

Part II: Traditional Patterns

162

Triangles
163

Chart 9-7 Northrop Grumman, 1984-1990, weekly.

Figure 9-14 Failed bearish right-angled triangle.

i n t o a r e c t a n g l e . However, t h e moves following right-angled b r e a k d o w n s
a r e usually a little m o r e d r a m a t i c . T h e r e are two types of failure. T h e first
develops w h e n t h e price p e n e t r a t e s t h e h o r i z o n t a l t r e n d l i n e a n d t h e n
moves b a c k t h r o u g h it. E x a m p l e s for a b o t t o m a n d a t o p are shown in
Figs. 9-13 a n d 9-14. In Fig. 9-13, t h e false u p s i d e b r e a k o u t develops close

Figure 9 - 1 3

Failed bullish right-angled triangle.

to the apex, so it is possible to set a fairly close stop under the rising trendline. It would even make sense to exit the position once the price slips
below the horizontal line again. However, the odds of a failure increase
greatly when the rising line in a bullish pattern is penetrated. This is
because this line is often a strong area of support. False breakouts often
result in a strong move in the opposite direction to that expected, so it is
even possible to go short on a break of the rising trendline. The buy stop
would then be placed above the line at Xand moved progressively higher
until a better place is found at a lower level. Figure 9-14 shows another
false break, this time from what looked like a descending pattern. The distance between the breakout point and the descending line is quite substantial, which means that any short positions, triggered from the false
downside break, would be more timely covered on a rally above the horizontal trendline at X.
An example of an upside failure is shown in Chart 9-7, for Northrop
Grumman. This 1988 failure is a classic case because the breakout developed
against the direction of the main trend. The scene had been set earlier with
a breakdown from an almost three-year right-angled triangle top. Volume
for the failed pattern contracted as the pattern was developing, a normal
phenomenon. However, it expanded in a very deceptive way on the breakout, thereby offering a false sense of security. Just after the price fell below
the rising trendline, volume expanded rapidly. This was a clear-cut signal,
since this bear market characteristic left no doubt that the sellers were now
in control.

Part II: Traditional Patterns

165

Triangles

164
accompanied by expanding volume, this really emphasizes the fact that
market participants have made a mistake. The high volume indicates that
a large number of people are trying to unwind their positions, while others are
trying to climb on board the new trend before the train has left the station.

The Underlying Psychology of Failed Right-Angled Triangles

Figure 9-15 Failed bullish right-angled triangle.

The second way in which an ascending triangle can fail is when the ris
ing or falling trendline is penetrated prior to a breakout through the horizontal line. In Fig. 9-15, buyers are gaining confidence as each decline is
terminated at a higher level. Technically oriented people see this price
action and buy in anticipation of a successful pattern completion. Such
hence, there is the potential for a sharper than average decline. An example
for a failed ascending triangle is shown in Fig. 9-16. If such a failure is

Figure 9 - 1 6

Failed bearish right-angled triangk

In the case of a failed ascending triangle, confidence rises as the price rallies to the horizontal trendline, but each decline is smaller than its predecessor. A move above the overhead resistance indicated by the horizontal
line is therefore expected, and positions are taken accordingly. At the same
time, sellers feel quite comfortable liquidating every time the price rises to
the line. Since this happens on a number of occasions, there is no pressure
to accept a lower price—that is, until the ascending line is violated. Then
everybody wants out. On the one hand, the original sellers realize that it is
no longer possible to attain the higher prices associated with the horizontal trendline. On the other hand, those who bought in anticipation of the
upside break lose heart. Such a rationale explains why such failures are so
often followed by substantial price moves.
Chart 9-8, for Linear Technology, offers a good example. In this
instance, the price looked as if it were in the process of forming a descending right-angled triangle. The June-August period even took on the air of

Chart 9-8

Linear Technology, 1981-1983, daily.

166

Part II: Traditional Patterns

Triangles
167

Chart 9-9 WorldCom, 1993, daily.

a symmetrical triangle. However, the descending triangle was never completed, and prices broke to the upside. Chart 9-9, for WorldCom, on the
other hand, shows two examples of a failed ascending and one of a failed
descending triangle. Notice how the volume shrinks as all three triangles
are formed. It also expands as the two on the right experience false breakouts. Also featured is a large rectangle top formed in early 1994.
Finally, Analog Devices, in Chart 9-10, experienced a failed ascending pattern between late 2001 and early 2002.

Summary
Symmetrical Triangles Quick Review
• Price characteristics: A narrowing trading range confined between two converging trendlines moving in different directions.
• Volume considerations: Volume contracts as the pattern forms. Upside
breakouts should be accompanied by expanding volume.

Chart 9-10 Analog Devices, 1998-2003, weekly.

• Measuring implications: The maximum depth of the pattern is projected
in the direction of the breakout. Alternatively, for bullish patterns, draw
a line parallel to the pattern's lower trendline and anchor it at the initial
rally. The extended line becomes the price objective. For bearish patterns,
draw a line parallel to the pattern's upper trendline and anchor it at the
initial decline. The extended line becomes the price objective.
• Strongest breakouts: Come from a point one-half to two-thirds of the distance
between the start of the pattern and the apex.
• Signs of false breakouts: Weak volume accompanying an upside breakout.

Right-Angled Triangles Quick Review
• Price characteristics: A narrowing trading range confined between two converging trendlines, one of which is at a right angle to the vertical axis. The
price is expected to break through the right-angled line in the direction
of the sloping trendline.
• Volume considerations: Volume contracts as the pattern forms. Upside
breakouts should be accompanied by expanding volume.
• Measuring implications: The maximum depth of the pattern is projected
in the direction of the breakout. Alternatively, for bullish patterns, draw
a line parallel to the pattern's lower trendline and anchor it at the initial

168

Part II: Traditional Patterns

rally. T h e extended line becomes the price objective. For bearish patterns,
draw a line parallel to the pattern's u p p e r trendline a n d a n c h o r it at the
initial decline. T h e e x t e n d e d line becomes the price objective.

10

Strongest breakouts: Come from a point one-half to two-thirds of the distance
between the start of the pattern a n d the apex.
Signs of false breakouts: Contracting or weak volume accompanying an
upside breakout.
Pattern failures: Occasionally come with a violation of the sloping trendline.

Broadening
Formations

B r o a d e n i n g formations signify a very unstable technical situation a n d typically develop after a t r e n d has b e e n underway for some time. It's almost as
if the battle between buyers a n d sellers is out of control, because these patterns exhibit wider a n d wider price fluctuations. They contrast well with triangles, w h e r e the trading r a n g e gradually narrows to a very balanced
position p r i o r to the breakout.
B r o a d e n i n g formations occur when a series of three or m o r e price fluctuations widens o u t in size so that peaks a n d troughs can be c o n n e c t e d with
two diverging trendlines. Triangles come in two varieties, a n d so do broade n i n g formations. These variations are called orthodox a n d right-angled. T h e
right-angled type is sometimes referred to as a broadening formation with a
flat top (the accumulation version) or aflat bottom (the distribution variety).
Let's start with the classic or o r t h o d o x pattern.

Orthodox Broadening Formations
An example of an orthodox broadening top is shown in Fig. 10-1. It comprises
three rallies, with each succeeding peak being higher than its predecessor.
T h e three peaks are separated by two bottoms, with the second bottom being
lower than the first. O r t h o d o x b r o a d e n i n g formations are associated with
market peaks r a t h e r than troughs, although some of the textbooks tell us
that they can develop at the e n d of bear trends. O r t h o d o x b r o a d e n i n g formations are sometimes called reverse triangles, because that is essentially
what they are. On those rare occasions when they appear at market bottoms,

169

Part II: Traditional Patterns

170

Figure 10-1

Orthodox broadening formation.

they are sometimes called megaphone bottoms because their appearance is
similar to that of a megaphone.
Some patterns are ideal for trading because they offer clear demarcation
lines for entry and convenient low-risk stop points in case things go wrong.
A right-angled triangle that breaks out close to the apex is a good example.
Unfortunately, this is not the case with the orthodox broadening formation,
since these patterns are extremely difficult to detect until some time after
the final top has been formed. Also, there is no clearly definable level of
support, the violation of which serves as a convenient benchmark. The violent and emotional nature of both price and volume swings further compounds the confusion and increases the complexity of defining these
situations. Obviously, a breakout is difficult to pinpoint under such conditions, but if the formation is reasonably symmetrical, a decisive move below
the descending trendline joining the two bottoms, or even a decisive move
below the second bottom, usually serves as a timely warning that an even
greater decline is in store.
Measuring implications are similarly difficult to determine, but normally
the volatile character of a broadening top formation implies the completion of a substantial amount of distribution. Consequently, price declines
of considerable proportion usually follow the successful completion of such
patterns.
The problem is that the breakdown point from such formations develops well after the final turning point of the previous move, so if you are
long, you will probably say to yourself, "I can't sell here; it's declined far

Broadening Formations

171

too m u c h . I'll wait for a b o u n c e . " Unfortunately, this is n o t a g o o d strategy, since t h e c o m p l e t i o n p o i n t of such formations is often j u s t the tip of
the iceberg. T h e reason lies in the fact t h a t t h e o r t h o d o x p a t t e r n usually
develops after a very lengthy bull move, w h e n prices have increased substantially from t h e previous b e a r m a r k e t low. Consequently, t h e r e is no
s h o r t a g e of sellers who are able to cash in on their profits. T h e volatility
d u e to t h e diverging n a t u r e of t h e f o r m a t i o n also adds to the feeling of
instability. W h e n prices are unstable, h o l d e r s who would otherwise be reasonably calm a n d c o n t r o l l e d have a g r e a t e r t e n d e n c y to p a n i c . It is n o t
unlike the idea that p e o p l e living in a violent-crime-ridden n e i g h b o r h o o d
will have a g r e a t e r t e n d e n c y to resort to violent solutions t h a n s o m e o n e
living in a quiet s u b u r b .
Having spent a page or two describing the o r t h o d o x b r o a d e n i n g formation, it must be p o i n t e d out that two classic texts, Edwards a n d Magee
a n d H. M. Gartley, b o t h agree that apart from the 1929 top, when such patterns were r a m p a n t , t h e r e is n o t m u c h evidence of this formation's existence. I n d e e d , after looking t h r o u g h 20 years of data on the Nasdaq 100
a n d most of the S&P 500, I failed to c o m e up with what I would regard as
either useful examples or situations in which where the r i s k / r e w a r d was
favorable. It would appear, therefore, that this formation is pretty well
extinct. Certainly the lack of good examples makes it impracticable. W h a t
is n o t impracticable is the right-angled b r o a d e n i n g variety, which is covered
next.

Right-Angled Broadening
Formations
T h e easiest types of b r o a d e n i n g formations to detect are those with a "flattened" b o t t o m or top, as shown in Figs. 10-2 a n d 10-3.
These patterns are sometimes referred to as right-angled broadening formations. Since the whole c o n c e p t of widening price swings suggests highly
emotional activity, volume patterns are difficult to characterize, although at
market tops, activity is usually heavy d u r i n g the rally phases. T h e patterns
at both bottoms a n d tops are similar to head-and-shoulders patterns except
that the "head" in the b r o a d e n i n g formation is always the last part of the
pattern to be formed. A bear signal comes with a decisive downside breakout. Volume can be heavy or light, but if activity expands at this point, this
is an additional bearish factor.
Since a b r o a d e n i n g formation with a flattened top is an accumulation pattern, volume expansion on the b r e a k o u t is an i m p o r t a n t r e q u i r e m e n t , as

172

Part II: Traditional Patterns

Figure 10-2 Right-angled broadening top formation.

shown in Fig. 10-4. In my experience, these formations pack a p u n c h far
greater than their size would suggest. They are really head-and-shoulders
Tops (bottoms) where the situation is so bearish (bullish) that the price does
n o t have any time to trace o u t a right shoulder.

Broadening Formations

Figure 10-4

173

Right-angled broadening bottom formation.

The Underlying Psychology
B r o a d e n i n g formations with a flat top usually a p p e a r as the culmination of
a persistent downtrend. At the start of the formation, bearish sentiment is
typically quite excessive. T h e price initially sells off, then subsequently rallies to the horizontal level of resistance. W h e n the bears see that the resistance has held, they are e m b o l d e n e d to p u t o u t m o r e shorts, a n d a decline
sets in. T h e price t h e n rallies again a n d finds resistance in the area at the
top of the pattern, b u t the advance from the second low is n o t sufficient to
cause the shorts to cover. I n d e e d , when the price fails to rally t h r o u g h the
resistance, the bears are even m o r e e m b o l d e n e d to p u t out even m o r e
shorts.
As the price falls below the previous low, latecomers are attracted to the
short side. After all, they can see that in the past few weeks a n d m o n t h s ,
there was easy money to be m a d e from shorting, especially as the news backg r o u n d is so negative that prices are "sure" to go down a lot m o r e . This is
e n o u g h to panic weak holders, who experience the give-up phase a n d liquidate to anyone "kind" e n o u g h to take the security off their h a n d s . If bearish sentiment was widespread coming into the pattern, it is now universal,
as virtually everyone is a believer in the b e a r trend. Those who are left are
either strong holders with a positive long-term belief in the eventual outcome or short sellers. T h e b o t t o m line is that there is little or no selling pressure b u t lots of potential buying pressure from the mercurial shorts.

Figure 10-3 Right-angled broadening bottom formation.

174

Part II: Traditional Patterns

T h e gasoline is on the fire. All that is n e e d e d is a match. T h a t could be
provided by u n e x p e c t e d good news, a change in the direction of the overall market, or some o t h e r reason. Perhaps the news is particularly b a d a n d
the price does n o t decline. This type of action will spook the shorts, who
argue that if b a d news won't cause a further decline, n o t h i n g will. T h e reason is immaterial. W h a t matters is that the price starts to rise, a n d rise
quickly, fueled by some short covering and bargain hunting. It then explodes
t h r o u g h the resistance represented by the horizontal trendline, a n d the pattern is completed. Buyers are reluctant to e n t e r at this level because the
price has already risen substantially from the low. However, those with a longterm conviction are n o t deterred. Also, traders who are still struggling to
cover their short positions c o n t i n u e to buy. Remember, those who are short
are n o t thinking a b o u t the downside risk; they are worried a b o u t the unlimited upside catastrophe that will occur if prices continue to rally. Virtually
everyone who shorts has a relatively brief time horizon, so the rally coming
off a b r o a d e n i n g low has the feel of a h u g e u n e x p e c t e d bull market. T h e r e
is only o n e solution: Cover before prices go any higher!
T h e reason why the rally extends probably arises from the utterly and completely bearish environment at the low. Anyone who was previously long and
wanted to get out has already d o n e so. As a result, the security in question
is held either by long-term believers or by shorts. O n e isn't going to sell, a n d
the o t h e r literally has to buy.
T h e psychology at tops is exactly the opposite. H e r e , the formation starts
after a persistent advance, which encourages widespread optimism as the
price makes successive new highs d u r i n g the development of the b r o a d e n ing part of the pattern. Naturally, these progressively h i g h e r rallies discourage the shorts. They probably place stops j u s t above resistance, which
they j u d g e to be at the previous high. After a couple of attempts at this, they
give u p . This m e a n s that at the final top, t h e r e is a very small short position.
Large short positions act as s u p p o r t for declines; small ones do not.
Since the pattern develops after a fairly lengthy advance, new buyers are
attracted every time the price rises to a new high because there is no feeling
of downside risk. Unfortunately, these new players are of m u c h p o o r e r quality than, for example, the type of person who held on through the final throes
of a b r o a d e n i n g bottom. These participants are attracted by the good news
and attractive prospects being painted by the media, brokers, and others. T h e
progressively stronger rallies associated with this pattern also result in careless decisions, since rising prices bail out the greedy and inexperienced buyers who b e c o m e accustomed to a one-way street. These are i n d e e d LIFO
buyers—last in, first out. Consequently, when the price reaches its final peak,
holdings are concentrated a m o n g u n i n f o r m e d a n d weak holders. To make
matters worse, there is very little cushion in the form of short positions. Prices

Broadening Formations

175

then begin to decline rather rapidly. Since a quick sell-off of this n a t u r e is
unexpected, few are able to get out at the beginning. As the price falls
t h r o u g h the support at the horizontal trendline, it attracts m o r e selling, but
any short covering that would normally have taken place at the support trendline is n o t available as a result of previous short squeezes. Therefore, prices
continue to fall. T h e ensuing decline is fairly persistent because it takes a long
time to liquidate the weak holders who b o u g h t the security in the heady days
when the top was forming.

Measuring Implications
T h e m e a s u r i n g objectives for these patterns are taken from the m a x i m u m
distance between the peak (or the bottom, in the case of accumulation) of
the formation a n d the horizontal line. T h e distance is t h e n projected from
the b r e a k o u t p o i n t in the direction of the breakout.
Right-angled b r o a d e n i n g formations can experience r e t r a c e m e n t or pullback moves j u s t like o t h e r patterns. Because they are fairly violent a n d
unstable, these r e t r a c e m e n t s can b e extremely sharp u n d u n n e r v i n g .
Fortunately, they are normally short-lived.
Chart 10-1 shows a right-angled b r o a d e n i n g top for Intel. In this case, the
downside objective was r e a c h e d d u r i n g the b r e a k o u t move. It is unusual to

Chart 10-1 Intel, daily.

176

Part II: Traditional Patterns

Chart 10-2 Ericsson, daily.

see such a quick reversal after this pattern has been completed, since prices
normally drop or rally far more than the indicated objective. In this respect,
the top in Ericsson in Chart 10-2 is a more typical example. Note also the
retracement move that developed right after the breakdown. This would

Chart 10-3 Aetna, daily.

Broadening Formations

177

Chart 10-4 U.S. Dollar Index, 1975-1983, monthly.

have b e e n quite u n n e r v i n g to anyone who h a d g o n e short on the breakout,
since there was no logical low-risk resistance level above which to place a stop.
A n o t h e r example of a right-angled b r o a d e n i n g b o t t o m appears in Chart
10-3, for Aetna. This time the b r o a d e n i n g or diverging part of the formation is m o r e controlled a n d n o t that d e e p . T h e b r e a k o u t a n d subsequent
move are also m o r e constrained.
Generally speaking, the b r o a d e n i n g variations pack a great deal m o r e
p u n c h than an equivalent-sized head-and-shoulders pattern. Just take a look
at Chart 10-4, featuring a monthly close of the Dollar Index. In this instance,
the price rallied very quickly to five times the objective.

W h e r e t o D r a w t h e Lines
A lot of the time, it is necessary to use a little poetic license when constructing these patterns. Let's consider the accumulation variety in this explanation.
What we are really trying to construct is a reverse right-angled triangle, as shown
in Fig. 10-5. Note that because of the jagged nature of the price action, it is
not possible to j o i n all the rallies and reactions exactly. Often we have to compromise on an approximate area of resistance for the horizontal part of the
formation, for example. If you think of the underlying psychology of the pattern as described earlier, everything still fits. T h e unfortunate thing is that there
are rough approximation points rather than black-and-white signals that give
us greater confidence in the completion of these patterns.

Part II: Traditional Patterns

178

Broadening Formations

179

Chart 10-6 Philadelphia Gold and Silver Share Index, daily.

Figure 10-5

Right-angled broadening bottom formation featuring shape.

Charts 10-5 a n d 10-6 b o t h feature bottoms for the Philadelphia Gold a n d
Silver Share I n d e x . In C h a r t 10-5, each of t h e lines is temporarily b r e a c h e d
once b u t t h e r e is no d o u b t that the formation reflects the b r o a d e n i n g concept with a flat top. T h e initial rally (A) in C h a r t 10-6, on the o t h e r h a n d ,
presents us with a bit m o r e of a challenge; however, if t h e a n g l e d line is

Chart 10-5 Philadelphia Gold and Silver Share Index, daily.

b r o u g h t back to the apex, it is evident that it forms p a r t of the p a t t e r n .
Note that t h e a n g l e d or diverging line was a p p r o a c h e d or t o u c h e d on
n u m e r o u s occasions a n d therefore r e p r e s e n t e d substantial support. It also
gave t h e p a t t e r n m o r e credibility, so w h e n the b r e a k o u t t h r o u g h the horizontal t r e n d l i n e did take place, it was followed by a very worthwhile move.
T h e multiyear bottom in the Dollar Index (Chart 10-4) also presented a
small challenge in construction, since b o t h lines were e x c e e d e d o n c e .
However, there can be no denying the exceptionally strong rally that followed.

Right-Angled Broadening Formations as Consolidation Patterns
These b r o a d e n i n g formations can also develop as consolidation patterns,
as shown in Fig. 10-6 for an u p t r e n d . Chart 10-7, for the c o p p e r price, shows
a consolidation right-angled top. T h e achievement of triple the downside
objective indicates that these formations should be respected as m u c h as the
tops a n d bottoms. N o t e the b r o a d e n i n g b o t t o m that formed at the e n d of
the decline. T h e angle of descent was quite sharp. If you look at the examples in this chapter, you will see that there is a very r o u g h correlation
between t h e angle of t h e b r o a d e n i n g p a r t a n d the speed of the ensuing
move. Thus, the sharper the angle (the greater the volatility), the m o r e likely
it is that a precipitous decline or explosive rally will follow.

Part II: Traditional Patterns

180

Figure 10-6

Broadening Formations

181

Consolidation right-angled broadening bottom formation.
Figure 10-7

Failed right-angled broadening top formation.

Failed Broadening Formations
B r o a d e n i n g formations occasionally fail to work. Possibilities are shown in
Fig. 10-7 a n d 10-8. Unfortunately, t h e r e does n o t a p p e a r to be a reliable
or timely p o i n t b e y o n d which it is safe to say that t h e p a t t e r n has failed to
o p e r a t e unless a m i n o r p e a k or t r o u g h develops d u r i n g t h e b r e a k o u t

decline or rally. The best defense in such cases is to adopt the 50 percent
rule, in which the halfway mark of the final reaction (Fig. 10-7) or rally
(Fig. 10-8)—i.e., the dashed lines—is used as the give-up point. Some patterns may still work out after the 50 percent plus retracement has taken
place, but generally speaking, a breakout that is followed by such a strong

Chart 10-7 Copper, 2000-2002, weekly.

Figure 10-8 Failed right-angled broadening bottom formation.

182

Part II: Traditional Patterns

Broadening Formations

Chart 10-8 Avery Dennison, 1980-2003, monthly.

Figure 10-9 Broadening wedge.
c o u n t e r m o v e is n o t really in t h e spirit of t h e formation a n d s h o u l d be
treated accordingly.
Chart 10-8, for Avery Dennison, shows a failed o r t h o d o x b r o a d e n i n g formation. Up to the e n d of 1990, it looked like a large top. However, by the
beginning of 1991, the stock h a d experienced a sharp rally. Since the e n d of
1990 also m a r k e d a low for the overall market, it is evident that the stock rallied in sympathy with the rest of the list. Generally speaking, it is very difficult
for a stock to complete a major top such as this o n e if the overall market is
turning at the same time. Of course, it usually is n o t possible to gauge a market t u r n until well after the fact. However, if you can, be wary of patterns that
forecast prices continuing to move in the direction of the previous prevailing
trend. In this case, it was not until the price experienced a series of rising peaks
a n d troughs as it broke above the dashed trendline that the balance of evidence, as far as this pattern was concerned, moved to the bullish side.

Broadening Wedges
Sometimes the horizontal line in the "right-angled" b r o a d e n i n g formation
is set at a slight angle, as shown in Figs. 10-9 a n d 10-10. Note that b o t h lines
are actually moving in the same direction. T h e principles of interpretation
for these patterns are the same as for the right-angled variety. T h e breakout

Figure 10-10 Broadening wedge.

183

184

Chart 10-9 Advanced Micro, 1987-1988, daily.

Chart 10-10 Amerisour Bergn, daily.

Part II: Traditional Patterns

Broadening Formations

185

Chart 10-11 ADC, 1980-1990, weekly.

implications are similar in that broadening wedges continue to pack a substantial punch when they are completed.
An example of a top is shown in Chart 10-9, for Advanced Micro. Note the
sharp drop that follows the breakdown. Another example is shown in Chart
10-10, for Amerisour Bergn. This time the decline stops at approximately
three times the price-objective multiple. Note that there is also a nice double
bottom, the completion of which tells us that the decline is over. Finally, a
rounding top (see Chapter 11) forms at the very right-hand side of the chart.
The last chart in this chapter, Chart 10-11, shows a broadening wedge for
ADC. In this instance, the full potential of a very bearish-looking pattern
was not realized. An indication of the abbreviated decline was given when
the price held above its secular bull market trendline and broke out from
an almost one-year rectangle.

Summary
Orthodox Broadening Tops Quick Review
• Price characteristics: A trading range following a substantial bull market that
is bounded by two diverging trendlines. Each line should be touched on
at least two occasions.
• Frequency: Very rare pattern.

186

Part II: Traditional Patterns

Broadening Formations

187

Right-Angled Broadening Tops Quick Review

Broadening W e d g e s Quick Review

• Price characteristics: A trading r a n g e following a rally, b o u n d e d by two
diverging trendlines. T h e lower line is at or n e a r a right angle to the vertical axis. Each line should be t o u c h e d on at least two occasions. Poetic
license is often r e q u i r e d in constructing these lines, since they are n o t
always t o u c h e d exactly, as is t h e case with most o t h e r patterns.

• Price characteristics: Similar to those of right-angled b r o a d e n i n g formations
except that the right-angled line experiences a slight slope a n d is n o t
drawn at 90 degrees.

• Volume considerations: Volume is usually h i g h e r on t h e first two peaks.
E x p a n d i n g activity on the downside b r e a k o u t is particularly bearish.
• Measuring implications: T h e m a x i m u m d e p t h of the p a t t e r n is projected
in t h e direction of t h e breakout. This is usually well e x c e e d e d by these
very bearish formations.
• Retracement moves: Since these are very dynamic patterns that are fraught
with volatility, retracements are typically short b u t very sharp.
• Signs of false breakouts: Very difficult to spot. A rally above a previous m i n o r
high or a r e t r a c e m e n t in excess of 50 p e r c e n t of the b r e a k o u t decline can
be used.

Right-Angled Broadening B o t t o m s Quick Review
• Price characteristics: A t r a d i n g r a n g e following a l o n g decline, b o u n d e d
by two diverging trendlines. T h e u p p e r line is at or n e a r a right angle
to t h e vertical axis. Each line s h o u l d be t o u c h e d on at least two occasions. Poetic license is often r e q u i r e d in constructing these lines, since
they are n o t always t o u c h e d exactly, as in t h e case with most o t h e r patterns.
• Volume consideration: Upside b r e a k o u t is particularly bullish.
• Measuring implications: T h e m a x i m u m d e p t h of the p a t t e r n is projected
in the direction of the breakout. This is usually well e x c e e d e d by these
very bullish formations.
• Retracement moves: Since these are very dynamic patterns that are fraught
with volatility, retracements are typically short b u t very sharp.
• Signs of false breakouts: Very difficult to spot. A decline below a previous
m i n o r low or a r e t r a c e m e n t in excess of 50 p e r c e n t of t h e b r e a k o u t rally
can be used.

• All o t h e r characteristics are similar.

Miscellaneous Patterns

189

11

Miscellaneous
Patterns
Figure 11-1

Diamonds
D i a m o n d patterns really consist of a small o r t h o d o x b r o a d e n i n g formation
preceding a symmetrical triangle. An example is shown in Fig. 11-1. Edwards
a n d Magee describe a d i a m o n d as a h e a d a n d shoulders with a V-shaped
neckline, which is what it really is. Figure 11-2 shows the same example as
Fig. 11-1, b u t this time with a h e a d a n d shoulders type of interpretation.
Diamonds tend to develop m o r e at tops than at bottoms a n d usually require
some poetic license to construct. This is because the rally highs a n d decline
lows do n o t usually m a t c h up exactly with the two diverging a n d converging sets of lines required to construct an idealized pattern. Sometimes they
d o , b u t in most cases they do not.
Rectangles, with their two parallel lines of support a n d resistance, are
pretty easy to define. Diamonds, on the o t h e r h a n d , often involve some creativity in their construction. Unless you are careful, this inability to be m o r e
specific can lead to problems. To p u t it a n o t h e r way, it is easy for the novice
to visualize a d i a m o n d formation that is n o t actually there.
T h e measuring r e q u i r e m e n t works on the same principle as that for other
formations. Figure 11-1 shows this principle in action. Chart 11-1, featuring
BMC Software, offers an example of a d i a m o n d formation that t u r n e d out
to be a consolidation d u r i n g an u p t r e n d . Chart 11-2, for the same company,
shows a n o t h e r possibility, this time for a top. T h e dashed line is there to
point out the similarities between a d i a m o n d and a head-and-shoulders top.
This is because the dashed line is really the neckline of a h e a d and shoulders.

188

Diamond top.

Note that the downside objective from the d i a m o n d was achieved on the
initial decline.
Generally speaking, d i a m o n d s are n o t a very useful type of pattern. First,
they are quite rare—certainly less c o m m o n than, say, head-and-shoulders
formations. Second, a valid d i a m o n d formation is very difficult to identify
and therefore can result in misleading conclusions.

Figure 11-2 Diamond top with "butterfly" head and shoulders.

190
Chart 11-1 BMC Software, 1995-1996, daily.

Part II: Traditional Patterns

Miscellaneous Patterns

191

Giant Wedges
T h e subject of wedges is covered in C h a p t e r 12, on smaller price patterns.
These formations generally develop as short-term contra-trend consolidation patterns d u r i n g t h e course of an o n g o i n g trend. T h e giant wedges
described h e r e are typically m u c h larger a n d tend to develop at the e n d of
a trend. Occasionally they encompass the whole trend. C o m p a r e d to, say
head-and-shoulders formations or triangles, these formations are m o r e of
a collector s item. However, when they can be correctly identified, they are
often followed by very sharp reversals. A giant wedge consists of two trendlines that converge in the direction of the prevailing trend. They differ from
symmetrical triangles in that the trendlines from which a giant wedge is constructed move in the same direction. Symmetrical triangles, on the o t h e r
h a n d , consist of o n e rising a n d o n e falling line. Examples of giant wedges
lor both tops a n d bottoms are featured in Figs. 11-3 a n d 11-4

Chart 11-2 BMC Software, 1993-1994, daily.

T h e angle of convergence between the two lines is sometimes very slight
In such situations, there is only a marginal difference between a wedge a n d
a t r e n d channel, where t h e two lines are approximately parallel. T h e distinction between these two concepts is not i m p o r t a n t because a break
below the lower line or a rally above the u p p e r o n e has essentially the same
implications in b o t h formations. As a general rule, t h o u g h , we may observe
that the closer the two lines are at the b r e a k o u t p o i n t relative to where they
were at the start of the pattern, the stronger t h e signal. This is because t h e
narrowing of the pattern reflects a finer balance between buyers a n d sellers

Figure 11-3 Giant wedge top.

192

Part II: Traditional Patterns

Miscellaneous Patterns

193

Marketplace Examples

Figure 11-4 Giant wedge bottom.

Consequently, the breakout, when it comes, indicates a m o r e decisive victory.

Psychological Rationale
W h e n you think a b o u t it, a giant wedge starts off with extremely wide price
fluctuations, t h e n gradually reflects a greater balance between buyers a n d
sellers. W h e n the b r e a k o u t from a falling giant wedge comes, the balance
is b r o k e n in favor of the previous losers, i.e., the buyers. Short positions have
to be quickly u n w o u n d simultaneously with new money coming in. As a
result, an explosive advance develops. This process is reinforced by the fact
that participants b e c o m e progressively m o r e bearish as the wedge is formed,
since they are used to a series of declining peaks a n d troughs. As the wedge
narrows, a sense of calm sets in as the bears anticipate a downside breakout. T h e lack of volatility therefore lures t h e m into a sense of false security.
For technically oriented traders, it is possible to qualify any upside risk by
using a break above the u p p e r trendline of the wedge as a stop loss point.
Consequently, when the break does c o m e , everybody is a buyer, a n d this
often results in an above-average rally. T h e same rationale in reverse could
be applied to a bearish rising giant wedge in that a break to the downside
catches market participants completely by surprise. In many instances the psychology associated with breakouts from these giant patterns resembles that
associated with failures of right-angled triangles, discussed in Chapter 9.

Chart 11-3, for KLA Technology, features two giant wedges, b o t h of which
e m b r a c e complete bear markets. Each primary decline has five waves a n d
ends with an explosive move to the upside. Notice also that the lower lines
do n o t exactly touch the lows. This indicates that it is m o r e imperative to
identify the "flavor" of the price action than to try to capture precise turning points. Having said that, we really want the price to touch or c o m e very
close to the u p p e r line, since that will be the o n e that signals a reversal in
t r e n d w h e n it has b e e n successfully penetrated. If this line is constructed in
a loose m a n n e r , it will be necessary to give the price m o r e r o o m d u r i n g the
breakout, since the demarcation point will be less precise. This could lead
to unacceptable risk in the event of a false breakout. T h e example on the
left in Chart 11-3 shows that the u p p e r line is p e n e t r a t e d at approximately
the same p o i n t as the peak of the rally c o m i n g off the bottom. T h u s we get
a giant wedge signal that develops simultaneously with a reversal in the
downward peak-and-trough progression. T h e m i n i m u m ultimate price
objectives have also b e e n calculated using the classic m a x i m u m d e p t h
t e c h n i q u e . In the case of the example on the left, the projection was pretty
accurate so far as the initial rally was c o n c e r n e d .
Chart 11-4 shows a rising giant wedge for Intel. This o n e took j u s t u n d e r
six m o n t h s to form. Note how the ultimate low following the breakdown
developed at approximately four times the m i n i m u m downside objective.

Chart 11-3 KLA Technology, 1994-1999, daily.

194

Part II: Traditional Patterns

Miscellaneous Patterns

Chart 11-5 Biomet, 1982-1985, daily.
Chart 11-4 Intel, daily.

The downside breakout was followed by a pretty sharp decline. This formation is very close to a channel, since the two lines do not converge that
much. It also represents an exception to the rule stated earlier that the narrower the lines at the breakout point, the sharper the ensuing decline is
likely to be, thereby underlining the fact that technical analysis deals in probabilities rather than certainties.
Biomet, in Chart 11-5, shows another bullish falling giant wedge. In this
case, volume picks up noticeably on the upside breakout. Note also that the
breakout rally is a reasonably sharp one.
Chart 11-6, also for Biomet, shows an example of a consolidation giant
wedge. If the June bottom was interpreted as a head, you could call part of
this formation a consolidation inverse head and shoulders. I don't think it
really matters what the pattern is called. The fact is, the price broke to the
upside and represented a valid signal that the uptrend was to be resumed.
Finally, Sterling Bancorp, in Chart 11-7, provides an example of a bullish
giant wedge, where both the single and double multiples of the projection
proved to be strong resistance points. If you look very carefully, you can see
that the price action between the end of 1999 and the breakout point was
really a downward-sloping head-and-shoulders bottom. The upper line
forming the wedge represented the neckline. The breakout therefore
offered a triple signal: completion of the pattern, a rising series of peaks and
troughs, and, of course, a breakout from a giant wedge.

Chart 11-6 Biomet, daily.

195

196

Part II: Traditional Patterns

Miscellaneous Patterns

197

Chart 11-7 Sterling Bancorp, 1996-2001, weekly.

Figure 11-5 Rounding (saucer) bottom.

Rounding Bottoms (Saucers)
and Tops
Most patterns offer clearly definable b r e a k o u t points, b u t r o u n d i n g patterns
do not. W h e n plotted on a chart, they look interesting, b u t from a practical point of view they are n o t very useful because they are usually identifiable only well after the fact. Figures 11-5 a n d 11-6 show the formation of a
saucer ( r o u n d i n g bottom) a n d a r o u n d i n g top. A saucer pattern occurs at
a market b o t t o m , while a r o u n d i n g top develops at a market peak. A classic saucer is constructed by drawing a circular line, which roughly approximates an elongated or saucer-shaped letter U, u n d e r the lows. As the price
drifts toward the low p o i n t of t h e saucer, investors lose interest a n d downward m o m e n t u m dissipates. This lack of interest is also reflected by the level
of activity, which almost dries up at the time the price is reaching its low
point. As the formation is completed, b o t h price a n d volume experience a
rapid acceleration to the upside. Occasionally they experience a sideways
trading r a n g e at higher levels. If that h a p p e n s , the b r e a k o u t from this consolidation offers a timely buy signal. Risk can be controlled by placing a stop
below the lower level of the trading range. According to Edwards a n d
Magee, r o u n d i n g bottoms have a tendency to develop in low-cap stocks,
where they take many m o n t h s to complete.
T h e price action of the r o u n d e d top (or inverted bowl, as it is sometimes
called) is exactly opposite to that of t h e saucer p a t t e r n , b u t the volume

characteristics are the same. This means that if volume is plotted below the
price, it is almost possible to draw a complete circle, as shown in Fig. 11-6. T h e
tip-off about the bearish implications of the r o u n d e d top is the fact that volu m e shrinks as prices reach their highest levels, then expands as they fall. Both
these volume characteristics are bearish and are discussed in greater detail in
Chapter 5.

Figure 11-6 Rounding top.

198

Part II: Traditional Patterns

R o u n d i n g tops a n d bottoms are fine examples of a gradual changeover
in the d e m a n d / s u p p l y balance that slowly picks up m o m e n t u m in the direction opposite to that of the previous trend. Quite clearly, it is difficult to
obtain b r e a k o u t points for these patterns, since they develop slowly a n d do
n o t offer any clear support or resistance levels on which to establish a potential b e n c h m a r k . Even so, it is worth trying to identify t h e m , since they are
usually followed by substantial moves. R o u n d i n g a n d saucer formations can
also be observed as consolidation p h e n o m e n a a n d can take as little as three
weeks or as m u c h as several years to complete.
Chart 11-8, for Bed Bath & Beyond, shows a classic r o u n d i n g top. Volume
shrinks as the high is being reached a n d expands greatly as prices accelerate on the downside. O n c e again, the p r o b l e m with a pattern such as this is
that there is no clearly definable b e n c h m a r k beyond which you can say that
the trend has reversed to the downside, or even that the pattern is invalid.
Chart 11-9, for German American, shows a n o t h e r r o u n d i n g top. This time
the volume configuration is n o t consistent, b u t the pattern works anyway.
T h e price tried to trace out a small r o u n d i n g b o t t o m b u t failed. This was
fairly obvious once it h a d slipped below the s u p p o r t trendline.
Synovus Financial (Chart 11-10) traces o u t a sort of consolidation saucer
pattern between September 1993 a n d April 1995. T h e two converging lines
show that the pattern could also be categorized as a right-angled triangle.
O n c e again, it does n o t matter what the p a t t e r n is called. It's the fact that
it worked that's important. As you can see, the battle between buyers a n d

Miscellaneous Patterns

Chart 11-9 German American, 1995-1999, weekly.

sellers gradually moved in favor of the buyers as prices moved higher. This
idea of a saucer b o t t o m b e i n g a consolidation formation is an i m p o r t a n t
o n e , since it reflects a long period of controlled profit taking as buyers gradually gain sufficient confidence to enable t h e m to push prices higher.

Chart 11-10 Synovus Financial, 1992-1995, weekly.
Chart 11-8 Bed Bath & Beyond, daily.

199

200

Part II: Traditional Patterns

Chart 11-11 Sterling Financial, 1994-1998, weekly.

Miscellaneous Patterns

201

denied. T h u s it is possible to extend the idea of a saucer b o t t o m or rounding top to e m b r a c e a circular line that reflects a series of gradually rising
bottoms or peaks, d e p e n d i n g on the n a t u r e of the situation. T h e c o n c e p t
of a gradual reversal in which buyers gain the u p p e r h a n d is still the same.

Cup with a Handle

Sterling Financial (Chart 11-11) also offers an example of a consolidation
r o u n d i n g bottom. In this case, the saucerlike pattern is really the right shoulder of a consolidation reverse head-and-shoulders formation.
Chart 11-12, for Wachovia, really features a consolidation reverse h e a d
a n d shoulders. However, the r o u n d i n g n a t u r e of the pattern c a n n o t be

This pattern has been m a d e famous by William O'Neil a n d is described in his
How to Make Money in Stocks (McGraw-Hill, 1995). T h e pattern develops as a
bullish one, usually in a continuation format. Figure 11-7 shows that it takes
the form of a big U (the c u p ) , followed by a rally and a small r o u n d i n g platform (the h a n d l e ) . T h e left-hand part of the cup usually marks the culmination of a strong rally a n d is often associated with heavy volume. T h e bottom
of the cup can take the form of a r o u n d i n g bottom, as in Fig. 11-7, or some
ranging action, as in Fig. 11-8. T h e next step in the development of this pattern is a rally on e x p a n d i n g volume, followed by a period of profit taking in
which b o t h volume a n d price go quiet. Finally, the handle is completed and
prices explode to the upside.
If this p a t t e r n is going to fail, the signal to look for is a break below the
lower part of the h a n d l e . If the price eventually breaks above the u p p e r level
of the h a n d l e , the situation will again b e c o m e bullish. Any breakouts that
develop with shrinking volume, t h o u g h , should be r e g a r d e d with suspicion.
Chart 11-13 shows a c u p with a h a n d l e formation for ADC Telecom. T h e
b r e a k o u t above the h a n d l e is n o t a c c o m p a n i e d by m u c h of an expansion

Chart 11-12 Wachovia, 1992-1995, weekly.

Figure 11-7 Cup with a handle.

Part II: Traditional Patterns

r

Miscellaneous Patterns

203

202
Chart 11-14 Adelphia, daily.

Figure 11-8 Cup with a handle.

in volume, b u t t h e price certainly doesn't suffer. T h e c u p in Chart 11-14,
for Adelphia, is m o r e of a V formation. William O'Neil clearly states that
the c u p should be r o u n d e d , so that a good solid base can be formed for
later price advances. In this case the base is really formed d u r i n g the h a n d l e

stage on very quiet action. W h e n this e x t e n d e d period of profit taking is
completed, b o t h price a n d volume explode to the upside.

Chart 11-13 ADC Telecom, 1990-1991, daily.

Shakeout Tops and Bottoms
A shakeout p a t t e r n is really a variation of o t h e r patterns. At tops, it represents a consolidation following a strong rally a n d a subsequent sharp reaction. I call these patterns shakeouts because the price is shaken o u t of the
consolidation on its way to its bear market low. Since these formations are
really part of a m u c h larger reversal process, they are typically followed by
an eventual decline that is greater than would be suggested by the size or
depth of the pattern. Figure 11-9 shows an example in which a head-and-shoulders consolidation is the straw that breaks the camel's back. In Fig. 11-10, the
shakeout pattern is a rectangle. These formations also a p p e a r at bottoms,
for which an example is shown in Fig. 11-11. H e r e we can see that an a b r u p t
reversal is subsequently followed by a trading range. T h e ranging action in
this example is a rectangle, but it could j u s t as easily have b e e n a reverse
h e a d a n d shoulders or a triangle. Like tops, these b o t t o m i n g formations are
often followed by advances that last m u c h longer a n d take prices m u c h further than would be expected from their size a n d d e p t h . Examples of these
p a t t e r n s in action a r e shown in Charts 11-15 a n d 11-16. N o t e also in

Miscellaneous Patterns

204

Figure 1 1 - 9

Shakeout top.

Figure 1 1 - 1 1

Chart 1 1 - 1 5

Figure 1 1 - 1 0

Shakeout top.

Shakeout bottom.

AMR, 1996-2002, weekly.

205

Part II: Traditional Patterns

206
Chart 11-16 Yahoo, 1998-1999, daily.

Miscellaneous Patterns

207

four to six m o n t h s to form. Occasionally they envelop a complete primary
trend.
• Measuring implications: T h e m a x i m u m d e p t h of the pattern is projected
in the direction of the breakout.
• Special characteristics: Giant wedges are unusual, b u t when they are completed, they are often followed by explosive long-term moves.

Rounding Tops and B o t t o m s Quick Review
• Price characteristics: R o u n d i n g bottoms consist of a saucerlike pattern,
where the price slowly b u t surely gathers substantial upside m o m e n t u m .
Tops consist of a slow r o u n d i n g process that gradually picks up m o m e n t u m on the downside.
• Volume characteristics: Both tops a n d bottoms are a c c o m p a n i e d by a saucerlike p a t t e r n in volume.
• Special characteristics: Because of the r o u n d i n g n a t u r e of these patterns,
there are no definable benchmarks beyond which a clear-cut breakout can
be identified.
C h a r t 11-15 the completion of an upward sloping consolidation h e a d a n d
shoulders, where the initial low developed at exactly twice the measuring
objective.

Cup with a Handle Quick Review

Summary

• Price characteristics: A cup-shaped consolidation p a t t e r n that forms after a
rally. T h e h a n d l e forms p a r t of the way up the right-hand side of the cup
a n d is a period of profit taking. T h e pattern is c o m p l e t e d with a break
above the h a n d l e .

Diamonds Quick Review
• Price characteristics: A p a t t e r n that consists of an o r t h o d o x b r o a d e n i n g formation followed by a symmetrical triangle; w h e n c o m b i n e d , the two patterns have the a p p e a r a n c e of a d i a m o n d . They develop as b o t h reversal
a n d continuation formations b u t are difficult to identify correctly.
• Volume considerations: E x p a n d i n g volume on an upside b r e a k o u t is particularly bullish.
• Measuring implications: T h e m a x i m u m d e p t h of the p a t t e r n is projected
in the direction of the breakout.

Giant Wedges Quick Review
• Price characteristics: Price action b o u n d e d by two converging trendlines
moving in the same direction. These are large formations that take at least

• Volume considerations: T h e two key ingredients are low volume d u r i n g the
formation of t h e h a n d l e a n d e x p a n d i n g volume on the upside breakout.
• How to recognize failures: Shrinking volume on the b r e a k o u t or a price pullback below the lower level of the h a n d l e .

PART III
Short-Term
Patterns

12

Smaller Patterns
and Gaps
Most of the patterns discussed in this chapter are relatively small and are of
the continuation variety. They reflect controlled profit taking during an
advance and controlled digestion of losses during a decline. Flags, pennants,
and wedges all consist of a trading range bounded by two trendlines sloping in a similar direction. Typically the slope develops in a countercyclical
way. This means that bullish patterns will slope in a downward direction and
develop in a bull market. Similarly, bearish patterns will slope in an upward
direction and form during primary bear moves.

Flags
A flag is a quiet parallel trading range accompanied by a trend of declining volume. Such formations usually interrupt a sharp, almost vertical
price rise or decline. As the name implies, this formation looks like a flag
on the chart. Since they are continuation patterns, flag completions
involve a breakout in the same direction as the previous trend. Examples
for both an up and a down market are shown in Fig. 12-1. Essentially, flags
take the form of a narrow trading range in which the rally peaks and reaction lows can be connected by two parallel lines. The lines move in a countercyclical direction. In the case of a rising market, the flag is usually
formed with a slight downtrend, but in a falling market it has a slight
upward bias. Flags may also be horizontal, in which case they are really a
special form of rectangle.
The classic texts tell us that in a rising market, an idealized pattern usually separates two halves of an almost vertical rise. Volume is normally
211

212

Part III: Short-Term Patterns

Smaller Patterns and Gaps

213

move is usually well worthwhile from a trading point of view. More to the
point, it is usually pretty fast.

Measuring Implications

Figure 12-1 Flags.

extremely heavy j u s t before the point at which the flag formation begins.
As the formation develops, activity gradually contracts to almost nothing. It
t h e n explodes as the price works its way out of the c o m p l e t e d formation.
On the daily charts, flags can form in a period as short as five days or take
as long as t h r e e to five weeks. It is also true to say that flags are starting to
b e c o m e an e n d a n g e r e d species. Searching for examples for this book, especially a pause within a vertical rise or fall, b e c a m e quite a challenge.
Whenever a concept becomes this difficult to demonstrate with real-live
examples, I start to question its usefulness.
T h e formation of the flag in a d o w n t r e n d is also a c c o m p a n i e d by declining volume. This type of flag represents a trading range with an upward bias
in price, so the volume implication is bearish in nature, i.e., rising price with
declining volume. W h e n the price violates the lower part of the flag, the
sharp slide continues a n d volume tends to pick u p . However, it n e e d n o t be
explosive, because prices can j u s t as easily fall due to a lack of bids as they
can because of strong selling pressure. Only upside breakouts in bull markets require heavy volume.
It is important to make sure that the price and volume characteristics agree.
For example, in a bull trend, the price may consolidate following a sharp rise,
in what appears to be a flag formation, but volume may fail to contract appreciably. In such cases, great care should be taken before coming to a bullish
conclusion, since the price may well react on the downside. A flag that takes
m o r e than four weeks to develop should also be treated with caution. This is
because these formations are, by definition, temporary interruptions of a
sharp u p t r e n d . A period in excess of four weeks represents an unduly long
time for profit taking, a n d therefore has a lower probability of being a true
flag.
Flags are usually reliable patterns from a forecasting point of view. Not
only is the direction of the ultimate b r e a k o u t indicated, b u t the ensuing

Technical folklore has it that flags form at the halfway point of a move. O n c e
the b r e a k o u t has taken place, a useful m e t h o d for setting a price objective
is to estimate the size of the move in the period immediately before the flag
formation began. This distance is t h e n projected in the direction of the
b r e a k o u t from the b r e a k o u t point. In technical j a r g o n , flags are said to fly
at half-mast, i.e., they are halfway up the move. Since they take a relatively
short period to develop, flags do n o t show up on weekly or monthly charts.
Chart 12-1, for American Electric Power, features a bearish flag that develo p e d d u r i n g a major down move. Note that the volume level shrank to
almost n o t h i n g as the flag was being formed. It then exploded on the day
of the breakout. Unfortunately, there was no easy way to get out because of
the h u g e gap. Normally we would expect to see at least some trading within
the flag trading range on the day of the breakout.
Chart 12-2, for Adelphia, shows two flags. T h e first, on the left, took barely
a week to form. Volume started to pick up on the day of the breakout, a n d
by the second day e x p a n d e d to a very heavy level. T h e second o n e develo p e d in a period of slightly m o r e than t h r e e weeks. In this case, the volume

Chart 12-1 American Electric Power, 1995-1996, daily.

214

Part III: Short-Term Patterns

Chart 12-2 Adelphia, daily.

pattern differed from the classic pattern, since it e x p a n d e d d u r i n g the formation of the late March-early April rally. Normally we expect activity to
gradually shrink as the p a t t e r n is being formed. Only on the b r e a k o u t is it
expected to e x p a n d .
C h a r t 12-3, for A d a p t e c , features a bullish flag. O n c e again volume
increased substantially on t h e upside b r e a k o u t b u t did n o t e x p e r i e n c e the
n o r m a l noticeable profit-taking shrinkage d u r i n g t h e formation of the
pattern.
Finally, Chart 12-4 shows a n o t h e r example of a bullish flag for Adaptec.
H e r e t h e r e is a very definite contrast between the shrinking volume during
the flag's formation a n d the e x p a n d i n g activity at its completion. Note also
that while the measuring implication was m o r e than achieved, the actual
level of the projection did serve as a resistance level for a few sessions.

Pennants
A p e n n a n t develops u n d e r exactly the same circumstances as a flag a n d has
similar characteristics. T h e difference is that this type of consolidation formation is constructed from two converging trendlines, as shown in Fig. 12-2.
In a sense, the flag corresponds to a rectangle and the p e n n a n t to a triangle,
because a p e n n a n t is, in effect, a very small triangle. T h e difference between
the two formations is that a triangle consists of a trading range b o u n d by

Smaller Patterns and Gaps

215

Chart 12-3 Adaptec, daily.

two converging trendlines that p o i n t in different directions. In the case of a
p e n n a n t , the two trendlines b o t h move in t h e same direction. If anything,
volume tends to contract even m o r e during t h e formation of a p e n n a n t than
d u r i n g that of a flag. In every o t h e r way, however, p e n n a n t s are identical to

Chart 12-4 Adaptec, daily.

216

Part III: Short-Term Patterns

Smaller Patterns and Gaps

217

Chart 12-6 Vitesse, daily.

Figure 12-2 Pennants.

flags: measuring implication, time taken to develop, volume characteristics,
a n d so on.
Chart 12-5, featuring Adobe, displays a rwo-week p e n n a n t . In this case, tne
trading range is extremely tight a n d volume contracts until the breakout.

Chart 12-5 Adobe, daily.

T h e next example of a p e n n a n t is provided by Vitesse in Chart 12-6. It
forms part of a r o u n d i n g top consolidation formation. In this instance, volu m e contracts d u r i n g t h e formation o f t h e p e n n a n t b u t does n o t e x p a n d
on the gentle downside breakout.
Chart 12-7, for Veritas Software, shows a different type of formation. In
this case we see a triangle formation with converging lines moving in the
same direction. However, the two converging lines in a classic p e n n a n t
should be moving in a contra-trend direction, b u t in this instance, they are
b o t h rising d u r i n g an o n g o i n g rising t r e n d instead of falling. T h e underlying psychology, t h o u g h , is still the same, since the battle between buyers and
sellers gradually b e c o m e s m o r e balanced as the trading range a n d accompanying activity contract. Finally o n e side scores a decisive victory as prices
break to the upside on very heavy volume. T h e r e is also a flag that develo p e d in the May-June period.
Chart 12-8, also for Veritas, shows a m o r e classic bullish p e n n a n t . Note
how the price explodes d u r i n g the breakout. Later on we see a flag, a n d
finally a consolidation triangle. T h e contrast between the p e n n a n t , with its
two trendlines pointing in the same direction, a n d the triangle, where the
trendlines point in different directions, is also a p p a r e n t from this chart.
USA Networks (Chart 12-9) presents us with a n o t h e r flag. It also features
a p e n n a n t ; o n c e again the two lines slope in the same direction as the main
trend, as in Chart 12-7. In this case the " p e n n a n t " has a slight upward slope,
so it is n o t a classic example, which should, of course, point in a contra-trend

218

Part III: Short-Term Patterns

Smaller Patterns and Gaps

219

Chart 12-9 USA Networks, daily.
Chart 12-7 Veritas Software, daily.

Chart 12-8 Veritas Software, 1994-1995, daily.

direction. However, whenever we see this kind of a situation in which a very
tight a n d narrowing balance between buyers a n d sellers edges u p , the
upside b r e a k o u t is often explosive. This example was no exception to that
rule. Volume explodes on the breakout, a n d the rally is c a p p e d with a twobar reversal (see C h a p t e r 16 for an e x p l a n a t i o n ) .

Wedges
A wedge is very similar to a triangle in that two converging lines can be constructed from a series of peaks a n d troughs, as shown in Fig. 12-3, but,
whereas a triangle consists of o n e rising a n d o n e falling line, or o n e rising
or falling line a n d o n e horizontal line, the converging lines in a wedge both
move in the same direction, as for a p e n n a n t . A falling wedge represents a
temporary i n t e r r u p t i o n of a rising trend, a n d a rising wedge is a temporary
interruption of a falling trend. It is n o r m a l for volume to contract d u r i n g
the formation of b o t h types of wedge. Since they can take anywhere from
two to eight weeks to complete, wedges sometimes occur on weekly charts,
but they are too brief to a p p e a r on monthly charts unless they are giant
wedges, as described in C h a p t e r 11.
Rising wedges are fairly c o m m o n as bear market rallies. Following their
completion, prices usually break very sharply, especially if volume picks up
noticeably on the downside.

220

Part III: Short-Term Patterns

Smaller Patterns and Gaps

221

Chart 12-10 Yahoo, daily.

Figure 12-3 Wedges.

The wedge and the pennant are very similar, since they both consist of converging trendlines that move in a contra-trend direction. The difference is
that the breakout point of a pennant forms very close to or even right at the
apex. The two projected lines for the wedge, on the other hand, would meet
way in the future-in many instances, literally off the charts. Figure 12-4 puts

Figure 12-4 A pennant vs. a wedge.

us straight on this one, as you can see that the projected dashed lines for the
wedge meet well after the breakout point. Compare this to the pennant,
which is much more akin to a triangle. Sometimes the difference between
these two formations is hard to judge. For example, the pattern in Chart 12-6
was called a pennant, but it could arguably have been called a wedge. If anything, wedges generally appear to take longer to form than pennants.
Chart 12-10, for Yahoo, shows a rising wedge that develops after the first
decline. You can see that the lines are not even close to a meeting point as
the breakout develops. Note also that volume picks up the day of the breakout. During a bear trend, it is possible for prices to fall of their own weight.
However, when activity picks up, it indicates that the sellers are motivated
and makes the whole situation that much more bearish.
Chart 12-11, featuring Vitesse, offers a good example of a small rising
wedge in terms of price action. Volume characteristics, though, are not
typical, since there is no contraction as the wedge is forming and no
expansion on the breakout. This example shows that it is possible for
wedges, like other formations, to develop without the classic volume configuration. In most situations, though, the gradual shrinking of activity
as the pattern forms and the explosion on the breakout usually result in
a more powerful move. Also, when volume contracts during the formative period, it offers a valuable clue that the pattern will turn out to be
valid.

222

Part III: Short-Term Patterns

Smaller Patterns and Gaps

223

Chart 12-11 Vitesse, daily.

Figure 12-5 Breakaway gap.

Gaps
A gap occurs when the lowest price in a specific trading period is above the
highest level in the previous trading period or when the highest price in a
specific trading period is below the lowest price in the previous trading
period (Fig. 12-5). On a daily bar chart, the trading period would be a day;
on a weekly chart, a week; a n d so forth.
Gaps do n o t a p p e a r on line or close-only charts, b u t are confined to bar
charts. They are r e p r e s e n t e d by an empty vertical space between o n e trading period a n d another, a n d they reflect highly emotional periods. They
most commonly form in overnight trading as good or b a d news is digested
by the market. Daily gaps are far m o r e c o m m o n than weekly ones because
a gap on a weekly chart can occur only between Friday's a n d Monday's price
range; i.e., it has a l-in-5 chance relative to a daily chart. Monthly gaps are
even rarer, since such "holes" on the chart can develop only between
monthly price ranges. T h e most typical place to find gaps on intraday charts
is at the o p e n . I will have m o r e to say on that point later.

The Importance of Gaps as Emotional Points
T h e places where gaps start a n d terminate are potential pivotal points on a
chart because they r e p r e s e n t high e m o t i o n . If you have an a r g u m e n t with

a friend a n d o n e of you shouts really loudly at o n e point, you will b o t h tend
to r e m e m b e r that particular m o m e n t because it represents an emotional
extreme. T h e same principle can be applied to technical analysis, since
charts are really a reflection of psychological attitudes. This means that when
the price returns to the area of previous gaps, the u p p e r a n d lower points
of those gaps have the potential to b e c o m e important support and resistance
levels where short-term trends may be temporarily reversed.

T h e P s y c h o l o g i c a l R a t i o n a l e for W h y M o s t G a p s
A r e E v e n t u a l l y Filled
A gap is said to be closed, or "filled," when the price reverses a n d retraces
the whole r a n g e of the gap. On daily charts this process sometimes takes a
few days, a n d at o t h e r times takes a few weeks or m o n t h s . In some rare
instances the process is never completed.
T h e r e is an old saying that the market abhors a vacuum, which means that
most gaps are eventually filled. It is certainly true that almost all eventually
are, b u t exceptions definitely occur. T h e underlying psychology is akin to,
say, a h u s b a n d a n d wife having a spirited disagreement. Tempers on b o t h
sides rise to a pretty high level as accusations a n d counteraccusations escalate. T h e n , almost invariably, after some time has elapsed a n d as o n e or both
parties p u t pride aside a n d see the e r r o r of their ways, the relationship
returns to normal, a n d an a t t e m p t is m a d e to close the emotional gap in
the relationship.

224

Part III: Short-Term Patterns

Since it can take m o n t h s or even years to fill a gap, trading strategies
should n o t be i m p l e m e n t e d solely on the assumption that a gap will be filled
in the immediate future. In the majority of cases, some attempt is made to fill the
gap, b u t quite often a partial filling on a subsequent test is sufficient. Most
gaps are closed because they are emotional affairs a n d reflect strong psychological motivation on the part of traders—we could say excess fear or
greed, d e p e n d i n g on the direction of the trend. Decisions to buy or sell at
any cost are n o t objective ones. This m e a n s that people are likely to have seco n d thoughts when things have cooled down. T h e second thoughts, in this
case, are r e p r e s e n t e d by the closing of the gap, or at least a good attempt
at closing it.
Gaps should be treated with respect, b u t their importance should n o t be
overemphasized. Those that occur d u r i n g the formation of a price pattern,
known as common gaps or area gaps, are usually closed fairly quickly a n d do
n o t have m u c h technical significance. A n o t h e r insignificant type of gap
results from a stock's going ex-dividend.
T h e r e are three o t h e r types of gaps that we n e e d to examine. These are
breakaway, runaway, a n d exhaustion gaps.
Breakaway Gaps
A breakaway gap is created w h e n a price breaks out of a price p a t t e r n or
some o t h e r trading r a n g e configuration. An example of an upside breakaway gap is shown in Fig. 12-5. Generally speaking, the p r e s e n c e of the gap
emphasizes the bullishness or bearishness of the breakout, d e p e n d i n g on
which direction it takes. Even so, it is still i m p o r t a n t for an upside breako u t to be a c c o m p a n i e d by a relatively high level of volume. Gap breakouts
that occur on the downside are n o t r e q u i r e d to be a c c o m p a n i e d by heavy
volume.
It should n o t be c o n c l u d e d that every gap b r e a k o u t will be legitimate
because the "sure things" do n o t exist in technical analysis. However, a gap
that is associated with a b r e a k o u t is m o r e likely to be valid than o n e that is
n o t so associated. If a gap does t u r n out to be a whipsaw, t h e n this will usually be signaled sooner r a t h e r t h a n later. Since most gaps are filled, and
there is rarely a reason why you have to buy, it could be argued that it is better to wait for the price to at least a t t e m p t to fill the gap before committing
money. After all, if you miss out because the price does n o t experience a
retracement, all you have lost is an opportunity. Certainly you will experience some frustration, b u t at least you will n o t have lost any capital. With
markets there is always a n o t h e r opportunity. If you have the patience and
the discipline to wait for that opportunity, you will be m u c h better off in
the long r u n . T h e p r o b l e m , especially in this day a n d age of shrinking time

Smaller Patterns and Gaps

225

spans, is that most of us are n o t blessed with the patience a n d discipline that
we so badly require for successful trading a n d investing.
T h e d a n g e r of buying on a gap b r e a k o u t is that you will get caught up in
the emotions of the crowd. This buy-at-any-cost mentality is likely to result
in d i s c o u r a g e m e n t w h e n the price inevitably retraces to the downside as
emotions calm down. T h e advice is n o t that you should never buy a gap
breakout, b u t that you should think very carefully a n d mentally p r e p a r e
yourself for the high probability that the price will correct, thereby placing
your position temporarily u n d e r water.
Breakaway gaps that develop d u r i n g the early stages of a primary bull
m a r k e t are m o r e likely to be valid t h a n those that develop after a long
price advance. This is because y o u n g bull markets have a t r e m e n d o u s
a m o u n t of upside m o m e n t u m . This m e a n s that t h e r e is less likelihood of
indecisiveness b e i n g reflected in the charts in the form of r e t r a c e m e n t
moves a n d t r a d i n g r a n g e s . O n t h e o t h e r h a n d , breakaway gaps that
develop at t h e e n d of a bull move are m o r e likely to indicate e m o t i o n a l
e x h a u s t i o n as t h e sold o u t bulls literally give up on any possibility of b e i n g
able to buy again at lower prices. T h e same principle in reverse applies to
b e a r trends.
In Technical Analysis of Stock Trends, Edwards a n d Magee have a slightly different take. T h e i r advice a b o u t w h e t h e r to buy a breakaway gap rests on the
volume configuration. They state that if volume is high j u s t prior to the gap
a n d shrinks as the price moves away from the u p p e r part of the gap, then
there is a 50-50 c h a n c e of a retracement. On the o t h e r h a n d , if volume
expands at the u p p e r part of the gap as prices move away from it, t h e n the
odds of a r e t r a c e m e n t or gap-closing effort are substantially less. Such characteristics, they imply, should be b o u g h t into.
I think this can be taken a little further by setting a three-step rule for
buying b r e a k o u t gaps. A theoretical example is shown in Fig. 12-6. First, it's
i m p o r t a n t for the gap to develop at the b e g i n n i n g of a move, which implies
that it should be p r e c e d e d by at least an intermediate decline. In o t h e r
words, if a gap is to r e p r e s e n t a sustainable c h a n g e in psychology, it must
have some pretty bearish psychology (as witnessed by the preceding decline)
to reverse. Second, the gap day should be a c c o m p a n i e d by exceptionally
heavy volume. This again reflects a change in psychology because the bulls
are very m u c h in control. Third, an attempt to close the gap should be m a d e
within two to four days, a n d the price should take out the high of the day
of the gap. If an a t t e m p t to close the gap fails, so m u c h the better. T h e idea
of the test is that m a r k e t participants have h a d a chance to change their
(bullish) minds a n d did not. T h e part of the rule about the new high is really
a way of d e t e r m i n i n g w h e t h e r the market confirms the gap following the
successful test.

226

Part III: Short-Term Patterns

Figure 12-6 Three-step rule for buying breakaway gaps.

Chart 12-12, for Yahoo, shows a nice breakaway gap at the start of a major
market move. Volume on the day of the gap was exceptionally heavy. In this
case, the m o m e n t u m was so powerful that there was no r e t r a c e m e n t move
whatsoever. This gap h a d two things going for it: the heavy volume a n d the
fact that the m a r k e t itself was just c o m i n g off a major low.

Smaller Patterns and Gaps

Breakaway gaps are often followed by worthwhile moves, b u t there do not
a p p e a r to be any useful measuring yardsticks that can be applied. Perhaps
this is because these gaps develop very early, as the new t r e n d is j u s t getting
underway. As a result, all measuring objectives would be quickly attained and
exceeded. In such cases, t h e measuring objective is usually better obtained
from the p a t t e r n or o t h e r entity from which the breakaway gap develops.
Chart 12-13, for Amazon, shows a breakaway gap in early 2003. In this
instance, the gap was p r e c e d e d by a small decline. Rules 2 (exceptionally
heavy volume) a n d 3 (test of the u p p e r area of the gap, followed by a postgap-day high on the fifth day following the day of the gap) were also in force.
Note that in this case, the b r e a k o u t does n o t c o m e from a price pattern,
b u t from a down trendline.
In Chart 12-14, featuring Apple, rule 1 was satisfied, although the previous decline, in which the stock h a d come down from $30, is n o t shown.
Rules 2 (heavy volume) a n d 3 (test of the gap followed by a new high) are
m e t inside the ellipse. T h e r e was also a subsequent attempt to fill the gap
in D e c e m b e r 1993. Note how its u p p e r area b e c a m e a great s u p p o r t zone
that halted the decline. We see a n o t h e r gap develop in early 1994 on very
heavy volume a n d a subsequent test of its u p p e r area. What is striking is that
the confirmation is a kind of a double signal because it simultaneously takes
out the gap day's high a n d the high for the previous move. It would be nice
to say that the price went on to double after this, b u t that was n o t the case;

Chart 12-13 Amazon, daily.
Chart 12-12 Yahoo, 2002-2003, daily.

227

228

Part III: Short-Term Patterns

Chart 12-14 Apple, 1993-1994, daily.

it barely m a d e the $19 level before declining to the lower level of the initial gap (on the left), or j u s t over $12.
A downside breakaway gap appears in Chart 12-15, for Amazon. This
b r e a k o u t developed at the e n d of a bear m a r k e t rally. In this case the gap
day was associated with exceptionally heavy volume a n d closed near its low.
A weak attempt to close the gap developed o u t of a small r o u n d i n g top.
Eventually the level of the post gap low was taken out. This would have been
the signal for any diehards to liquidate.

Continuation, or Runaway, Gaps
Runaway gaps occur during a straight-line advance or decline, when price quotations are moving rapidly and emotions are running high (see Fig. 12-7a).
Sometimes they are closed very quickly, e.g., within a day or so. Alternatively,
they may tend to remain open for m u c h longer periods and generally n o t be
closed until the market makes a major or intermediate swing in the opposite direction to the price m o v e m e n t that was responsible for the gap. This
type of gap often occurs halfway between a previous b r e a k o u t a n d the ultimate d u r a t i o n of the move. For this reason, continuation gaps are sometimes called measuring gaps. Occasionally m o r e than o n e continuation gap
develops d u r i n g a price move.
Continuation gaps are far m o r e likely to show up in thinly traded markets or stocks than in well-seasoned issues. This is probably because the door

Smaller Patterns and Gaps

229

Chart 12-15 Amazon, daily.

of opportunity is m u c h narrower, a n d if everyone is trying to get in or out
at the same time, there are fewer people who can be a c c o m m o d a t e d at the
desired price. Consequently, their buying (selling) d e m a n d s can be accomm o d a t e d only at m u c h h i g h e r (lower) prices.
Chart 12-16, featuring Yahoo, shows a runaway or measuring gap in
September. Note how it develops about halfway up the move, as indicated
by the arrows. As soon as the objective was obtained, the price quickly
reversed a n d r e t u r n e d to the u p p e r area of the gap. T h e chart also features
a breakaway gap that was retraced about 20 trading days later. T h e n the price
exploded to the upside.

Figure 12-7

Runaway and exhaustion gaps.

230

Part III: Short-Term Patterns

Smaller Patterns and Gaps

231

Chart 12-17 Yahoo, daily.
Chart 12-16 Yahoo, daily.

Chart 12-17 also features Yahoo. H e r e we can see that the runaway gap
in January l a u n c h e d a rally that eventually t o p p e d out at four times the measuring objective. T h e initial objective also worked, since it halted the rally
for a couple of days. Note also the two breakaway gaps that started the move
a n d the failed attempt to completely close the first of these.

Exhaustion Gaps
W h e n a price move contains m o r e than o n e runaway gap, this indicates that
a very powerful t r e n d is in force. T h e presence of a second or third gap
should also alert the technician to the fact that the move is rapidly maturing. H e n c e , there is a possibility that a second or third runaway gap will be
the final o n e . An exhaustion gap is, therefore, associated with the terminal
phase of a rapid advance or decline a n d can be the last in a series of runaway gaps (Fig. 12-76). Alternatively, an exhaustion gap may merely develop
after a long, protracted price move.
In effect, the breakaway gap represents the start of a move, the runaway
gap is in the middle, a n d the exhaustion gap indicates the terminal phase.
Exhaustion gaps are therefore associated with rapid a n d protracted price
moves a n d signal the b e g i n n i n g of the giving up phase, as buyers convince
themselves that a buying opportunity on a correction will never develop.

In a downtrend, sellers persuade themselves that a relief rally will never allow
t h e m to get out with a smaller loss. Indeed, in Profits in the Stock Market, H. M.
Gartley argues that downside exhaustion gaps often develop between o n e
a n d t h r e e days prior to a selling climax session.
Spotting exhaustion gaps is not easy, for how do you know at the time that
the gap is n o t a runaway gap? First of all, if the gap develops close to the
b e g i n n i n g of a move, it is m o r e likely to be a breakaway type, so the probability of its being an exhaustion gap can be m o r e or less ruled out. T h e
odds are even greater if the objective called for by a price p a t t e r n or trendline b r e a k o u t has n o t yet b e e n achieved.
O n e clue that an exhaustion gap may be forming is an unusually heavy level
of volume in relation to the price change for that day. In such a case, volume
usually works up to a crescendo, well above previous levels. Sometimes the
price closes near the vacuum (or gap) and well away from its extreme reading. If the next day's trading creates an "island," with the gap day being completely isolated from the previous day's trading by a vacuum, this is usually an
excellent sign that the gap day was in fact the turning point. This indicates
only temporary exhaustion, b u t it should be a red flag to highly leveraged
traders that they should liquidate or cover their positions.
Alternatively, the day of the gap or o n e of the subsequent sessions may
develop into a one-day price pattern, such as those described in subsequent
chapters. In that event, the gap a n d the price action will reinforce each

232

Part III: Short-Term Patterns

other, thereby increasing the probabilities of a near-term reversal. In any
event, an exhaustion gap should n o t be r e g a r d e d as a sign of a major reversal, b u t merely as an indication that, at the very least, some form of consolidation should be expected.
If you are m o n i t o r i n g an oscillator that is overbought (or oversold, in the
case of a decline) or has experienced a divergence or two with the price,
this indicates an overextended price trend. Such a combination is m o r e
likely to be associated with an exhaustion gap than an oscillator that is at a
neutral reading.
Edwards a n d Magee p o i n t o u t that runaway gaps are usually left o p e n for
some time a n d typically require an intermediate or even primary t r e n d in
the opposite direction to the gap before they are closed. On the other h a n d ,
since an exhaustion gap is indicative of the terminal phase of a trend, it is
usually closed within a few sessions. Such action provides strong evidence
that the gap is n o t of the measuring or continuation variety. Even then, however, it should be appreciated that such gaps in a n d of themselves indicate
only a pause in the c u r r e n t t r e n d or a short-term reversal. They are n o t signals of a major trend reversal. It's possible that an exhaustion gap will appear
at a major turning point, b u t that is likely to be m o r e of a coincidence. O t h e r
tools, such as long-term m o m e n t u m , should be used for the identification
of a major t r e n d reversal.
Finally, exhaustion gaps are occasionally referred to as wide gaps. This is
because they tend to be wider than runaway gaps. In this sense, like heavy
or light volume, the term wide is a relative o n e . A "wide" bar can be recognized only in relation to previous chart action a n d is therefore a matter of
personal j u d g m e n t a n d experience.
Chart 12-18, for Apple, shows an extremely wide gap in early August. Note
the incredible expansion of volume. T h e next day the price experienced a
very small gap, which was immediately closed. After a long r u n like this, the
quick gap closing is the first sign of exhaustion. Within a couple of sessions
the wide gap below it was also closed, an even stronger sign of fatigue.
Chart 12-19, for Adobe, shows two types of gap. T h e rally begins with a
breakaway gap on good volume. An a t t e m p t is m a d e to close the gap, and
w h e n that fails, a confirmation develops as the price makes a new high. T h e
measuring implication for the runaway gap is r e a c h e d in a fairly precise
manner, b u t n o t before we see a n o t h e r gap. T h e tip-off that this was an
exhaustion gap came when the gap was quickly closed d u r i n g the next session. This particular gap-closing day was an outside bar, wThich is discussed
as a reversal p h e n o m e n o n in Chapter 13.
A second exhaustion gap develops at the extreme right-hand part of the
chart. Note the confirmation as the price breaks above a small base and
closes the complete gap within a few sessions.

Smaller Patterns and Gaps

Chart 12-18 Apple, daily.

Chart 12-19 Adobe, 1997-1998, daily.

233

232

Part III: Short-Term Patterns

other, thereby increasing the probabilities of a near-term reversal. In any
event, an exhaustion gap should n o t be r e g a r d e d as a sign of a major reversal, b u t merely as an indication that, at the very least, some form of consolidation should be expected.
If you are m o n i t o r i n g an oscillator that is overbought (or oversold, in the
case of a decline) or has experienced a divergence or two with the price,
this indicates an overextended price trend. Such a combination is m o r e
likely to be associated with an exhaustion gap than an oscillator that is at a
neutral reading.
Edwards a n d Magee p o i n t out that runaway gaps are usually left o p e n for
some time a n d typically require an intermediate or even primary t r e n d in
the opposite direction to the gap before they are closed. On the other h a n d ,
since an exhaustion gap is indicative of the terminal phase of a trend, it is
usually closed within a few sessions. Such action provides strong evidence
that the gap is n o t of the measuring or continuation variety. Even then, however, it should be appreciated that such gaps in a n d of themselves indicate
only a pause in the c u r r e n t t r e n d or a short-term reversal. They are n o t signals of a major trend reversal. It's possible that an exhaustion gap will appear
at a major turning point, but that is likely to be m o r e of a coincidence. Other
tools, such as long-term m o m e n t u m , should be used for the identification
of a major t r e n d reversal.
Finally, exhaustion gaps are occasionally referred to as wide gaps. This is
because they t e n d to be wider than runaway gaps. In this sense, like heavy
or light volume, the term wide is a relative o n e . A "wide" bar can be recognized only in relation to previous chart action a n d is therefore a matter of
personal j u d g m e n t a n d experience.
Chart 12-18, for Apple, shows an extremely wide gap in early August. Note
the incredible expansion of volume. T h e next day the price experienced a
very small gap, which was immediately closed. After a long r u n like this, the
quick gap closing is the first sign of exhaustion. Within a couple of sessions
the wide gap below it was also closed, an even stronger sign of fatigue.
Chart 12-19, for Adobe, shows two types of gap. T h e rally begins with a
breakaway gap on good volume. An a t t e m p t is m a d e to close the gap, and
when that fails, a confirmation develops as the price makes a new high. T h e
measuring implication for the runaway gap is r e a c h e d in a fairly precise
manner, b u t n o t before we see a n o t h e r gap. T h e tip-off that this was an
exhaustion gap came when the gap was quickly closed d u r i n g the next session. This particular gap-closing day was an outside bar, which is discussed
as a reversal p h e n o m e n o n in Chapter 13.
A second exhaustion gap develops at the extreme right-hand part of the
chart. Note the confirmation as the price breaks above a small base and
closes the complete gap within a few sessions.

Smaller Patterns and Gaps

Chart 12-18 Apple, daily.

Chart 12-19 Adobe, 1997-1998, daily.

233

234

Part III: Short-Term Patterns

Intraday G a p s
T h e r e are really two types of opening gap in intraday charts. T h e first develops as prices o p e n beyond the trading parameters of the previous session,
as in Chart 12-20. I'll call these classic gaps, since these are the ones that also
a p p e a r on the daily charts.
T h e second, m o r e c o m m o n type of gap develops only on intraday charts
as the o p e n i n g price of a new day gaps well away from the previous session's
closing bar. I'll call these gaps intrabar gaps because they fall between two
bars calculated on an intraday time frame. For example, in Chart 12-21, the
price o p e n e d up higher a n d created a gap. However, if you look back, you
will see that the trading range of the previous day (contained within the box
on the left) was not exceeded at the o p e n i n g price, a n d thus on a daily chart
t h e r e would have b e e n no gap.
If you are a trader with a two- to three-week time horizon who is using
intraday charts, you should a p p r o a c h gaps in a different way from s o m e o n e
with a one- or two-day time horizon.
People in the first category should try to avoid initiating trades at the time
the gap is created. This is because almost all gaps are eventually closed.
Sometimes this h a p p e n s within a couple of hours, a n d at other times it can
take two or three weeks. Consequently, if you buy on an o p e n i n g gap on the
upside, as in Chart 12-21, you r u n the risk that the gap will soon be closed.

Chart 12-20 March 1997 bonds 15-minute bar. (Source: pring.com.)

Smaller Patterns and Gaps

235

Chart 12-21 March 1997 bonds 15-minute bar. (Source: pring.

T h e p r o b l e m is, you do n o t know whether it will be closed in two days or
four weeks.
Intraday traders are also advised to step aside when the market opens
sharply h i g h e r or lower. In the case of stocks, this is caused by an o r d e r
imbalance. T h a t m e a n s that the market makers are forced to go short so
that they can satisfy the unfilled d e m a n d . They naturally try to get the price
a little higher at the o p e n i n g so that it will c o m e down a little, enabling them
to cover all or part of the short position. T h e process will be reversed in the
case of a lower opening. T h e key, then, is to watch what h a p p e n s to the price
after the o p e n i n g range. Normally, if prices work their way h i g h e r after an
upside gap a n d o p e n i n g trading range, this sets the tone of the market for
at least the next few hours, a n d often longer.
On the o t h e r h a n d , if the price starts to close the gap after a few bars,
then the t o n e becomes a negative o n e . In Chart 12-22, featuring Merrill
Lynch, t h e r e is an o p e n i n g gap on the Wednesday. After a bit of backing
a n d filling, the price gradually worked its way lower t h r o u g h o u t the day. T h e
signal that the o p e n i n g could be an aberration developed after the price
slipped below the trendline. Note how the trendline b e c a m e resistance for
the rest of the session. Thursday saw a n o t h e r o p e n i n g gap, but this time
there was very little in the way of a trading range, since the price c o n t i n u e d
to climb. Again the rally away from the o p e n i n g bar set the t o n e for the rest
of the day. On Friday a n o t h e r gap appears, but this time the o p e n i n g trading

236

Part III: Short-Term Patterns

Smaller Patterns and Gaps

237

Chart 12-22 Merrill Lynch 7.5-minute bar. (Source: pring.com.)

Figure 12-8 Island reversal.

O n e - D a y Island

range is resolved on the downside as the price breaks below the $86 level.
Once again this proves to be resistance for the rest of the day.

Island Reversals
An island reversal is a c o m p a c t trading range created at the e n d of a sustained move a n d isolated from previous price behavior by an exhaustion gap
and a breakaway gap. A typical island reversal is shown in Fig. 12-8. However,
it should be stated at the outset that islands are a pretty rare p h e n o m e n o n
in the charts a n d are in a n d of themselves short-term p h e n o m e n a . They can
appear, t h o u g h , at the e n d of an intermediate or even a major move a n d
form part of an overall price p a t t e r n , such as t h e top (or b o t t o m ) of a
head-and-shoulders formation (or an inverse head-and-shoulders p a t t e r n ) .
Islands occasionally occur as one-day p h e n o m e n a .
Chart 12-23, featuring Apple Computer, shows an island bottom. However,
since it developed within the confines of a bear market, the island failed.
Initial confirmation of this developed when the second island was completed
with a downside gap right at the e n d of April. Final confirmation occurred
when the price broke below the dashed trendline, which was in fact the
neckline of a consolidation h e a d a n d shoulders. In this case, the second
island r e p r e s e n t e d the bulk of the head. This emphasizes that a large number of islands are really patterns within patterns.

A one-day or one-bar island has the same characteristics as the island reversal j u s t described. However, this particular pattern, as its n a m e implies, consists of only o n e isolated a n d lonely bar.
A one-day island for Altera is featured in Chart 12-24. Note the very large
volume on the day of the island a n d the fact that the price closed very n e a r
the low for the day. Also, volume on the second day was even heavier. An
expansion of volume on the downside is usually bearish, b u t such an expansion the day after a one-day island was the kiss of death.

Chart 12-23 Apple, 1996-1997, daily.

238
Chart 12-24 Altera, daily.

Part III: Short-Term Patterns

Smaller Patterns and Gaps

239

Finally, Chart 12-25, for Yahoo, shows a small, bearish island top. Later
on we see a runaway gap followed by a one-day island. You can see how the
gap day closed on its h i g h a n d was a c c o m p a n i e d by exceptionally heavy volu m e . T h e day following the one-day island experienced even greater volu m e . T h e e x t r e m e right-hand part of the chart shows a two-bar reversal
followed by a breakaway gap to the downside. (For an explanation of twobar reversals, please see C h a p t e r 16.)

Summary
Flags Quick Review
• Price characteristics: A small trading range b o u n d e d by two parallel lines that
slope in the opposite direction to the prevailing trend.
• Volume configuration: Volume shrinks as the p a t t e r n is being formed a n d
expands in the case of an upside breakout. Expansion in volume is n o t a
r e q u i r e m e n t for a downside break b u t will e n h a n c e the validity of such a
break.
• Type: Appears as a continuation pattern.
• Measuring implication: T h e flag often appears halfway up or down a price
trend.
Chart 12-25 Yahoo, daily.
Pennants Quick Review
• Price characteristics: A small trading range b o u n d e d by converging trendlines sloping in the same direction, which should be opposite to the direction of the main trend.
• Volume configuration: Volume shrinks as the pattern is being formed a n d
expands in the case of an upside breakout. Expansion in volume is n o t a
r e q u i r e m e n t for a downside break b u t will e n h a n c e the validity of such a
break.
• Type: Appears as a continuation pattern.
• Measuring implication: T h e p e n n a n t often appears halfway up or down a
price trend.

W e d g e s Quick Review
• Price characteristics: A small trading range b o u n d e d by two converging lines
that slope in the opposite direction to the prevailing trend. Wedges differ

240

Part III: Short-Term Patterns

from p e n n a n t s in that they usually take longer to form a n d their theoretical apex is a long way from the b r e a k o u t point.
• Volume configuration: Volume shrinks as the pattern is being formed a n d
expands in the case of an upside breakout. Expansion in volume is n o t a
r e q u i r e m e n t for a downside break b u t will e n h a n c e the validity of such a
break. Breakouts are often followed by very sharp moves.
• Type: Appears as a continuation pattern.
• Measuring implication: T h e m a x i m u m distance between the two converging lines is projected in the direction of the breakout.

13

Outside Bars

Gaps Quick Review
• Price characteristics: A vacuum or hole in the chart that develops because
of high emotion. Gaps are usually closed a n d represent potential support
a n d resistance areas.
• Volume configuration: High volume on a gap increases its potential significance.
• Type: Breakaway gaps develop at the b e g i n n i n g of a move, runaway gaps
in the middle of a move, a n d exhaustion gaps at the e n d of a move. Island
reversals are small price patterns (or o n e bar, in the case of a one-bar
island) that are isolated from the main price trend by two gaps. They often
signal the termination of an intermediate move. Ex-dividend a n d area
gaps have little significance.

Introduction to One- and Two-Bar
Patterns
T h e price patterns covered in earlier chapters take some time to complete—
usually at least 15 bars. They all reflect changes in the relationship between
buyers a n d sellers, indicating that there has b e e n a change in the underlying sentiment.
This a n d subsequent chapters describe m u c h smaller patterns that are
confined to o n e or two bars. These formations can be extremely useful
because they trigger signals at a relatively early stage in the d e v e l o p m e n t of
a new t r e n d a n d usually offer very practical b e n c h m a r k s , above or below
which it is possible to place low-risk stops.
Historically, the patterns described in Chapters 13 to 16 have b e e n called
one- a n d two-day patterns or one- and two-week patterns. With the advent of
intraday charts, these titles no longer encompass the majority of these patterns. Therefore we will use the term bar rather than day to describe them.
A key factor d e t e r m i n i n g the significance of any pattern is the pattern's
size. These one- a n d two-bar patterns do n o t take very long to form, so they
only have a short-term influence on prices. For example, a pattern consisting of a one-day bar would be expected, u n d e r n o r m a l circumstances, to
affect the price over, say, a 5- to 10-day period. A two-bar pattern created
from 10-minute bars would influence the t r e n d over the course of the next
50 minutes to an h o u r or so. Even t h o u g h this is a relatively short period,
the m o r e I study these patterns, the m o r e impressed I b e c o m e with their
ability to signal short-term t r e n d reversals reliably.
When a trend has reached maturity and the long-term technical picture is
consistent with a turn, these one- and two-bar patterns often develop literally

241

242

Part III: Short-Term Patterns

at the final turning point of the move. In a sense, the one- and two-bar patterns
become dominoes or reverse dominoes as they become the final straw that tips
the long-term technical balance in a new direction; but more on that later.

Not All Patterns Are Created
Equal
We tend to think of the signals from these patterns as being either buy or
sell indications, a n d that is fine as far as it goes. However, it's equally important that we attempt to j u d g e their quality, since n o t all patterns are created
equal. This m e a n s that we should consider these signals in terms of stars or
shades of gray. For example, p a t t e r n A might contain the characteristics of
a specific formation, b u t pattern B could reflect t h e m in a stronger way. Thus
A could be a two-star pattern, b u t B might be classified as five-star. Technical
analysis deals in probabilities, so a five-star signal would offer h i g h e r odds
of a valid trend reversal than, say, a two-star signal, a n d so forth. Just as there
are no guarantees in technical analysis, t h e r e are no guarantees that a fivestar signal will always be valid, or that a one-star signal will be followed by a
weak move; it's merely that the odds favor this.
W h a t we are doing is h u n t i n g for clues to the strength a n d significance
of a change in sentiment that is being signaled by a particular reversal phen o m e n o n . I could say the word help, for example, b u t if I shout it from the
rooftop, you will get the message that I n e e d h e l p far m o r e clearly. T h e same
principle operates in the marketplace. Consequently, one- a n d two-bar patterns should be i n t e r p r e t e d as shades of gray rather than as black or white
because some patterns offer stronger signs of exhaustion than others.

General Interpretive Principles
T h e r e are several interpretive g r o u n d rules that apply to all these one- and
two-bar formations:
1. In o r d e r for these formations to be effective, there must be something
for t h e m to reverse. This m e a n s that top reversals should be p r e c e d e d
by a meaningful rally, a n d b o t t o m formations should be p r e c e d e d by a
sharp sell-off. As a general rule, the stronger the p r e c e d i n g trend, the
m o r e powerful the effect of the one- a n d two-bar price pattern.
2. These formations generally reflect an exhaustion point. In the case of an
u p t r e n d , such patterns develop when buyers have temporarily pushed
prices up too far a n d n e e d a rest. In the case of a downtrend, there is little, if any, supply, because sellers have completed their liquidation. Such

Outside Bars

243

patterns are almost always associated with a reversal in the prevailing
trend.
3. Not all patterns are created equal. Some show all of the characteristics
described later in a very strong way. Others reflectjust a few of these characteristics in a mild way. What we might call a five-star p a t t e r n , with all
the characteristics, is m o r e likely to result in a strong reversal than, say,
a two-star p a t t e r n that has mild characteristics. It is therefore necessary
to apply a certain degree of c o m m o n sense to interpretation of these patterns, r a t h e r than j u m p i n g to the immediate conclusion t h a t the presence of o n e of t h e m guarantees a quick, profitable price reversal.
4. Occasionally, it is possible to observe some form of confirmation closely
following or even d u r i n g the d e v e l o p m e n t of a one- or two-bar pattern.
Examples would include the completion of a larger p a t t e r n or, m o r e
likely, the violation of a trendline. Such events n o t only a d d significance
to the p a t t e r n but also increase the odds on this being a valid signal.

One- and Two-Bar Western
Patterns Compared to Japanese
Candlesticks
Before we move on to a m o r e specific study of the individual patterns, it is
i m p o r t a n t to u n d e r s t a n d that many of t h e m bear a close resemblance to
Japanese candlesticks. I say "close" because candlestick interpretation places
great emphasis on the o p e n i n g a n d closing prices. Many J a p a n e s e patterns
are d e t e r m i n e d by these two points relative to the o p e n i n g a n d closing
prices of the previous candle. O p e n i n g a n d closing prices are also important in bar chart interpretation, but so is the entire trading range, since it
captures the complete emotional m a k e u p of the bar. A comparison between
the two m e t h o d s of charting, where appropriate, will be m a d e in this a n d
the following t h r e e chapters. First, t h o u g h , a very quick heads-up on candles. For a review of the subject in m o r e d e p t h , please see my Introduction
to Candlestick Charting (workbook a n d CD-ROM tutorial) or the classic work
on the subject, Steve Nison's Introduction to Candlesticks.

Overview of Candlesticks
An individual candlestick consists of a rectangle with two p r o t r u d i n g lines,
o n e above the rectangle a n d the o t h e r below it. T h e rectangle is called the
real body a n d reflects the o p e n i n g a n d closing prices. T h e two vertical lines
are called wicks a n d represent the trading between the o p e n / c l o s e and the

244

Part III: Short-Term Patterns

Outside Bars

245

Chart 13-2 A white candlestick.
Chart 13-1 A black candlestick.

high a n d low. Candlesticks come in two colors, black a n d white. T h e color
that is plotted is d e t e r m i n e d by the relationship between the o p e n a n d the
close. If the o p e n is higher than the close, the rectangle is colored black, as
in Chart 13-1. If the close is higher than the o p e n , the real body is colored
white, as shown in Chart 13-2. In a very general sense, white candles are considered bullish a n d black ones bearish. T h a t is a very simplistic statement,
however, a n d needs to be qualified by the circumstances in which they
develop. Since candlesticks a n d bar charts display identical information,
there is n o t h i n g in a candlestick pattern that c a n n o t be found in a bar chart,
a n d vice versa. T h e principal difference is that candlesticks exhibit certain
charting p h e n o m e n a in a simpler way than bars do. They also place greater
emphasis on the o p e n i n g a n d closing levels. So, you may ask, why use bars?
T h e answer is that several other p h e n o m e n a can be recognized m o r e easily
on a bar chart. Also, because bars are thinner, m o r e data can be readably
displayed on a chart. Because psychology is the prime d e t e r m i n a n t of prices
in any freely traded market, highs a n d lows n e e d to be taken into account.
They are not ignored in candles, b u t they are easier to spot in the bars.

represent exhaustion. An example of a top reversal is shown in Fig. 13-1 a n d
an example of a b o t t o m in Fig. 13-2.

D e t e r m i n i n g Their S i g n i f i c a n c e
T h e guidelines for deciding on the potential significance of an outside bar
are as follows:

Outside Bars
An outside bar is a bar whose trading r a n g e totally encompasses that of its
predecessor. These patterns develop after b o t h down- a n d u p t r e n d s a n d

Figure 13-1 Outside bar at a top.

Part III: Short-Term Patterns

246

Outside Bars

247

period's low. T h u s a subtle c h a n g e in the strength or inclination of buyers a n d sellers has taken place. No longer are buyers willing or able to
push prices higher.

Figure 1 3 - 2

Outside bar at a bottom.

1. The wider the outside bar relative to the preceding ones, the stronger the signal.
This arises from the fact that the outside bar is supposed to reflect a
c h a n g e in the balance between buyers a n d sellers. In the u p p e r part of
Fig. 13-3, the buyers have h a d t h e u p p e r h a n d until the p e r i o d of the
outside bar. They c o m e into the bar expecting m o r e of the same, but
by the close sellers have m a n a g e d to take the price below the previous

Figure 1 3 - 3

Measuring the significance of outside bars.

2. The sharper the rally (reaction) preceding the outside bar, the more significant the
bar. It stands to reason that if a change in sentiment is going to take place,
t h e r e must be s o m e t h i n g to change. Therefore, the stronger the trend
p r e c e d i n g the outside bar, the stronger the implied sentiment dominating that trend. W h e n opinions move to an extreme, they are very vulnerable to change. In o t h e r words, if the p r e c e d i n g t r e n d is strongly
positive, all the bullish arguments will be fully discounted. Thus, at some
point, buyers will b e c o m e exhausted a n d the market will be vulnerable
to profit taking. This subtle c h a n g e is reflected in the outside day. An
example is shown in the lower part of Fig. 13-3.
3. The more bars encompassed, the better the signal. In most situations, the outside bar will encompass only o n e o t h e r bar. However, when it encompasses several bars, the signal that the balance has shifted from buyers
to sellers at a top or from sellers to buyers at a bottom becomes that m u c h
stronger. T h e encompassed bars b e c o m e a small price pattern in themselves. An example is shown in Fig. 13-4.
4. The greater the volume accompanying the outside bar relative to previous bars, the
stronger the signal. This again goes back to the idea that the trend preceding this pattern has b e e n d o m i n a t e d by o n e side or the other. T h e bar
itself indicates that this domination is no longer in force. If volume is particularly heavy, then m o r e market participants are voicing their views.

Figure 13-4

Outside bar encompassing several bars.

Part III: Short-Term Patterns

248

Strength in numbers, as reflected by the increased activity, increases the
significance of the outside bar.
5. The nearer the price closes to the extreme point of the bar that is away from the
direction of the previous trend, the better. For example, if the previous t r e n d
was down a n d the price closes very n e a r the high, this is m o r e favorable
than if it closes n e a r the low, a n d vice versa. This arises because the outside bar is supposed to signal a reversal in sentiment a n d a c h a n g e in
trend. T h e fact that the closing in this example develops n e a r the high
merely emphasizes the strength of the buyers, thereby adding to the validity of the signal.
If the close develops n e a r t h e high in a rising t r e n d or n e a r t h e low in a
falling trend, t h e n the outside bar is n o t consistent with a change in psychology. In this case, the outside b a r will b e c o m e a consolidation, n o t a
reversal pattern. An example is shown in Fig. 13-5, where the outside bar
appears after a small consolidation, b u t the close takes place n e a r the high
of the bar. In this instance, the sellers tried h a r d to push the price lower
and extend the correction. By the e n d of the period, t h o u g h , the buyers
h a d taken over a n d were able to close t h e price very n e a r to its high.
W h e n considering outside bars or any of the o t h e r one- a n d two-bar price
patterns, it is i m p o r t a n t to ask yourself the question, "What is the price
action of this bar telling me a b o u t the underlying psychology?" Wide bars,
sharp p r e c e d i n g rallies or reactions, a n d high volume all suggest a change
in the previous t r e n d of sentiment.

Figure 13-5

Consolidation outside bar.

Outside Bars

249

Not all one- a n d two-bar p a t t e r n s are followed by a reversalin trend. Some,
for example may be followed by a change in trend, as prices consolidate after
an up or down move.

Marketplace Examples
Chart 13-3 features an example of an outside bar for Merrill Lynch. It's really
a five-star signal, since it has pretty well all of the characteristics of a strong
reversal.
T h e price was in a persistent d o w n t r e n d d u r i n g the afternoon of t h e
twenty-first. T h e n a strong bar, totally encompassing the trading range of
the bar that traced o u t the low point for the move, developed. This was a
pretty strong statement because the bar o p e n e d near its low a n d closed
almost at its high. Note also t h e very high volume that accompanied this formation.
One- a n d two-bar patterns are n o t always followed by an immediate reversal in price. I n d e e d , they a r e often subject to some retracing or waffling
action prior to the onset of a strong reversal. This type of action is, of course,
consistent with o t h e r price patterns. This chart shows an example of this, as
the price experiences a two-bar correction prior to taking off on the upside.
This can be a good thing if you are thinking of taking a long position.
Remember, the stop should be placed below support, which in this case is

Chart 13-3 Merrill Lynch 10-minute bar.

250

Part III: Short-Term Patterns

the extreme low of the outside bar. T h e r e t r a c e m e n t allows you to enter the
trade with a lower risk, b u t with the same potential reward. Generally speaking, though, I would prefer to see a small correction, for often the m o r e
g r o u n d that is retraced, the weaker the signal. Even so, we c a n n o t say that
the outside bar has b e e n cancelled until the price actually moves beyond its
extreme point, t h o u g h anything greater than a 50 percent retracement is
cause for concern.
Several outside bars are a p p a r e n t in C h a r t 13-4. Example A is a g o o d
o n e because it is p r e c e d e d by a relatively strong rally a n d is reasonably
wide. E x a m p l e B fails completely. T h e reasons lie in the fact that the close
is right on t h e low, t h e b a r is n o t m u c h larger t h a n the previous o n e , a n d
the bar is n o t p r e c e d e d by m u c h of a decline. While in a strict technical
sense this is an outside bar, it is n o t signaling m u c h in t h e way of a c h a n g e
in s e n t i m e n t .
Even t h o u g h the price rallied for a very short period, example C was not
really successful either. It's true that it was p r e c e d e d by a decline, was a reasonable size, a n d closed at its high. However, I purposely p u t in this example to show that even when a substantial n u m b e r of the requirements are
present, this does n o t guarantee the success of a pattern.
Chart 13-5 shows two m o r e examples of outside bars for the DJIA in March
of 2001. T h e first represents a reversal from a d o w n t r e n d to an u p t r e n d . It's
always nice when we can see a one- or two-bar pattern confirmed by additional

Chart 13-4 S&P Composite 5-minute bar.

Outside Bars

251

Chart 13-5 DJIA 60-minute bar.

evidence pointing to a t r e n d reversal. In this case, the price breaks out from
a small base j u s t after the outside bar. T h e second example on the chart,
which developed in April, marks the top of a very sharp rally. Note that the
outside bar is very wide, encompasses the three previous bars, a n d also violates a good up trendline. Add to this the fact that it also experienced a whipsaw break above the horizontal trendline a n d you have all the ingredients
of a dramatic c h a n g e in sentiment.
Chart 13-6 is a daily chart of the Nasdaq 100. T h e outside bar on the left
has a lot of the ingredients of a valid pattern. It is p r e c e d e d by a good rally,
a n d the bar itself is wide a n d encompasses several o t h e r bars. It also o p e n e d
in the direction of the then-prevailing trend a n d closed on its low. T h e question naturally arises as to why it failed. O n e explanation is that in really
strong up- or downtrends, o n e - a n d two-bar pattern reversals represent contra-trend signals, a n d contra-trend signals often result in whipsaws. Note that
in this case, the outside bar was the market's a t t e m p t to close a gap that h a d
o p e n e d up several trading days earlier. T h e lower part of the bar therefore
r e a c h e d a s u p p o r t area.
In addition, it's i m p o r t a n t to r e m e m b e r that we are dealing in probabilities, not certainties, in technical analysis. Failures can a n d do exist. That's
why it is always necessary to look over your shoulder a n d mentally rehearse
where you are going to get o u t should the low-probability losing scenario
develop. In this case, the stop would be placed above the u p p e r point of
the outside day.

252

Part III: Short-Term Patterns

Chart 13-6 Nasdaq 100, daily.

T h e second outside bar canceled out a bearish outside bar that h a d
formed two days earlier. Usually cancellations are followed by a pretty
strong move, a n d this example was no exception.
Chart 13-7, for cocoa, shows two outside bars. T h e first should have b e e n
expected to reverse to the upside, but a couple of factors got in the way. First,
the bar whipsawed above a down trendline. Second, t h e close was n e a r the
low for the week, instead of the high. While this was technically an outside
bar, a n d it was p r e c e d e d by a sharp downtrend, these o t h e r factors m o r e
than canceled out its bullish aspects.
T h e second outside bar developed at the low for the decline. H e r e we see
a reversal of the factors that h a d caused the earlier failure. T h e price closed
above the down trendline almost at the high for the week, for a classic
reversal.
C h a r t 13-8 features a n o t h e r weekly example for cocoa. This time the outside bar is the third in a series of three outside bars. T h e effect is that they
reinforce each other, for a very strong signal. In addition, notice that the
final outside bar also closes above the resistance trendline j o i n i n g several
lows. This indicates that the break was a whipsaw, which a d d e d even m o r e
icing to the bullish cake.
Finally, Chart 13-9 shows the very significant October 2002 bear market
low in the U.S. stock market, where an outside bar was accompanied by very
heavy volume.

Outside Bars

Chart 13-7 Cocoa, 1991-1992, weekly.

Chart 13-8 Cocoa, 1999-2001, weekly.

253

254

Part III: Short-Term Patterns

Chart 13-9 S&P Composite, 2002-2003, daily.

Outside Bars

Chart 13-10 DJIA 60-minute bar.

Using Outside Bars as Dominoes or Reverse Dominoes
Earlier in this chapter, it was established that outside bars, like o t h e r onea n d two-bar patterns, have an influence on prices for only a short period of
time. However, when the longer-term technical structure is finely balanced
a n d likely to turn, these patterns can act as a d o m i n o at m a r k e t tops or a
reverse d o m i n o at bottoms. They can therefore be used to anticipate a reversal in trend, which is of far greater significance than the move implied by
the pattern itself.
Chart 13-10, featuring a 60-minute bar for t h e DJIA, offers an example.
H e r e we can see an outside bar following a small rally. It encompasses three
previous bars, so we would expect to see a decline for p e r h a p s 5 to 10 bars.
In this case, t h o u g h , the decline is m u c h greater than that. T h e r e are two
reasons for this. First, a closer review of t h e action shows that the u p p e r e n d
of the bar was actually a whipsaw breakout. Whipsaws are typically followed
by sharp moves in the opposite direction to that expected from t h e breakout, as traders scramble to square their positions. Second, the lower part of
the bar succeeds in violating a seven-day up trendline, a long time on an
hourly chart. T h e outside bar was therefore in a perfect position to tip the
balance of t h e longer-term technical position.
O u r second "domino" example features a daily chart of the DJ Transports
in C h a r t 13-11. T h e s m o o t h e d RSI (a nine-day RSI s m o o t h e d with an

Chart 13-11 DJ Transports, daily.

255

256

Part III: Short-Term Patterns

Outside Bars

257

eight-day simple moving average) has started to flatten in early January, so
the question becomes whether it will roll over completely, thereby giving a
short-term sell signal. A clue is provided by the price, which experiences a
bearish outside bar around January 12. This was not an outstanding example, but it was sufficient to push the immediate picture into a negative mode
and turn the smoothed RSI into a bearish mode.

Outside Bars and Candles
There is no direct counterpart to outside bars in Japanese candlestick interpretation. A close neighbor is the engulfing pattern, featured in Figs. 13-6
and 13-7. In this situation, it is the open and closing prices that are the determining factors, because such patterns require the real body of the second
candle to "engulf or fully encompass that of its predecessor. The wicks are
totally ignored in this concept. If there were such a thing as the outside bar
for a candlestick, it would include the wicks as well as the real body. For
example, Fig. 13-6 does not reflect a Western outside bar, but Fig. 13-7 does.
The underlying psychological principles of the engulfing pattern are basically the same in that they indicate a change in the buyer/seller balance following a worthwhile trend. The two differences are, first, the Japanese use
the opening and closing prices for the signal, whereas Western bar charts
use the extreme points on a bar. Second, the engulfing candle should be
the opposite color from the candle it engulfs. For tops it would be black and
for bottoms, white.

Figure 13-6 Engulfing (top) pattern.

Figure 13-7 Engulfing (bottom) pattern.

A close second to the engulfing pattern is the dark cloud cover (at tops)
and the piercing white line (at bottoms). These are shown in Figs. 13-8 and
13-9. At market tops, the first candle should have a reasonably long white
real body. The second should have a black real body that has a higher opening, but that closes more than halfway down the first candle. At market bottoms, the exact opposite holds for a piercing white line, as in Fig. 13-9. Once
again the wicks (highs and lows) are totally ignored.

Figure 13-8 Dark cloud cover.

Part III: Short-Term Patterns

258

14
Inside Bars
Figure 13-9 Piercing white line.

T h u s there are some similarities to outside bars in J a p a n e s e candlesticks,
b u t there is no exact match.

Summary
Outside Bars Quick Review
. Price characteristics: A single bar whose trading range encompasses its predecessor.
• Requirement: Should be p r e c e d e d by a sharp up- or d o w n t r e n d .
. Factors enhancing its significance: T h e wider the better; the m o r e bars it
encompasses, the higher the volume, the stronger the p r e c e d i n g trend.

Inside bars are the o p p o s i t e of outside bars, since they form totally within
the trading r a n g e of t h e p r e c e d i n g bar. An outside bar indicates a strong
reversal in sentiment. An inside bar, on the o t h e r h a n d , reflects a closer balance between buyers a n d sellers following a sharp up or down move in which
o n e o r the o t h e r d o m i n a t e d .
In Fig. 14-1, we see a rally in which buyers are very m u c h in control.
During the period in w h i c h the inside bar is formed, buyers are unable to
push prices to a new h i g h . For the first time d u r i n g the rally phase, the buy
side seems to have lost s o m e m o m e n t u m . By the same token, sellers have
n o t taken over, since t h e y are u n a b l e to push the price below t h e previous
bar. In effect, the b a l a n c e between the two parties is very close, a n d neither
is in control. Such action is a subtle indication that the prevailing t r e n d is
about to change. Since inside bars do n o t reflect as obvious a reversal in sentiment as their outside c o u n t e r p a r t s , these patterns are often followed by a
small trading range, as o p p o s e d to a reversal in trend. This trading range
is characteristically followed by an actual t r e n d reversal. An example of an
inside bar at a market b o t t o m is shown in Fig. 14-2.

• Factor suggesting failure: W h e n the close is at the extreme of the bar in the
direction of the prevailing trend.
. Measuring implications: N o n e , b u t the pattern should have an influence on
prices for 5 to 10 bars.
. Japanese candlestick match: N o n e , b u t engulfing patterns, piercing white
lines a n d dark cloud cover have close similarities.

Guidelines for Determining
the Significance of an Inside Bar
T h e i m p o r t a n c e of an inside bar is d e t e r m i n e d by the following factors:
1. The sharper the trend preceding the pattern, the better. In this situation, the
strong t r e n d indicates the d o m i n a n c e of o n e side or the o t h e r a n d is the
setup for the reversal. If n e i t h e r side were in control, sentiment would
be indecisive, a n d t h e r e would be n o t h i n g to reverse. For example, during a sharp rally, m a n y traders have large p a p e r profits. W h e n they sense

259

260

Part III: Short-Term Patterns

Inside Bars

261

Figure 14-1 Inside bar at a top.
Figure 14-3 Smaller inside bar at a top.
a reversal, it's an easy decision to cash in those profits, rendering the price
vulnerable. T h e opposite would be true for short sellers following a
decline. If the inside bar is n o t p r e c e d e d by a worthwhile move, this motivation will be absent.
2. The zvider the first bar and its immediate predecessors in relation to previous bars
and the inside bar, the better. This is because the wide bar indicates that the
prevailing t r e n d is reaching a climax. An extensive trading range within
a bar indicates volatility, a n d volatility is often a precursor of a trend reversal. Also, at the m a t u r e stage of a strong trend, wide bars indicate the
probability of exhaustion on the part of those pushing the trend. W h e n

the wide bar is followed by a substantially narrower bar, this suspicion is
confirmed.
3. The smaller the inside bar relative to its predecessor, the more dramatic the change
in the buyer/seller balance, and therefore the stronger the signal. T h u s the inside
bar in Fig. 14-3 is a m u c h b e t t e r example than those in Figs. 14-1 a n d
14-2.
4. Volume on the inside bar should be noticeably smaller than that for the preceding
bar, since this bar indicates a m o r e balanced situation. If we see heavy volu m e accompanying the bars p r e c e d i n g the inside bar, this indicates great
enthusiasm on the part of those p u s h i n g the trend. T h e n , if volume
shrinks noticeably, it m e a n s that n e i t h e r party is in control. Two examples are shown in Fig. 14-4. T h e example on the left also shows a fourbar consolidation following the inside bar. T h e price subsequently moves
to the upside.

Marketplace Examples

Figure 14-2 Inside bar at a bottom.

Chart 14-1 shows two examples of inside bars. T h e first marks the e n d of
the sharp S e p t e m b e r - N o v e m b e r decline. Note the substantial width of the
first bar a n d the paltry r a n g e of the actual inside bar. During the formation
of the first bar, volume is very heavy a n d the price declines sharply. T h e r e
is no d o u b t h e r e that sentiment is strongly on the bearish side. T h e n , on

262

Part III: Short-Term Patterns

Inside Bars

263

Chart 14-2 S&P Composite 5-minute bars.

Figure 14-4 Inside bars with volume.

the second day, volume dries up appreciably a n d the trading range is dramatically r e d u c e d , indicating a fine balance between buyers a n d sellers.
While this m a r k e d the b o t t o m of the move, the next short-term trend was
essentially a sideways o n e . Quite often we find that with inside bars, there
is a change as opposed to a reversal in trend.

Chart 14-1 Oxford Industries, 2000-2001, daily.

T h e second inside p a t t e r n developed j u s t u n d e r halfway up the rally a n d
was followed by a sideways trading range. T h e move was t o p p e d off by a twobar reversal, which is discussed in C h a p t e r 16.
Chart 14-2 shows some m o r e inside bars, this time on a five-minute bar
chart for the S&P Composite. T h e first example (A) was followed by a pretty
good rally, although the quality of the signal was n o t that great, since there
was not m u c h of a contrast between the actual inside bar a n d its predecessor. Example B was an outright failure, since the price continued to advance.
This goes to show that when a strong t r e n d is underway, price patterns can
a n d do fail. I n d e e d , their very failure can often be a clue to the strength of
the trend. It is often a good idea to see w h e t h e r one- a n d two-bar patterns
can be confirmed by o t h e r evidence, such as a trendline violation. In this
case there was no such p e n e t r a t i o n .
Finally, example Cwas a classic. T h e two final bars e x p a n d e d considerably in size. T h e actual inside bar was relatively small. Note also how the
o p e n a n d close developed at almost identical prices, thereby confirming the
idea of a very fine balance between supply a n d d e m a n d .
Chart 14-3 shows an inside bar for weekly cocoa. Note how it was closely
followed by an outside bar. In this case, the inside bar indicated that there
was a fine balance between buyers a n d sellers, a n d the outside bar indicated
that buyers were now in control. Later on we see a n o t h e r inside bar, right
at the top of the move. This was followed by a small two- to three-week

264

Part III: Short-Term Patterns

Chart 14-3 Cocoa, 1991-1992, weekly.

decline, which was part of an overall trading range. Note how the power of
the inside bar was later canceled by an outside bar.
Chart 14-4 features the Australian Gold Share Index. An inside bar
appears right at the e n d of a sharp decline. I n d e e d , the index h a d b e e n
declining for four straight sessions prior to the inside bar. Since these were
all relatively wide bars, it indicated that the bears were very m u c h in control. T h e inside bar was therefore the first sign that this one-way street h a d
ended.
In Chart 14-5, for the Sydney All Ordinaries, an inside bar appears right
at the first July low. Strictly speaking, it was n o t an inside bar, since the low
was set at approximately t h e same place as that for the previous bar.
However, it definitely indicated that sellers were n o t strong e n o u g h to send
prices any lower. W h e n this h a p p e n e d for a second day a n d the index closed
on its high, m o r e proof was given. T h e top of this two-day rally was capped
by a n o t h e r "almost" inside day, a n d a four-day correction followed.
Finally, Chart 14-6 shows some m o r e inside days. In the November example, the highs for the wide bar a n d the inside bar were almost identical.
T h e r e was no d o u b t who was in control on the day of the wide bar. However,
since the buyers were unable to push prices higher the next day, it was evid e n t that the market h a d r e a c h e d a m o r e even state of balance. A bearish
inside day in O c t o b e r called the top of a small rally. However, after a quick
decline had taken place, what looked like a perfectly legitimate bullish inside

Inside Bars

265

Chart 14-4 Australian Gold Share Index, daily.

bar did n o t work. Based on the price action contained in this chart alone,
there is no g o o d e x p l a n a t i o n for this failure. That's why it is always important to look at o n e or two trends above that which is b e i n g traded. Such an
investigation m i g h t have revealed that the d o m i n a n t short-term trend (two
Chart 14-5 Sydney All Ordinary Index, daily.

266

Part III: Short-Term Patterns

Inside Bars

267

Chart 14-6 Eurodollar, daily.

Figure 14-5 Bullish harami.

to six weeks) was declining at this point a n d that the inside bar was a countercyclical signal. Even if that were n o t so, we n e e d to r e m e m b e r that technical analysis is far from perfect a n d that perfectly good rules can be a n d
are b r o k e n from time to time. I would also a d d that of all the o n e - a n d twobar price patterns, this o n e appears to be the least reliable. In this instance,
waiting for a trendline violation would have m a d e sense. Finally, a n o t h e r
inside bar develops a b o u t halfway up the O c t o b e r advance. O n c e again it
failed to trigger a consolidation or a reversal, probably because the shortterm t r e n d was strongly positive.

of at an extreme. Generally speaking, a "good" h a r a m i should consist of a
relatively wide real body followed by a very narrow real body. Unlike in the
engulfing p a t t e r n s , t h e color of t h e second candle is n o t i m p o r t a n t
although opposite colors are preferable.

Inside Bars and Japanese
Candlesticks
T h e inside bar corresponds to the h a r a m i in J a p a n e s e candlestick interpretation (see Figs. 14-5 a n d 14-6). However, it is different because the
H a r a m i ignores the high a n d low a n d considers only the o p e n a n d close.
This m e a n s that theoretically either or b o t h wicks of a h a r a m i could encompass the p r e c e d i n g candle. In situations where b o t h wicks encompassed the
entire previous candle, the h a r a m i would c o r r e s p o n d to a Western-style outside day. It would not be a particularly strong o n e , however, because the
o p e n a n d close would develop in the middle of the trading r a n g e instead

Figure 14-6 Bearish harami.

268

Part III: Short-Term Patterns

Summary
Inside Bars Quick Review

15

• Price characteristic: A persistent t r e n d that culminates with a small bar that
is totally encompassed by its substantially wider predecessor.
• Volume characteristics: Possibly heavy volume on the wide bar; lower activity on the inside bar.
• Significance increased: T h e sharper the i n c o m i n g trend, the wider the first
bar in relation to previous bars, the smaller the inside bar relative to the
outside bar, a n d the greater the volume accompanying the first bar.
• Implication: Inside bars are followed by either a 5- to 10-bar consolidation
or an actual reversal. It's nice to see confirmation, such as from a trendline break.

Key Reversal,
Exhaustion, and
Pinocchio Bars
Key Reversal Bars
A key reversal bar develops after a p r o l o n g e d rally or reaction a n d indicates
very strong exhaustion characteristics on the part of the d o m i n a t i n g party.
By the time the price experiences a really strong key reversal bar, the trend
should have accelerated, even taking on parabolic tendencies occasionally.

Basic Characteristics
T h e classic p a t t e r n has t h e following characteristics:
1. T h e price o p e n s strongly in the direction of the prevailing trend. Ideally,
the o p e n will show up as a gap on the chart.
2. T h e trading range is very wide relative to the p r e c e d i n g bars.
3. T h e price closes n e a r or below the previous close (or n e a r or above the
previous close in a d o w n t r e n d reversal).
4. Volume, if available, should be climactic.
Fi gures 15-1 a n d 15-2 show two examples of key reversals, o n e at a top
a n d the o t h e r at a b o t t o m .

269

270

Part III: Short-Term Patterns

Key Reversal, Exhaustion, and Pinocchio Bars

271

a n n o u n c e s earnings that are above expectations, this could cause the stock
price to gap up at the o p e n i n g . However, if, at the e n d of the day, the price
has r e t u r n e d to a level that is n e a r or j u s t below t h e previous close, the news
has obviously had no new net positive effect on the price. Since the key reversal day is p r e c e d e d by a strong u p t r e n d , this is really the last gasp of the buyers as weak, u n i n f o r m e d traders, a n d investors react to the "great" news. It
is also the give-up phase for short sellers, many of w h o m panic out of the
security in t h e frantic activity associated with the early phase of the bar's formation. After all, if g o o d news c a n n o t push prices higher, what will?

Figure 15-1 Key reversal bar at a top.

Psychological Underpinning
T h e strong o p e n i n g emphasizes the urgency with which t h e d o m i n a t i n g
party regards the situation. At tops, buyers reach a new high in their level
of greed, a n d at bottoms, sellers are at the h e i g h t of panic. T h e gap opening is often associated with unexpectedly good news at a top or bad news at
a bottom. T h e way a m a r k e t deals with pro-trend developments can be crucial in assessing the short-term technical picture. For example, if a company

Figure 15-2 Key reversal bar at a bottom.

At bottoms, the gap lower o p e n i n g panics the weak longs into liquidation,
while short sellers are fully confident as they initiate new positions. F u r t h e r
declines are such a foregone conclusion that the only questions r e m a i n i n g
are how low prices will go a n d when they will get there. W h e n the price
returns to a level n e a r the previous close, the careless shorts are locked in
at lower prices. Since long liquidation is pretty well complete, prices no
longer fall, a n d the shorts are forced to cover. This is a surprise to t h e m ,
because the b a d news associated with the gap lower o p e n i n g should have
p u s h e d prices lower. However, if bad news c a n n o t invoke further liquidation, what can?
T h e wide-bar aspect of the key reversal bar is i m p o r t a n t because it implies
volatility, a n d volatility is often associated with market t u r n i n g points. W h e n
prices o p e n strongly in t h e direction of the prevailing trend, experience a
wide trading range, a n d close n e a r to where they started, this m e a n s that a
lot of p e o p l e are locked into losing positions. T h e h i g h e r the volume, the
greater the n u m b e r of p e o p l e who are locked in. It also indicates that buyers in an u p t r e n d tried to p u s h the price h i g h e r but were able to do so only
temporarily. They gave it their very best, b u t they failed, because at t h e e n d
of the bar, prices were little c h a n g e d from the previous close. T h e opposite
would be true for sellers in a downtrend.
Since key reversals develop after a sharp, almost parabolic rally or reaction, they t e n d to reflect stronger emotions than the average one- or twobar reversal p a t t e r n s . F o r this reason, their effect is usually far m o r e
significant, often lasting far longer than the usual 5 to 10 bars.
Key reversal bars are often followed by sharp reversals in trend, but we
occasionally find that after several m o r e bars have developed, a r e t r a c e m e n t
move sets it. An example is shown in Fig. 15-3 for a top. This r e t r a c e m e n t
acts as a kind of test. If it fails, it becomes a reinforcement of t h e c h a n g e in
psychology that has taken place. T h e r e t r a c e m e n t also offers sellers at a top
a n o t h e r opportunity to get o u t a n d buyers at a b o t t o m a second chance to
get in.
Normally, the e x t r e m e level of the key reversal b a r is n o t exceeded, b u t
if it is, t h e n the pattern is canceled. Since this is an emotional p o i n t on t h e

272

Part III: Short-Term Patterns

Key Reversal, Exhaustion, and Pinocchio Bars

Chart 15-1 Barrick Gold, 1999-2000, daily.

Figure 15-3 Key reversal bar showing a retracement.

chart, it is an important potential support/resistance level, a n d therefore it
represents an intelligent place to set a stop. In the case of a top, this will be
slightly above the key reversal bar, and in the case of a bottom, just below it.
Chart 15-2 Merrill Lynch, 1998, daily.
Marketplace Examples
The left-hand part of Chart 15-1 shows a classic example of a key reversal for
Barrick Gold. Note how the short-term rally is climaxed by an explosion of volu m e and a wide key reversal bar. T h e opening price is very near the high. Quite
often a key reversal is followed by a sharp change in trend a n d a subsequent
retracement. That is exactly what h a p p e n e d in this case, as the price rallied in
the fourth and fifth sessions following the key reversal. T h e termination of this
brief advance was signaled by an outside day. These one-bar patterns have only
short-term significance as a general rule, but often they can prove to be the
first d o m i n o in a major trend reversal. It all depends on the maturity of the
trend in question combined with the position of the longer-term indicators.
In this instance, the key reversal marked the top for at least six months.
T h e second example of a key reversal, to the right, is also a good o n e in
that volume expands along with a fairly wide bar. However, it is not preceded
by m u c h of a rally. Thus the bullish expectations were n o w h e r e n e a r as
strong as on the late S e p t e m b e r reversal day. This pattern would certainly
not warrant as many stars as the first one.
Chart 15-2 features the 1998 b o t t o m for Merrill Lynch. T h e actual day of
the low was a classic case of a key reversal bar. T h e volume also cooperated

273

274

Part III: Short-Term Patterns

by e x p a n d i n g to a crescendo as the stock b o t t o m e d . It was even possible to
construct a small down trendline. W h e n this trendline was violated, it confirmed the signal being given by the reversal bar. T h e r e was no way of knowing from the bar itself that a b e a r market low in the stock h a d j u s t b e e n seen;
it merely indicated that a worthwhile rally was likely. T h e extremely heavy
volume would have told us that this particular key reversal would probably
reverse the t r e n d for a lot m o r e than the n o r m a l 5 to 10 days. On the other
h a n d , had it b e e n possible to establish that the overall market h a d b o t t o m e d
on the same day, it could legitimately have b e e n argued that the character
of this p a t t e r n a n d its accompanying volume signaled a bear m a r k e t b o t t o m
for the stock. If you look closely you can see that this particular b a r was also
an outside day. In fact, it could be a r g u e d that it bears m o r e resemblance
to an outside b a r than to a key reversal, since the price closed well above
the previous close. This is a semantic point, since we are really c o n c e r n e d
with identifying technical p h e n o m e n a that reflect significant changes in psychology. This bar certainly does, regardless of what it is termed.
Chart 15-3 shows that t h e July 2002 b o t t o m in the S&P was signaled by a
key reversal bar. T h e o p e n i n g price appears to be at the same level as the
previous day's close, but the chart does n o t indicate the initial weakness
experienced by many issues. This is because the o p e n i n g price does not
reflect stocks that have a mismatch of orders. Since the o p e n i n g was weak,
these stocks experienced a delayed opening, and the "opening" price for the

Key Reversal, Exhaustion, and Pinocchio Bars

275

S&P included only those stocks that were trading at the 9:30 a.m. opening.
T h e trading bar was extremely wide c o m p a r e d to anything else on the chart
and was preceded by a very sharp a n d persistent decline. T h e potential for a
reversal in bearish psychology was definitely present. Perhaps the key to this
situation was the record volume experienced during this session. Record volu m e by definition does n o t h a p p e n every day. W h e n it develops after a long
bear trend, it is typically a sign of a new bull market. In this case, the index
subsequendy m a d e a slightly lower low in October, b u t the July low saw m o r e
issues reach 52-week lows a n d was therefore a m o r e intense bottom.

Exhaustion Bars
Exhaustion bars develop after a really sharp up or down move. They are a form
of key reversal a n d are also a variation on the one-bar island reversal, described
in Chapter 12. However, they differ e n o u g h to warrant their own category.

Basic Characteristics
T h e r e q u i r e m e n t s for an exhaustion bar are as follows:
1. T h e price opens with a large gap in the direction of the then-prevailing
trend. T h a t t r e n d is typically quite strong.

Chart 1 5 - 3

NYSE Composite, 2002, daily.

2. T h e bar is extremely wide relative to previous bars. As a general rule, the
relative width of these bars is far greater than the average key reversal.
3. T h e o p e n i n g price develops in the lower half of the bar in a d o w n t r e n d
a n d in the u p p e r half in an u p t r e n d . In o t h e r words, sentiment is at or
very close to its most e x t r e m e at the start of the bar.
4. T h e closing price should be b o t h above the o p e n i n g a n d in the top half
of the b a r in a d o w n t r e n d a n d in the lower half a n d below the o p e n i n g
in an u p t r e n d . This offers a few pointers that sentiment, at least d u r i n g
the bar's formation, has b e g u n to reverse.
5. T h e bar is completed with a gap to the left still in place. This differs from
the key reversal bar, where there is either no gap or an extremely small one.
Examples of exhaustion bars for b o t h a b o t t o m a n d a top are featured
in Figs. 15-4 a n d 15-5. They differ from the one-bar island reversal (see
C h a p t e r 12) in that there is no gap between the exhaustion bar and its successor.
Also, it is n o t u n c o m m o n for t h e o p e n i n g a n d closing prices in a one-bar
island to be n e a r each o t h e r a n d fairly close to the gap.

Part III: Short-Term Patterns

276

Key Reversal, Exhaustion, and Pinocchio Bars

277

or less persistently lower prices, often reflected by wide bars. The large gap
as the price opens in what will turn out to be the lower part of the exhaustion bar reflects the culmination of this bearish psychology and the reality
that the sellers are very much in control. However, the fact that prices often
drop much further after the opening but then close in the upper half of
the bar offers a subtle hint that the selling may be over and that perhaps it
is the buyers' turn to take command. The wide high-volume trading range
also suggests the kind of frenzied activity associated with a turn.
At market tops, the same kind of psychology is also present, except that
greed replaces fear and the opening-closing relationship is reversed.

Marketplace Examples

Figure 1 5 - 4

Exhaustion bar at a bottom.

Underlying Psychology
When trying to identify an exhaustion bar, look for a very strong trend that
is capped with an extension that takes the form of an extreme movement
in the price. The idea that the bar opens with a huge gap and closes in the
opposite direction reflects the concept of a reversal in psychology. For example, at a market low, we come into the exhaustion bar following a sharp
decline. On a daily chart, this could take the form of 5 to 10 days of more

Figure 15-5

Exhaustion bar at a top.

Chart 15-4 shows an exhaustion bar on a daily chart for Kellwood. It was preceded by a short but sharp decline and became a reversal point for a small
rally in an ongoing downtrend. It has the correct price characteristics: gap
down, low below the opening of the bar, and close more than halfway up.
Generally speaking, it would have been better for the opening to have been
in the lower part of the bar, but the extreme width more than made up for
this missing ingredient.
Warnaco, in Chart 15-5, had the correct characteristics vis-a-vis the openclose relationship. In this instance, the downside gap would have been a

Chart 1 5 - 4

Kellwood, 2 0 0 0 - 2 0 0 1 , daily.

278

Part III: Short-Term Patterns

Key Reversal, Exhaustion, and Pinocchio Bars

279

Chart 15-5 Warnaco, daily.

Chart 15-6 NYSE Composite 5-minute bar.

pretty horrific experience for the longs. T h e price h a d closed at $24 the
night before a n d o p e n e d at $19 on t h e day of t h e exhaustion bar, for a $5

because they give us a false sense of what is really going on. Examples are
shown in Figs. 15-6 a n d 15-7.

gap!
In retrospect, t h e bar t u r n e d o u t to be the first b o t t o m of a double-bottom formation. Note that the second b o t t o m consisted of an inside bar,
which was confirmed when the price broke above the down trendline.
On intraday charts, gaps almost always develop at the o p e n because of
some overnight change in psychology. This means that exhaustion bars tend
to be m o r e prevalent in these very-short-term charts. Chart 15-6 shows an
exhaustion bar for the NYSE Composite. It has all the characteristics: large
gap, close higher than the opening, wide range, a n d so on. It is also followed
by an inside bar, which a d d e d a further piece of evidence that the trend had
changed.

Pinocchio Bars
Basic Characteristics
A pretty c o m m o n form of exhaustion shows up when a security opens a n d
closes relatively quietly, b u t d u r i n g t h e bar the price breaks temporarily
above a trading range, level of resistance, or trendline a n d t h e n falls back
below it. T h e reverse would be true at a b o t t o m . I call these Pinocchio bars,

Underlying Psychology
Pinocchio bars are bars in which t h e bulk of the trading takes place outside
the previous a n d subsequent trading range. This means that upside breakouts lock in unwary buyers with a loss at the closing of the bar. Short sellers will be similarly t r a p p e d at bottoms. Alternatively, these patterns may
develop because the pros on the floor sense that there are a substantial number of stops below the market. T h e security is t h e n pressured to squeeze out
these unwary holders, a n d the price is subsequently free to move higher.
During t h e Pinocchio period of trading, when the price is temporarily
above resistance in an u p t r e n d or below support in a downtrend, either buyers or sellers have a chance to push prices in their direction, but they fail.
For e x a m p l e , if the optimists see an upside breakout, they should be
attracted to the security, a n d the trend o u g h t to continue. However, by the
e n d of t h e bar, the price returns to below the resistance a n d they b e c o m e
discouraged.
T h e character Pinocchio tells us w h e n he is lying because his nose gets
longer. In the case of the Pinocchio bar, it is the trading beyond the resistance or s u p p o r t level in question that indicates that the signal is false. We

280

Part III: Short-Term Patterns

Figure 15-6 Pinocchio bar at a top.

can also take this a step further by saying that the bigger the nose, the bigger the lie. In the case of security prices, the larger the trading range above
(below at bottoms) the o p e n i n g a n d closing prices, the m o r e false the signal a n d t h e greater t h e move in the opposite direction is likely to be. To
p u t it a n o t h e r way, the wider the bar, the greater the n u m b e r of traders that
have the potential to be locked in at losing prices.
Figure 15-8 reflects the idea that when a false break develops above a down
trendline, this is also indicative of exhaustion, since the price c a n n o t hold

Figure 15-7 Pinocchio bar at a bottom.

Key Reversal, Exhaustion, and Pinocchio Bars

281

Figure 15-8 Pinocchio bar at a trendline.

above the strong resistance reflected by the line. T h e same principle would
hold true for a false downside break of an up trendline.

Marketplace Examples
Chart 15-7 shows a break above the trading range that was nullified by the
time the bar closed. T h e price moved sideways for a few bars, but, as often
h a p p e n s following false breakouts, it then moved to the downside in a decisive way. Chart 15-8 shows a false break to the downside. O n c e again this
was followed by a strong move in the opposite direction to the Pinocchio
"lie."
Chart 15-9 displays a five-minute bar for the S&P December 1998 contract.
In this case, the price breaks above the two previous bars' highs, but by the
time of the closing it is back below the resistance. This is n o t as clear-cut a
case of t e m p o r a r y resistance violation as Chart 15-7. Even so, it is surprising how many false breaks above resistance (below support) formed byjust
two or t h r e e bars result in a Pinocchio-type whipsaw.
O n e i m p o r t a n t fact a b o u t an exhaustion move is that the extremity of the
bar in question often proves to be an i m p o r t a n t s u p p o r t or resistance point.
Therefore, it is often a good idea to place a stop loss a little bit beyond the
extremity of the Pinocchio bar—provided, of course, that t h e trade still
results in a reasonable risk/reward. An example is shown in Charts 15-10
and 15-11, which feature five-minute bars for the S&P futures. Chart 15-10
indicates a Pinocchio break above the dashed trendline. If a trader h a d h a d

282

Part III: Short-Term Patterns

Chart 15-7

S&P Composite 10-minute bar.

Chart 1 5 - 8

S&P Composite 10-minute bar.

Key Reversal, Exhaustion, and Pinocchio Bars

Chart 1 5 - 9

Chart 1 5 - 1 0

S&P Composite 5-minute bar.

S&P Composite 5-minute bar.

283

284

Part III: Short-Term Patterns

Chart 15-11 S&P Composite 5-minute bar.

a reason for going short, a stop above t h e e m o t i o n a l point—i.e., the
extremity of the Pinocchio bar—would have m a d e sense. Chart 15-11 shows
the same situation, b u t in this chart the trading p e r i o d has b e e n e x t e n d e d
to show that the Pinocchio high was i n d e e d a great resistance area, since it
t u r n e d back two subsequent rallies.
Chart 15-12 features a daily graph for the l u m b e r price. A Pinocchio bar
develops in early April. It may n o t have b e e n that obvious at the time, b u t
as the chart progresses, you can see that it was possible to e x t e n d the line
j o i n i n g the two pre-Pinocchio highs. This emphasized the fact that a false
upside break h a d taken place. Consequently, it would have b e e n possible
to have sold short when a violation of the dashed up trendline confirmed
the Pinocchio bar three days after its formation. Note the consolidation outside bar that developed a couple of days after the trendline violation.
In Chart 15-13, featuring the daily crude oil price, an inside day helps the
price b o u n c e off its early April low. After a few m o r e days, it would have been
obvious that that low was a b o t t o m of some kind. Since previous lows a n d
highs are good candidates for s u p p o r t / r e s i s t a n c e zones, it would have b e e n
possible to construct a horizontal trendline at the low, indicating potential
support. In fact, this trendline was t o u c h e d or a p p r o a c h e d twice a r o u n d
J u n e 8,just prior to being temporarily b r e a c h e d as the price experienced a
Pinocchio day. After the Pinocchio, the line again acted as support. T h e n
an outside day formed a n d the price broke above a nice down trendline,
thereby confirming the Pinocchio and the outside bar. Finally, note the small

Key Reversal, Exhaustion, and Pinocchio Bars

Chart 15-12 Lumber, daily.

Chart 15-13 Crude oil, daily.

285

286

Part III: Short-Term Patterns

Key Reversal, Exhaustion, and Pinocchio Bars

287

Chart 15-14 Kellogg, 1997-1998, daily.

Figure 15-9 Doji candlesticks.

resistance in the case of a whipsaw bottom. This is because the idea of a false
r e s i s t a n c e / s u p p o r t violation is crucial to the Pinocchio characteristic, yet
n o n e of the candlesticks described h e r e involve such a concept.

Summary
Pinocchio that m a n a g e d to temporarily break above the down trendline in
mid-May. It was initially followed by a decline, as we might have expected.
However, the good news is that it e n h a n c e d the significance of the down
trendline, so when that trendline was finally violated, a terrific rally followed.
Finally, Chart 15-14, for Kellogg, shows a Pinocchio that temporarily
pushes above two horizontal trendlines. This bearish action was confirmed
the next day as the price violated a good up trendline on extremely heavy
volume. T h e e x p a n d e d activity emphasized that t h e sellers certainly h a d the
u p p e r h a n d at this point.

Key Reversals Quick Review
• Price characteristics: A wide bar following a strong and persistent trend. In
the case of a top, it gaps higher at the opening and may move higher during the remainder of the bar, but it closes near the previous close. At bottoms, it gaps lower at the opening and may move lower during the
remainder of the bar, but it closes near the previous close.
• Volume characteristics: A heavy-volume day that often climaxes a previous
trend of rapidly rising volume.
• Failure signaled: When the price moves beyond the extreme point of the
bar.

Key Reversals, Exhaustion Bars,
Pinocchios, and Candlesticks
T h e r e are no candlesticks that are equivalent to key reversals or exhaustion
bars. A key reversal would show up as a very long real body following an
advance or decline. Nor are there any direct correlations to Pinocchio bars.
T h e closest is probably a gravestone, dragonfly, or long-legged doji, as shown
in Fig. 15-9. A doji is a candle in which the o p e n a n d close are either identical or extremely close to each other. A Pinocchio could also take the form
of a h a m m e r or inverted h a m m e r . In the Pinocchio context, it would be
i m p o r t a n t for the u p p e r wick to temporarily break above resistance in t h e
case of a whipsaw upside breakout, or for t h e lower o n e to break below

• Measuring implications: None. Key reversals often develop at the end of
intermediate and primary trends.

Exhaustion Bars Quick Review
• Price characteristics: A very wide bar that develops after a strong t r e n d a n d
opens with a large gap At bottoms, the o p e n i n g is in the lower half of the
bar, a n d the closing is in the top half. At tops, the o p e n i n g is in the top
half of the bar, a n d the closing is in the lower half. T h e r e is no gap on
the right-hand side of t h e p a t t e r n unless the exhaustion bar is a one-bar
island reversal.

288

Part III: Short-Term Patterns

• Volume characteristics: Usually volume is heavy a n d reflects a temporary selling climax.
• Failure signaled: W h e n the price moves beyond the e x t r e m e of the bar.

16

• Measuring implications: N o n e .

Pinocchio Bars Quick Review
• Price characteristics: A bar that temporarily breaks a resistance or support
area between the o p e n i n g a n d t h e closing. T h e longer t h e "nose," the
stronger the signal.
• Volume characteristics: Can develop on low volume, but high volume reflects
m o r e t r a p p e d individuals a n d tends to be m o r e significant.

Two- and ThreeBar Reversals

• Significance: T h e greater the r e s i s t a n c e / s u p p o r t barrier that is p e n e t r a t e d
by the "nose" a n d the stronger the i n b o u n d trend, the m o r e the significance is e n h a n c e d . Also, the longer the nose, the greater the significance.
• Failure signaled: W h e n the price moves beyond the extreme point of the
bar. T h e e x t r e m e p o i n t often represents strong support or resistance.
• Measuring implications: N o n e , b u t stronger moves tend to develop when a
Pinocchio is confirmed within three or four bars by a trendline violation.

Two-Bar Reversals
Basic Characteristics
A two-bar reversal is a classic way in which the charts signal exhaustion. Like
o t h e r one- a n d two-bar patterns, two-bar reversals develop after a prolonged advance or decline, b u t the reversal appears to be stronger than the
average one- or two-bar formation. Examples are shown in Figs. 16-1 a n d
16-2. T h e characteristics of these patterns are as follows:
1. They n e e d to be p r e c e d e d by a strong a n d persistent trend. Since twobar reversals are often followed by dramatic reversals, it is m o r e important for t h e m to be p r e c e d e d by a strong rally or reaction than for most
of the o t h e r one- a n d two-bar formations.
2. Both bars should stand out as having exceptionally wide trading ranges
relative to previous bars (see Figs. 16-3 a n d 16-4). O n c e again, wide bars
reflect volatility a n d emotional exhaustion. T h e m o r e we can see of this
type of action, the m o r e confidence we can have in the power of the pattern to signal an i m p o r t a n t reversal.
3. T h e o p e n i n g a n d closing of both bars should be near the extreme points
of the bar. This is important, as it reflects the total a n d u n e x p e c t e d
change in sentiment that has taken place. Longs are locked in at two-bar
reversal tops a n d shorts at bottoms.
4. An expansion of volume on b o t h bars e n h a n c e s the c o n c e p t of a change
in sentiment. High volume after a strong t r e n d is also indicative of buyer
289

290

Part III: Short-Term Patterns

Two- and Three-Bar Reversals

Figure 1 6 - 3

Two-bar reversals as rectangles.

Figure 16-1 Two-bar reversal at a top.
or seller exhaustion, d e p e n d i n g on the direction of the p r e c e d i n g trend.
It also m e a n s that m o r e traders are locked in with a loss at the pattern's
completion.
5. A lot of the time, two-bar reversals are followed by an immediate sharp
m o v e , with the price reversing on a dime. However, there a p p e a r to be
e q u a l l y as many occasions when a small r e t r a c e m e n t move develops prior
to the real trend emerging. Examples are shown in Figs. 16-5 a n d 16-6.
T h e retracements in these examples have b e e n t e r m e d resistance zones

F i g u r e 16-2

Two-bar reversal at a bottom.

for tops a n d support zones for bottoms. T h e retracements can be as little
as o n e bar or as many as three. However, the longer this process takes,
the weaker the pattern is likely to be. I do not have any h a r d a n d fast
rules as to the degree of r e t r a c e m e n t that is allowable, b u t 50 p e r c e n t
appears to be a good starting point. As a rule of t h u m b , I would say that
the smaller the retracement, the greater the ensuing decline (or advance
in the case of a two-bar reversal b o t t o m ) is likely to be, a n d the higher
the probability that the pattern will work. After all, if t h e r e has b e e n a

Figure 1 6 - 4

Two-bar reversals as rectangles.

292

Part III: Short-Term Patterns

Two- and Three-Bar Reversals

293

Since the protective sistop loss should be placed just beyond the extreme point
of the formation, a r e t r a c e m e n t also r e d u c e s the risk associated with the trade.

Underlying Psychcology

Figure 16-5 Two-bar reversals showing a resistance zone.

dramatic change in sentiment, we do n o t want the price to act indecisively,
since, o t h e r things being equal, indecision will indicate a weak signal.
On the other hand, when resistance a n d support zones develop after the
completion of a two-bar reversal, they offer good places above or below which
relatively low-risk stops can be placed. T h e retracements also offer the opportunity to enter a trade at a price closer to the extreme of the two-bar reversal.

The first bar of the f o r m a t i o n develops strongly in the direction of the thenprevailing trend. F o r a five-star signal in an uptrend, we need to see the
close of the bar at or very near the high. The bar should also be relatively
wide. At the opening g of the next period, buyers come in expecting more
of the same. This m e a n s that the price should open very close to the high
of the previous bar. However, the underlying concept of the two-bar reversal is that a change in psychology takes place and the second bar closes
slightly above or b e l o w the low of the first bar. Hence, the high expectations of participants at the opening of the bar are totally dashed by the end
of the period, i n d i c a t i n g a change in sentiment. People who bought at the
upper end of the t r a d i n g range of the two bars are locked into the position at higher prices s and are therefore a potential source of supply on the
way down.
The two-bar reversal at market bottoms works exactly the same way, but
in reverse. An e x a m p l e is shown in Fig. 16-2. In this case, the pattern should
be preceded by a p e r s i s t e n t decline, with the sellers remaining very much
in control. The first bar opens at or near its high and closes near its low as
this pessimistic t r e n d is maintained. However, although the second bar
opens close to the l o w of the first, it ends on a high note, indicating that it
is the buyers who n o w have the upper hand. Once again, it is important for
the two bars to be substantially wider than those preceding the pattern. If
they are a c c o m p a n i e d by very high volume, so much the better. The width
adds to the flavor of f the formation because wide bars imply volatility, and
volatility is associated with reversals. People tend to reach their maximum
level of anger right at the end of a row. Markets also tend to reach an
extreme in emotion, indicated at the end of a trend with wide bars and lots
of volume.
To be really effective, two-bar reversals should reflect a climactic experience. A specific p a t t e r n should therefore contain as many of the elements
described here as possible.

Marketplace E x a m p l e s

Figure 16-6 Two-bar reversal showing a support zone.

Charts 16-1 and 16-2 feature a five-minute bar for the S&P Composite. Chart
16-1 shows a classic pattern, with the highs and lows falling at exactly the
correct points following a good rally. Chart 16-2 demonstrates a small

294
Chart 16-1 S&P Composite 5-minute bar.

Chart 16-2 S&P Composite 5-minute bar.

Part III: Short-Term Patterns

Two- and Three-Bar Reversals

295

Chart 16-3 June 2001 gold, daily.

retracement move and violation of the dashed up trendline. Often we do
not get confirmation from one- and two-bar patterns, but in this case the
downside break indicated that the index had completed a small top, thereby
adding further significance to the two-bar reversal top.
Chart 16-3 shows a two-bar reversal for June gold in February of 2001. In
many cases, such patterns are followed by an immediate advance. In this
case, the advance was delayed a day, as an additional piece of evidence indicated a trend reversal. This came in the form of an inside bar that developed on the day after the two-bar reversal. Such double patterns are often
quite effective in signaling changes in trend. In this case, the pattern would
probably have been stronger without the inside bar. This is because inside
bars indicate a balance between buyers and sellers, and the two-bar reversal ideally indicates a dramatic reversal between them. By the same token,
it is possible to argue that the previous weeks had been dominated by sellers and that the inside bar offered another successful test of the idea that
the sellers were no longer in control. In this instance, the inside bar was followed by an outside bar (not labeled), so the situation was resolved in favor
of the bulls. Note also that the second bar in the two-bar reversal encompasses the first bar. In effect, this is an outside bar. The two-bar reversal pattern rule does not require an outside bar, but the very presence of such a
bar reinforces the idea that sentiment has definitely reversed. If we are looking at the quality of a signal, I would certainly add points for this, together
with the outside bar that followed the inside bar.

Part III: Short-Term Patterns

296

Chart 1 6 - 4

US Bancorp, daily.

Chart 16-4 shows a two-bar reversal at t h e climax of a rally in US Bancorp
in the fall of 2000. Note the strong expansion of volume, together with the
fact that volume on the second day was slightly higher than that on the first.
This provided an additional clue that the tide h a d t u r n e d in favor of the
sellers.
Sugar offers a very emotional two-bar reversal in Chart 16-5. See how the
first day gaps u p . It doesn't quite o p e n on the low, b u t the two bars definitely have the right flavor. T h e second bar opens right on its high a n d closes
at its low. T h r e e days later, it is a p p a r e n t that the two-bar reversal is part of
an island reversal. (For a full explanation of island reversals, please see
Chapter 12.)
A false upside breakout develops in Chart 16-6, for the Dollar/Swiss franc.
This adds to the significance of the two-bar reversal, since those who were
buying for the bulk of the two days were locked in above the breakout point.
This is a slightly different p a t t e r n from the classic o n e that develops after a
sharp rally, b u t the false b r e a k o u t a n d fact that the second bar of the twobar reversal was an outside bar certainly make up for what the rally is lacking. This pattern was also confirmed as the price broke below the dashed
support line.
Chart 16-7, for weekly live cattle, shows a n o t h e r false two-bar reversal
breakout. In fact, it shows three. T h e first, on the left (A), has the characteristics of a two-bar reversal without m u c h of a rally p r e c e d i n g it. Also, the
bars were n o t particularly wide. It was by no means a classic pattern, b u t it

Two- and Three-Bar Reversals

Chart 1 6 - 5

Sugar, daily.

Chart 16-6

Dollar/franc, daily.

297

298

Part III: Short-Term Patterns

Chart 16-7 Live cattle, 1993-1994, weekly.

does indicate that when breakouts turn out to be whipsaws a n d the two-bar
reversal forms part of the breakout, t h e n we should expect to see at least a
temporary reversal. Later on (B), we see a bullish two-bar reversal bottom.
O n c e again it is n o t p r e c e d e d by m u c h of a d o w n t r e n d , so it would n o t have
b e e n given very many stars. However, it did result in a nice rally. Finally, there
is a n o t h e r whipsaw (C), again characterized by two bars with a two-bar reversal flavor. T h e whipsaw a n d two-bar reversal was confirmed in the subsequent
week as the price violated a small (dashed) s u p p o r t trendline.

The Importance of Monitoring Several Time Frames
Quite often, patterns will develop in the daily charts that are n o t a p p a r e n t
in the weeklies or, say, the hourlies, or vice versa. A trader who is simply concentrating on o n e particular time frame, excluding all others, could easily
miss out on some valuable technical information. It therefore makes sense
to look at charts constructed above a n d below the time frame you are actually trading in.
For example, Chart 16-8, for Westvaco, shows a 15-minute time span.
Several two-bar reversals are featured in the rectangles. Chart 16-9, on the
other hand, shows the same stock, but this time with 5-minute bars. T h e
ellipses indicate the same time periods as the rectangles in Chart 16-8, b u t
they do not contain any technical p h e n o m e n a that indicate a reversal. Thus,

Two- and Three-Bar Reversals

Chart 16-8 Westvaco 15-minute bar. {Source: Telescan.)

Chart 16-9 Westvaco 5-minute bar. (Source: Telescan.)

299

300

Part III: Short-Term Patterns

anyone who was just looking at the 5-minute chart would have missed some
potential opportunities.
It may be annoying to traders who typically use daily charts to see such
small and, they could justifiably say, irrelevant time frames. However, the
same principle could be applied by considering hourly or weekly charts. This
is because the hourly charts could well reveal i m p o r t a n t price patterns, such
as a h e a d a n d shoulders, that are n o t a p p a r e n t with, say, three or four bars
of daily data. Similarly, by looking up o n e time frame to the weeklies, a dailybased trader could well gain some perspective c o n c e r n i n g the m o r e domin a n t trend. Perhaps the weeklies will reveal an outside bar or a two-bar
reversal that is n o t a p p a r e n t on the daily charts.

Two- and Three-Bar Reversals

301

Chart 16-11 DJ Transports, daily.

Two-Bar Reversals as a D o m i n o or Reverse D o m i n o
Sometimes these patterns develop at major j u n c t u r e points on b o t h the
intraday a n d the daily charts. This is because the two-bar reversals (or any
o t h e r one- or two-bar pattern, for that matter) act as dominoes or reverse
dominoes. We examined this concept briefly in Chapter 13 with outside bars.
Chart 16-10 features daily price action for the DJ Transports. T h e b o t t o m
p a n e l shows a short-term KST, a s m o o t h e d s u m m e d rate of c h a n g e , an
indicator that I developed many years ago (for a streaming audiovisual tutorial on the KST, go to www.pring.com, "Charting the KST"). A m o n g o t h e r

Chart 16-10 DJ Transports, daily.

things, this versatile series offers buy a n d sell signals when it crosses decisively above or below its 10-day moving average. In Chart 16-10, it is trying
to give a short-term buy signal, but it is n o t quite there. Since such signals
are typically followed by rallies lasting well over t h r e e weeks, they are worth
watching for. W h a t clue can we look for that will increase the probability of
a reliable signal? T h e answer lies in the price action, for as the chart ends,
the Transports complete a two-bar reversal. Chart 16-11 indicates that this
was timely information because the average rallies nicely a n d the KST eventually gives us a buy signal.

Three-Bar Reversals
Basic Characteristics and Underlying Psychology
A three-bar reversal is essentially a two-bar reversal separated by an exhaustion bar. An example of a top is shown in Fig. 16-7, a n d an example of a
b o t t o m in Fig. 16-8. C o m i n g into a top pattern, the buyers are very m u c h
in control. T h e second bar usually indicates exhaustion in the form of a
Pinocchio bar, with most of the trading taking place above the o p e n a n d
close. So it is evident that while buyers are able to push prices significantly
higher d u r i n g the bar's formation, they are unable to keep it there. In o r d e r
to reflect the a b r u p t c h a n g e in psychology that takes place, it is i m p o r t a n t

Part III: Short-Term Patterns

302

Two- and Three-Bar Reversals

303

This also underlines the idea of buyer or seller exhaustion, d e p e n d i n g on
the direction of the trend.

Marketplace Examples

Figure 16-7

Three-bar reversal at a top.

for the first a n d last bars to be as wide as possible relative to those preceding t h e m d u r i n g the trend leading up to the pattern. If the bar in the middle is a Pinocchio type, then the wider the trading above the o p e n a n d
closing price, the better, because that implies exhaustion. T h e opposite
would be true for a bottom. In addition, a three-bar pattern will gain a few
stars if the center bar is a c c o m p a n i e d by a noticeable expansion in volume.

Chart 16-12, featuring Cisco, shows a three-bar bottom reversal. It's not really
a classic formation because it is missing o n e element, a n d that is substantial width in the third bar. However, the center bar is accompanied by lots
of volume a n d is unusually wide, which m o r e than compensates for the lack
of width in the third bar.
Chart 16-13, for sugar, shows a n o t h e r three-bar reversal. This time the center bar turns out to be a one-day island. Note that the o p e n a n d close are
right at the low, so all the trading that occurred that day was d o n e at h i g h e r
levels. At the e n d of the day, virtually everyone who h a d b o u g h t was locked
in with a loss. This was hardly what they expected given the sharp run-up
p r e c e d i n g this pattern.
Cintas, in Chart 16-14, provides us with a nice three-bar reversal top. T h e
arrow shows the relatively high level of volume on the Pinocchio day. Note
that all the trading took place above the o t h e r two bars, as indicated by the
d a s h e d horizontal trendline. T h e violation of the up trendline the day after
this p a t t e r n was c o m p l e t e d provided strong confirmation that the formation was valid.
Chart 16-12 Cisco, daily.

Figure 1 6 - 8

Three-bar reversal at a bottom.

304

Chart 1 6 - 1 3

Chart 1 6 - 1 4

Part III: Short-Term Patterns

Sugar, daily.

Cintas, daily.

Two- and Three-Bar Reversals

Chart 1 6 - 1 5

305

Cnet, daily.

Chart 16-15, featuring Cnet, is missing some of the ingredients. First, the
center bar is n o t a true Pinocchio, since the o p e n a n d close are not close to
each other relative to the bar's width. Also, the close is higher than the open.
T h e volume on this center bar is pretty heavy, a n d the width of the flanking
bars is acceptable. Normally we would expect to see m o r e of a decline following a three-bar reversal than the eight or so days indicated on the chart.
T h e real point of showing this example, though, is that even a high-volume
center bar does n o t necessarily have to take on Pinocchio characteristics in
order to work, t h o u g h it is preferable for it to do so.

Quiet Three-Bar Reversal
A variation on the three-bar reversal is what I call a quiet three-bar reversal
(see Fig. 16-9). T h e "quiet" part refers to the center bar, which is n o t a
Pinocchio, b u t is very narrow relative to the two-bar reversal flanking it.
Volume, unlike that in the standard three-bar reversal, is often quite low.
This, of course, indicates a distinct lack of interest on the part of b o t h parties. It also acts as a contrast to the wide bar that precedes it, where the dominant party is urgently buying or selling, d e p e n d i n g on the direction of the
trend. Typically the o p e n a n d close of the center bar will be very n e a r each
other. T h e implication of this proximity is that both parties are pretty evenly
matched following the strong up- or downtrend that precedes it. T h e narrow

306

Part III: Short-Term Patterns

Two- and Three-Bar Reversals

307

Figure 16-10 A three-bar and a quiet three-bar reversal as rectangles.
Figure 16-9 Quiet three-bar reversal at a bottom.
bar therefore warns that whoever was previously in charge may no longer
be in control. It is a less strong signal than the Pinocchio bar, b u t a subtle
o n e nonetheless. T h e clincher comes on the third bar, which o p e n s close
to the first bar's closing price a n d closes n e a r that bar's o p e n i n g . T h u s , by
the time the third bar closes, t h e r e has b e e n little or no n e t gain, despite
all the volatility involved in the pattern's formation.
Figure 16-10 shows a comparison of a regular three-bar reversal with a quiet
three-bar reversal, with the trading ranges being represented by rectangles.
Chart 16-16, for Cisco, shows a quiet three-bar reversal following a shortterm rally. T h e arrow indicates that the center b a r was a c c o m p a n i e d by very
low volume, reflecting a fine balance between buyers a n d sellers following
the previous d o m i n a n c e of the buyers. This p a t t e r n was immediately confirmed by a trendline break, which left little d o u b t that a short-term downtrend was underway.

Three-Bar-Plus Reversal
This is a variation on the quiet three-bar reversal that includes m o r e than
o n e quiet bar, h e n c e the inclusion of the word plus. Examples are shown in
Figs. 16-11 a n d 16-12.
T h e c o n c e p t remains the same, w h e t h e r we have o n e , two, or t h r e e narrow trading bars. They all indicate a fine balance between buyers a n d sellers flanked by two very decisive wide bars.

In o r d e r to work well, these three-bar-plus patterns really n e e d to be prec e d e d by a sharp price m o v e m e n t because they should reflect a quick transition in sentiment between o n e t r e n d a n d another.
Apple C o m p u t e r p r o d u c e s a three-bar-plus reversal in Chart 16-17. I
would have liked to have seen a wider trading range for the bar p r e c e d i n g
the two center bars, b u t there was no question that the sessions p r e c e d i n g
Chart 16-16 Cisco, daily.

308

Part III: Short-Term Patterns

Two- and Three-Bar Reversals

309

Chart 16-17 Apple, daily.

IIIIIIIIIIIII
Figure 16-11 Three-bar-plus reversal at a top.
it were d o m i n a t e d by sellers. However, the final bar in the formation m o r e
than m a d e up for this. T h e volume configuration for this pattern shows a
r o u n d i n g formation, which is j u s t what we like to see. If the volume h a d
e x p a n d e d a bit m o r e on the final day, this would have b e e n great, b u t you
can't have everything!

Two-Bar Reversals and
Candlesticks
T h e r e are no well-known Japanese candlestick patterns that correspond to a
two-bar reversal. T h e closest is an engulfing pattern, described in Chapter 13.
Two-bar reversals would show up in the candles as two very wide real bodies
with little or no u p p e r a n d lower wicks (see Figs. 16-13 a n d 16-14). T h e first
real body would be colored in the direction of the prevailing trend, a n d the
second would have the opposite color. A top would therefore consist of a
black candle followed by a white o n e , a n d vice versa.

Three-Bar Reversals and
Candlesticks

Figure 1 6 - 1 2

Three-bar-plus reversal at a bottom

Three-bar reversals are close to a doji star, although the Pinocchio aspect
of the three-bar reversal means that the center bar is usually wider than in
the doji star. Examples of m o r n i n g a n d evening dojis are shown in Fig.1615. Quiet three-bar reversals are m o r e like the m o r n i n g a n d evening star.
A m o r n i n g star is featured in Fig. 16-16. T h e underlying psychology of all
these patterns is very similar: A strong trend gives way to a balance between

310

Part III: Short-Term Patterns

Two- and Three-Bar Reversals

311

Figure 16-15 Doji stars.

Figure 16-13 Candlestick rendition of a two-bar reversal top.

Figure 16-14 Candlestick rendition of a two-bar reversal bottom.

buyers and sellers, which then leads to a new trend favoring the party that
was dominated in the previous trend.
The three-bar-plus reversals are in many instances like tower tops and
bottoms (see Chart 16-18). These patterns involve a very narrow, saucerlike trading range flanked by two very wide real bodies of differing colors. The idea of the two wide candles sandwiching several smaller ones
is again indicative of two strong trends being separated by a fine balance
between buyers and sellers. The principal difference, as you can see from
Chart 16-18, is that the top of a tower takes a lot longer to form than a
quiet three-bar-plus reversal.

Figure 16-16 Morning star.

312

Part III: Short-Term Patterns

Chart 16-18 General Motors, 1994-1995, and a tower top. (Source: pring.com.)

Two- and Three-Bar Reversals

313

wide bars at a top or toward the low of the two wide bars at a bottom. T h e
first wide bar at a top o p e n s n e a r its low a n d closes near its high. T h e seco n d wide bar o p e n s n e a r its high a n d closes n e a r its low. For a bottom,
the first wide bar o p e n s n e a r its high a n d closes n e a r its low, a n d the seco n d wide bar o p e n s n e a r its low a n d closes near its high.
• Volume characteristic: Above-average volume on the two wide bars a n d the
Pinocchio bar. Small volume on t h e narrow bar in t h e quiet three-bar
reversal.
• Measuring implications: N o n e , b u t usually followed by a sharp short-term
reversal.
• Reliability: Usually good w h e n p r e c e d e d by a strong i n b o u n d trend.

Quiet Three-Bar-Plus Reversal Quick Review
T h e same as quiet three-bar reversals except that a single narrow bar is
replaced with two to four narrow bars.

Summary
Two-Bar Reversals Quick Review
• Price characteristics: Two wide bars of roughly the same height following a
strong trend. At a top, the first bar opens n e a r its low a n d closes n e a r its
high. T h e second opens n e a r its high a n d closes n e a r its low. For a bottom, the first bar opens n e a r its high a n d closes n e a r its low, a n d the seco n d o p e n s n e a r its low a n d closes n e a r its high.
• Volume characteristic: Both bars at tops a n d bottoms should be accompanied by higher-than-average volume.
• Measuring implications: N o n e , but usually followed by a strong short-term
reversal.
• Reliability: Usually good w h e n p r e c e d e d by a strong i n b o u n d trend.

Three-Bar Reversals Quick Review
• Price characteristics: Two wide bars of roughly the same height following a
strong trend. They are separated by either a Pinocchio-type bar or a narrow bar (quiet three-bar reversal) that forms toward the high of the two

PART IV
Miscellaneous
Issues

17

How to Assess
Whether a
Breakout Will Be
Valid or False
When most traders or investors commit money to the marketplace based
on a price pattern breakout, they always assume that the trade will be profitable. However, it is widely understood that most professional traders—or
should I say successful professional traders—expect to lose as many times
as they win. What makes them successful is the fact that the losing trades
are small relative to the winners. "Cut your losses and let your profits run"
is a hackneyed expression, but the principle definitely works. Throughout
this book, I have tried to emphasize that technical analysis does not deal in
certainties, but only in probabilities. It is therefore important that whenever
we take action, we look over our shoulders and ask ourselves the question,
What if? What if the pattern fails? Where should we get out, or what technical action would we have to see in order for the situation to revert from
bullish to neutral, or even bearish?
There are two reasons why we need to go through this process. First, price
trends are nothing more than people in action. People can and do change
their minds, and thus so do markets. Consequently, it's important to be prepared ahead of time to deal with such situations. Second, if we mentally
rehearse what our reaction might be if the technical evidence were to
change, we will be in a more objective psychological state if this actually happens. Failure to prepare in this way will result in small losses turning into
large ones. For example, if we buy a stock on a breakout, and the n e x t day
317

Part IV: Miscellaneous Issues

318

it sells off into the body of the pattern on bad news, what do we do? Most
people will hold on; after all, they have just experienced a loss when they
were expecting a profit. In most instances, they will convince themselves that
the trend is still up, even though the breakout has been invalidated. The
tendency of most people is to stubbornly hold on in the hope that the price
will come back. The right thing to do is to recognize that false breakouts
are a fact of life and a cost of doing business. Therefore, it is important to
anticipate potential adversity by planning ahead and placing stops in the market in the event that the worst does happen. This chapter is concerned with
learning to anticipate when a failure might develop and pointing out some
intelligent places to put stops. In the meantime, let's take a few moments
to demonstrate why cutting losses is so important.

Why Cutting Losses Is Important
Table 17-1 shows a hypothetical example of a trading experience that
begins with a large loss—50 percent, in fact. The next four trades gain a total
of 80 percent. You would think that this would be more than sufficient to
offset the initial loss and earn a nice profit. However, the table clearly shows
that the trader is barely back to breakeven. His record is 4-1 on the winning
side, and yet he has failed to make any money. Add to this the fact that we
have not included commissions and slippage from poor executions, and you
can appreciate why it is important to cut those losses to a minimum.
You may be saying to yourself that if our hypothetical trader had lost the
50 percent at the end instead of the beginning, he would have been OK.
However, Table 17-2 shows that the end result is virtually identical. A large
loss is a large loss however you cut it. The point of the exercise is to show
that it is mathematically much harder to come back from a large loss than
from a small one. Therefore, while we cannot avoid losses, we should at least
take the time when trading price patterns to plan ahead and provide for
false breakouts in our trading plan.

Table 1 7 - 1
Capital
Capital
Trade
Trade 11
Trade
Trade 22
Trade
Trade 33
Trade
Trade 44
Trade
Trade 55

100
100
50
50
75
75
82.5
82.5
90.8
90.8

Profit/Loss
Profit/Loss
%
%

Profit/Loss
Amount
Amount

(50)
(50)
50
50
10
10
10
10
10
10

(50)
(50)
25
25
75
75
83
9.1

Ending
50
75
82.5
90.8
(99.9)

How to Assess Whether a Breakout Will Be Valid or False

319

Table 1 7 - 2
Profit/Loss
Trade
Trade
Trade
Trade
Trade

1
2
3
4
5

Capital

%

100
150
165
181.5
199.65

50
10
10
10
50

Profit/Loss
Amount

Ending

50
15
16.5
18.15
99.825

150
165
181.5
199.65
(99|)

When Do False Breakouts Most
Commonly Occur?
The most common reason why price pattern breakouts fail is that they
develop in a contra-trend way. That means bullish breakouts in bear trends
and bearish breakouts in bull trends. If the security you are following is experiencing a primary bear trend, that trend is the dominant one, so far as the
short- and intermediate-term trends are concerned. That means that most
of the magnitude of these shorter trends is on the downside. Reactions are
large and rallies short. As a result, in most cases, bullish price patterns have
used up most, if not all, of their positive potential by the time the breakout
develops. After the breakout, the dominant primary trend once again takes
over, and prices slip. After they fall below the lower part of the pattern, a new
down leg in the bear market gets underway. The reverse would be true for
downside breakouts in a bull market. We cannot say that every contra-trend
breakout results in a whipsaw, but it is certainly surprising how many do,
This can be appreciated from Chart 17-1, featuring Echostar. The chart
has been divided into two parts, a primary bull market and a primary bear
market. The idea is to demonstrate how buyers dominate during a bull market and sellers during a bear market. Underneath is a nine-day RSI
smoothed with an eight-day moving average. [For a full explanation of the
RSI, please see Martin Pring on Market Momentum, Volume 1 (McGraw-Hill,
2002.)] We would expect bullish psychology to dominate during a primary
uptrend and bearish sentiment to have the upper hand during a negative
trend. Since oscillators, such as the smoothed RSI, closely reflect psychology, we would also expect to see the character of this indicator change as
the primary trend changes direction. The two thick sets of parallel lines
attempt to illustrate this by showing that the oscillator appears to have a bullish upside bias in an uptrend and a downside bias during a bear market.
The point here is that if an upside breakout takes place in a contra-trend
manner, it is not likely to hold because the oscillator does not have much
potential to remain overbought.

320

Part IV: Miscellaneous Issues

Chart 17-1 Echostar, 1997-2002, daily.

Charts 17-2 and 17-3 focus on the two ellipses highlighted in Chart 17-1, since
each of these contains a contra-trend breakout. T h e first reproduces the action
in the ellipse on the right and shows an upside breakout during the bear market. Note that the RSI was at an overbought reading at the time of the completion of the pattern. If you refer back to Chart 17-1, you can see that during
the primary bear trend, the price was extremely sensitive to overbought conditions. T h e actual high for the move, which was marginally above the breakout high, was associated with a negative divergence in the smoothed RSI.
If a purchase h a d b e e n m a d e on the basis of the breakout, where should
the protective stop have b e e n placed? T h e answer would be below the breakout level, since that m a r k e d the low point between the late August a n d late
S e p t e m b e r highs. It was also the point at which the dashed up trendline was
intersecting. A violation of that line would have confirmed the negative
divergence as well.
T h e o t h e r ellipse, featured in Chart 17-3, shows a downside breakout.
Referring back to Chart 17-1 again, you can see that this developed in a bull
market, when the oscillator was very sensitive to oversold conditions b u t n o t
to overbought ones. It is questionable w h e t h e r this was a whipsaw, since the
price did achieve the pattern's m i n i m u m ultimate downside objective.
However, since it reversed on a dime, the trader would have h a d to be
extremely nimble to make m u c h of a profit. If a short trade h a d b e e n initiated on the breakout, a protective stop should have b e e n placed above the

How to Assess Whether a Breakout Will Be Valid or False

Chart 17-2 Echostar, 1999-2000, daily.

Chart 17-3 Echostar, 1997-1998, daily.

321

322

Part IV: Miscellaneous Issues

p o i n t where the probabilities favored the pattern's no longer having any
downside influence on the price. In this instance, that would have o c c u r r e d
when the two dashed down trendlines were violated.
We must continue to emphasize the p o i n t that contra-trend breakouts do not
always result in whipsaws, and pro-trend breakouts are not always valid. However,
as is fairly evident, oscillators tend to be very sensitive to contra-trend overb o u g h t / o v e r s o l d extremes. This means that the ability of most contra-trend
breakouts to hold is relatively limited.

How to Assess Whether a Breakout Will Be Valid or False

323

Chart 17-4 Comcast, 1995-1996, daily.

Further Pointers to Invalid
and Valid Breakouts
For a breakout to stand a better chance of being valid, it needs to be decisive.
What is decisive is very m u c h in the eye of the beholder, a call made on the
basis of j u d g m e n t and experience. Unfortunately, it is not possible to give a
percentage beyond which a breakout can be said to be decisive because different securities have different volatility. A breakout for a low-volatility utility
would not have to be as decisive as, say, one for a highly volatile mining or technology stock.
A second consideration is time frame. A b r e a k o u t of 3 p e r c e n t on a 10m i n u t e bar may well constitute the whole move. On the o t h e r h a n d , such
a b r e a k o u t on a monthly chart would be reasonable, since the implied new
trend would be expected to last for n i n e m o n t h s or m u c h longer a n d have
far greater m a g n i t u d e than a t r e n d on a daily or intraday graph.
O n e b e n c h m a r k used by many technicians is to allow the b r e a k o u t to
maintain itself for at least two bars. If it holds for only o n e bar, it is argued,
t h e n this is a sign of exhaustion. H o l d i n g for two bars represents a sign of
strength a n d confirmation. This rule will certainly filter out a lot of whipsaws, b u t definitely n o t all of t h e m . T h e disadvantage is that by the time t h e
second bar has b e e n completed, prices may well have moved a long way from
the b r e a k o u t point. Unless you are long the security in question a n d have
to liquidate, or are short a n d forced to cover, this is n o t that important,
because the only thing you have lost is an opportunity. Markets being what
they are, t h e r e will always be a n o t h e r opportunity a r o u n d the corner. T h e
key is to have the patience a n d the discipline to wait for it.

Divergences and Breakouts
Sometimes prices work their way higher, b u t an oscillator does not. This provides us with a clue that the upside b r e a k o u t might fail. Chart 17-4 features

a breakout by Comcast. T h e price h a d b e e n rising in a series of higher peaks
between November 1995 a n d March 1996, but the 10-day rate of change had
experienced a series of lower peaks. This is known as a negative divergence
a n d is a bearish sign, since it indicates that the b r e a k o u t was n o t accompanied by m u c h in the way of upside m o m e n t u m . W h e n the price r e t u r n e d
to the body of the pattern a n d the rate of change peaked, that should have
b e e n e n o u g h to indicate a failed breakout. However, the issue was placed
beyond reasonable d o u b t a couple of sessions later when the price violated
the dashed up trendline.
T h e same principles are true for downside breakouts, b u t h e r e the conditions would be reversed: Falling prices a n d a false downside would be associated with a positive divergence between the price a n d the oscillator.

Volume and Breakouts
Having established what we might call m a c r o guides to potential whipsaws,
it is now time to turn o u r attention to smaller a n d m o r e subtle guides to
w h e t h e r a b r e a k o u t will lead to success or failure.
O n e of the most obvious of these guides is volume. It has already b e e n
established that it is better for upside breakouts to be a c c o m p a n i e d by a
noticeable expansion of activity, since this is an indication of enthusiasm on

324

Part IV: Miscellaneous Issues

Chart 17-5 Merrill Lynch, daily.

the part of buyers. However, many upside breakouts that are n o t accompanied by high volume are successful, so n o r m a l volume is n o t necessarily a
"pattern killer." It's just that heavy volume increases the odds that this is a
valid breakout. T h e thing to look o u t for in this regard is a definite decline
in volume as the price breaks to the upside. This is a bear market characteristic a n d is a definite red flag suggesting that the b r e a k o u t is developing
m o r e because of a lack of selling pressure than because of buying enthusiasm. Chart 17-5, featuring Merrill Lynch, shows a b r e a k o u t above a down
trendline, but on the day of the b r e a k o u t t h e r e is virtually no volume. T h e
price b o u n c e d a r o u n d for a few days before h e a d i n g south in a big way.
It is normal for volume to contract during declines, so a decline on downside breakouts does not provide any clues to whether the breakout will be invalid
or not. On the other hand, if volume picks up on the downside penetration,
then this is abnormal and indicates selling pressure. T h e expanding volume is
therefore a plus in terms of increasing the probabilities of a valid signal.

One- and Two-Bar Price Patterns
A false b r e a k o u t is, by definition, the e n d of a trend, a n d for this reason
many whipsaws show up as exhaustion moves. These are reflected on the
charts in the form of two-bar reversals, Pinocchio bars, a n d so forth. Such

How to Assess Whether a Breakout Will Be Valid or False

325

Chart 17-6 Merrill Lynch, 1981-1982, daily.

formations have only short-term implications, b u t their very presence at
b r e a k o u t time warns us of m o r e significant trouble than the pattern considered in isolation. This is because we do n o t expect to see an exhaustion
pattern as the price is breaking out, for exhaustion comes at the e n d of a
move, not the middle. W h e n the price breaks from a pattern, it is reasonable to anticipate an extension of that trend. T h e a p p e a r a n c e of a one- or
two-bar reversal should alert us to the fact that the probabilities of failure
have definitely increased. It is n o t possible to say that every time o n e of these
formations develops right after a breakout, the signal will turn out to be a
whipsaw, b u t it is definitely a red flag.
Chart 17-6, for Merrill Lynch, offers an example. T h e price breaks out
from a base, but the very next day it experiences a Pinocchio. This becomes
evident the day after, when it gaps down a n d closes well within the base. T h e
flavor is that of a three-bar reversal, but the gap, t h o u g h bearish, does n o t
exactly m e e t o u r definition of this formation. In a sense this chart is misleading. If you look carefully, you will see that the way the data are being
reported, the o p e n i n g a n d closing for each day are identical. Consequently,
t h e o p e n i n g data are incorrect, so we really do n o t know where any of the
openings took place. In reality this may n o t have b e e n a Pinocchio at all,
but prices certainly broke in the expected way.
Chart 17-7, for Knight Ridder, features a n o t h e r whipsaw. This time the
o p e n i n g a n d closing prices are r e c o r d e d , a n d they definitely reveal a

326
Chart 17-7 Knight Ridder, 1992-1993, daily.

Part IV: Miscellaneous Issues

How to Assess Whether a Breakout Will Be Valid or False

327

Pinocchio bar e n d i n g in a failed breakout. Note that the volume on the day
of the whipsaw was pretty heavy. T h e technical position certainly looked
good as t h e b r e a k o u t was taking place, b u t since t h e price e n d e d up w h e r e
it started, below the b r e a k o u t trendline, the volume b e c a m e a negative factor because it showed that t h e r e was overwhelming selling pressure at
higher prices. Note also the three-bar-plus reversal at the e n d of March that
signaled the start of the April rally.
Chart 17-8 shows another false breakout for Knight Ridder, this time on the
downside. T h e whipsaw was signaled by a very wide outside bar accompanied
by a h u g e level of volume. If there had b e e n any d o u b t that this formation
was going to fail, it would have been cleared up at the close of the outside
bar, because the price was well above the dashed down trendline at that time.

Anticipating Breakouts That
Are More Likely to Be Valid

Chart 17-8 Knight Ridder, daily.

Just as it is possible to look for signs that indicate a possible whipsaw, we can
also look for signs that increase the odds that a breakout will be valid. O n e
a p p r o a c h is to look for a potentially strong one- or two-bar price pattern
that develops just prior to a potential breakout point a n d use that as an indication that the next trend is likely to have the power to result in a move that
is strong e n o u g h to result in a valid breakout. In reality, we are looking for
o n e of those reverse domino-type effects discussed earlier.
Chart 17-9, for NCR, offers a good example. T h e price traces out an
upward-sloping head-and-shoulders pattern. At the time it would have b e e n
reasonable to anticipate a valid breakout because of the strong Pinocchio
bar that m a r k e d the top of the head. It was a classic pattern, first, because
it was p r e c e d e d by a very strong a n d persistent advance, a n d second,
because of its wide trading range. Finally, the h u g e volume that accompanied it was icing on the bearish cake. In actual fact, the Pinocchio bar a n d
the two bars s u r r o u n d i n g it were really an island reversal. T h e right should e r of the head-and-shoulders top was a good-faith effort to close the gap,
after which the price was free to make a new low. T h e chart also supports
the idea that exhaustion extremes are i m p o r t a n t potential resistance areas.
T h e horizontal trendline shows that the level experienced by the Pinocchio
extreme t u r n e d back the next advance several m o n t h s later.
T h e pattern in Chart 17-10, for Newmont Mining, is really a failed bullish
right-angled triangle. Its demise could have b e e n anticipated by the two-bar
reversal that developed on the rally preceding the downside breakout. This
two-bar formation was n o t a particularly strong pattern because it was n o t
accompanied by a lot of volume or p r e c e d e d by m u c h of a rally. However,

328
Chart 17-9 NCR, 1998-1999, daily.

Part IV: Miscellaneous Issues

How to Assess Whether a Breakout Will Be Valid or False

329

the bars were reasonably wide, and it developed at an important resistance
level. T h e two-bar reversal told us two things. First, the late May attempt at
an upside breakout would probably fail a n d require further regrouping.
Second, the a b r u p t turn of fortune, reflected by the characteristics of a good
two-bar reversal, ought to result in a test of the up trendline. W h e n this came
a n d failed two days later, the decisive downside breakout indicated that many
traders h a d anticipated a break above the solid trendline a n d were forced
to liquidate their positions. Incidentally, if this information h a d shown up
on the candle charts, it would have revealed a dark cloud cover, since the
o p e n i n g of the second day was above the close of the first, a n d the second
day closed m o r e than halfway down the real body (i.e., the difference
between the o p e n i n g a n d closing prices) of the first.
We have to use a little c o m m o n sense when using these patterns to predict a legitimate breakout. If they form a long way from the b r e a k o u t point,
they are far less likely to work, since they will n o t provide sufficient m o m e n t u m to act as a springboard for a successful breakout. Chart 17-11, for
Nabors, shows an example. We see a strong outside day on h u g e volume in
early May. This could certainly have b e e n used to anticipate a good rally,
but the outside bar developed so far from the b r e a k o u t trendline that it
would n o t have b e e n reasonable to use it as evidence to project a successful breakout. On the o t h e r h a n d , a two-bar reversal developed j u s t after a
whipsaw breakout. W h e n the price broke above the horizontal trendline for

Chart 17-10 Newmont Mining, 1995-1996, daily.
Chart 17-11 Nabors, daily.

330

Part IV: Miscellaneous Issues

a second time, this provided ample proof that this second b r e a k o u t would
n o t turn o u t to be a whipsaw. Thus we have a breakout, a cancelled breakout (whipsaw) on the first day of the two-bar reversal, a n d a cancelled whipsaw on the second. Although the two-bar reversal was not p r e c e d e d by m u c h
of a decline, the fact that the price was able to make a new high after such
an indecisive period was a strong signal indicating that the buyers were now
in control.

How to Assess Whether a Breakout Will Be Valid or False

331

Chart 17-13 Nasdaq 100 30-minute bar.

The Importance of Using
Perspective
It doesn't matter what time frame you are trading or investing in; it is always
a good idea to m o n i t o r the time frame above a n d the o n e below your chosen time frame. Considering the time frame above your chosen time frame
gives you the perspective of a m o r e d o m i n a n t trend a n d will help to either
validate any signals you obtain or give you a stronger a n d m o r e reliable indication of a whipsaw.
For example, Chart 17-12 shows a false b r e a k o u t for the Nasdaq on a 10m i n u t e bar. T h e false b r e a k o u t b e c a m e a p p a r e n t when the price broke
below the previous m i n o r low. On the o t h e r h a n d , if we look up to a higher
level, as in Chart 17-13, which features a 30-minute bar, we can see that the

Chart 17-12 Nasdaq 100 10-minute bar.

whipsaw was an outside bar. T h e two indications of whipsaws developed at
roughly the same time, but the outside bar being formed at a higher a n d
m o r e d o m i n a t i n g time frame represented a m u c h stronger signal.
Chart 17-14 features the U.S. G o v e r n m e n t 10-year yield. T h e head-andshoulders reversal indicates that the yield is h e a d e d higher, b u t this gives us
only part of the picture. That's because Chart 17-15 showed that on a weekly
basis, a large double bottom h a d yet to realize its full indicated upside potential. In o t h e r words, Chart 17-14 tells us that the situation is positive, but
Chart 17-15 indicates that it is actually substantially better than what is shown
by the shorter-term chart. Note that on the gap up day of the b r e a k o u t the
price action has many of the ingredients of an exhaustion bar. A cautious
trader would have b e e n justified in staying away from the trade. However,
when the price b r o k e above the high for the exhaustion day it was definitely
canceled. Unfortunately, t h e r e was n o t a low risk point u n d e r n e a t h which
a stop could have b e e n placed.
In this example we looked up to see what a longer-term t r e n d was signaling. In Charts 17-16 a n d 17-17, we look the o t h e r way.
T h e large ellipse in Chart 17-16 shows a three-bar-plus reversal. However,
for the p u r p o s e of this exercise, we are c o n c e r n e d with the trading activity
contained within the smaller ellipse. Looking at this daily chart, there is no
indication that a reversal is in progress; that does not come until the closing
of the three-bar-plus reversal. However, h a d we taken the trouble to look at

332

Part IV: Miscellaneous Issues

How to Assess Whether a Breakout Will Be Valid or False

Chart 1 7 - 1 4

10-year government bond yields, daily.

Chart 1 7 - 1 6

September 2003 Euro FX, daily.

Chart 1 7 - 1 5

10-year government bond yields, weekly.

Chart 1 7 - 1 7

September 2003 Euro FX 10-minute bar.

333

334

Part IV: Miscellaneous Issues

the intraday activity, as shown in Chart 17-17, we would have seen that the
price was forming a triangle bottom. Note that the breakout confirmed the
rising trend of the two m o m e n t u m indicators, which were both in the type
of subdued position that could support a good short-term rally. (For a description of the KST m o m e n t u m indicator, please go to the "Trader's Den" section at www.pring.com.) T h e example used h e r e compared a daily to an
intraday situation, but it could easily have been a monthly-weekly or a weeklydaily comparison. The important point is that longer-term charts do not always
reveal potential reversal patterns, but a study of short-term time frames often
does. Therefore it's a good idea to look not only at time frames above the one
that you are trading or investing in, but also at those below.

18

How Do Price
Patterns Test?
This c h a p t e r sets o u t the results of computer-recognized price patterns in
several U.S. industry groups. T h e research was based on criteria that I
defined a n d was c o n d u c t e d by Rick Escher a n d his team at Recognia Inc.
( h t t p : / / w w w . r e c o g n i a . c o m ) , to w h o m I owe a big d e b t of gratitude.
Historical data used for the testing were kindly provided by Bob Peltier, president of Commodity Systems Inc. (http://www.csidata.com), the world's
most comprehensive provider of commodity futures data. Recognia is an
investment research company devoted to stock screening using sophisticated
analytics, principally technical event analysis a n d chart p a t t e r n recognition.
T h e technology was originally developed for British military use, a n d
Recognia has e n h a n c e d it for the financial markets. Recognia's research and
its online tools are available t h r o u g h several brokerage houses a n d t h r o u g h
a special subscription service whose principal objective is price pattern
screening at www.pring.com.

Basic Problems in Price Pattern
Evaluation
At first glance, o n e might think that the recognition process is fairly simple. This is certainly true for evaluation of, say, moving-average crossovers
or o t h e r statistically derived indications of trend changes. Evaluating price
patterns, on the o t h e r h a n d , is n o t an easy process, because the interpretation of these patterns is a very subjective. Patterns first have to be recognized, a n d what is o n e m a n ' s head-and-shoulders top may n o t be another's.
As an example, I defined a head-and-shoulders top earlier as a final rally
335

Part IV: Miscellaneous Issues

Figure 18-la Symmetrical head and shoulders.

separated by two smaller rallies. Figure 18-la offers a classic symmetrical
example that virtually everyone could agree on. Figure 18-16, on the other
h a n d , meets the criteria of a final rally separated by two smaller ones, but
it would be far m o r e controversial because of the steepness of the neckline
a n d the extremely small left shoulder. A j u d g e in a vice case some years back
said when defining pornography, "I know it when I see it." T h e same is really
true of most price patterns, since very few are formed with the classic appearance defined in the textbooks.
Designing software to interpret price patterns is therefore an extremely
difficult a n d complex task. Recognia has a p a r a m e t e r that allows its patterns

Figure 18-lb Downward-sloping head and shoulders.

How Do Price Patterns Test?

337

to be classified according to a consensus rating—a fuzzy logic measure of a
pattern's closeness to the pattern ideal a n d of the strength of the prior trend
(Recognia calls this the " i n b o u n d t r e n d " ) . In o u r research, we originally
ignored t h e consensus rating, which m e a n t that all patterns, regardless of
the prior t r e n d a n d the pattern's closeness to the ideal, were included. It is
reasonable to ask the question, "Why not search for patterns with a consensus rating of 50 p e r c e n t or 100 percent?" T h e answer is that this resulted
in so few patterns that it r e n d e r e d the research statistically invalid. In the
e n d , we chose to use patterns that h a d a consensus rating of zero or higher.
This m e a n t that the prior trend was taken into consideration when selecting b o t t o m a n d top patterns, b u t that all patterns m e e t i n g the recognition
threshold were considered in the statistical analysis. From a practical point
of view, t h o u g h , the patterns r e t u r n e d by the Recognia search engine at
pring.com have b e e n filtered with a 50 p e r c e n t consensus rating.
O n c e a p a t t e r n has b e e n correctly identified, t h e r e is also t h e question
of m e a s u r i n g w h e t h e r t h e formation has b e e n successful. If a bullish pattern immediately rallies to its price objective, t h e r e will be little dispute
(Fig. 18-2fl). On the o t h e r h a n d , if the price breaks u p , declines well into
t h e b o d y o f t h e p a t t e r n , a n d s u b s e q u e n t l y a d v a n c e s t o t h e objective
(Fig. 18-26), is this a p r o f i t a b l e situation? After all it is q u i t e p r o b a b l e
t h a t in t h e i n t e r e s t of good risk m a n a g e m e n t , any responsible trader would
have liquidated t h e position d u r i n g the postbreakout decline, when it
a p p e a r e d that the p a t t e r n might have failed.
Recognia devised a way to get a r o u n d this p r o b l e m by treating all patterns
equally, basing the results on price as a percentage of the pattern's confirmation price a n d projected time periods of pattern length. In this way, it
was possible to statistically c o m p a r e patterns of differing duration a n d price

Figure 1 8 - 2 a

Head-and-shoulders bottom.

338

Part IV: Miscellaneous Issues

How Do Price Patterns Test?

339

The Concept of L

Figure 18-2b Head-and-shoulders bottom with hesitant breakout.

a n d to offer a statistical projection of reliability for all breakouts. H a d we
used some form of loss limitation, it is very p r o b a b l e that the results from
these patterns would have b e e n far better. I will explain the p a t t e r n length
concept later.

What Was Tested?
For the p u r p o s e of this test, we decided to use two patterns, head-and-shoulders a n d double bottoms a n d tops. Recognizing the influence of industry
trends on individual stocks a n d the i m p o r t a n c e of the direction of the primary trend, we chose four major sectors: financial, energy, transportation,
and retail. With the benefit of hindsight, we were able to categorize primarytrend bull a n d bear markets for each sector during the 1982-2003 period.
Stocks from individual sectors were identified from a database of close to
14,000 U.S. equities.
In o r d e r for a pattern to qualify as a legitimate reversal pattern, there h a d
to be s o m e t h i n g to reverse. For purposes of identification, therefore, the
Recognia consensus rating allows for the existence of what Recognia called
an i n b o u n d trend.
In o n e way, the results are biased in that we already knew the direction
of t h e primary trend, which is obviously n o t possible in real time. On the
o t h e r h a n d , since the testing permitted a theoretically unlimited loss from
short sales or a 100 p e r c e n t loss from long positions, the research results
are adversely affected, since they did n o t take into consideration the possibility of limiting the losses from failed patterns t h r o u g h p r u d e n t money
management.

T h e price behavior following a pattern's confirmation was evaluated in segments that reflected the time taken for the pattern to form. This permitted
patterns of differing duration to be c o m p a r e d on an equal basis. For example, if it took 50 days to complete a formation, this time duration was defined
as o n e L, or length. T h e period following the b r e a k o u t was t h e n measured
in units of L, up to a m a x i m u m of 5 L. In the example of a 50-day pattern,
5 L would be 250 days. This a p p r o a c h m a d e it possible to appraise perform a n c e over several time segments for individual patterns of varying length.
T h e price behavior following a pattern's b r e a k o u t was evaluated in segments of time that represent 10 p e r c e n t of the pattern's length (0.1 L).
T h u s , if a 52-day head-and-shoulders b o t t o m p a t t e r n was identified, t h e n
its p o s t b r e a k o u t prices were evaluated in segments of 5 days (52 divided
by 10 r o u n d e d down to a whole n u m b e r ) .
T h e closing prices for these 5 days were averaged, then the result was
divided by the closing price on the day of pattern confirmation (breakout).
Multiplying by 100 yielded the average price over the interval expressed as
a percentage of the pattern's closing price on the day of confirmation
(breakout). This process was c o n t i n u e d for each 5-day interval t h r o u g h to
5 L, or 50 intervals (10 intervals p e r L times 5 Ls equals 50).
O n e factor that could potentially affect t h e analysis was when some part
of the 5 L projection developed after a primary-trend reversal h a d taken
place. T h u s the results of, say, a bullish p a t t e r n could well be partially evalu a t e d after a b e a r m a r k e t was underway. This was addressed by terminating t h e analysis at t h e b u l l / b e a r b o u n d a r y . This e n s u r e d that the
post-turning p o i n t results were excluded from the calculation for subseq u e n t periods of 0.1 L.

Aggregate Results
In total, 5,235 patterns were identified. Chart 18-1 shows the average and
m e d i a n length for tops a n d b o t t o m s in b o t h bull a n d b e a r primary-trend
environments. T h e average length did n o t vary very m u c h , t h o u g h the time
taken to form patterns was slightly less in primary bear markets than in primary u p t r e n d s . For all 5,000 formations, the average was a r o u n d 49 trading days or 10 calendar weeks. This is interesting, since 10 weeks or 50 days
is a moving average that is widely used by many technicians.
Since we were able to split the data into primary bull a n d bear markets,
it was possible to categorize t h e results as pro- a n d contra-trend signals. Note
that the standard of success in this research is based on the percentage of

340

Part IV: Miscellaneous Issues

Chart 18-1 Tested pattern durations in different environments.

How Do Price Patterns Test?

341

Chart 18-2 Bullish breakouts in bull markets.

Pattern Durations

the objective being attained, n o t on profitability. This is a stricter assessment,
since a profitable pattern could fall well short of the objective.

Pro-Trend S i g n a l s
Aggregate results for all four patterns (head-and-shoulders tops a n d bottoms a n d d o u b l e tops a n d d o u b l e bottoms) are shown in Charts 18-2 a n d
18-5 to 18-7. Chart 18-2 reflects the b o t t o m i n g patterns d u r i n g primary bull
markets. T h e Faxis represents multiples of the price objective. T h u s , the
100 p e r c e n t line is the price objective, the 200 p e r c e n t line is twice the
objective, a n d so o n . Objectives were calculated by projecting the p o i n t
d e p t h of the p a t t e r n upward from the p o i n t of breakout. T h e Xaxis shows
the percentage of patterns reaching a particular level. Finally, the six curves
represent the results for a specific multiple of the pattern length (L). Thus,
at point A we can see that approximately 15 p e r c e n t of all patterns h a d
r e a c h e d their price objective within half the time that the pattern itself took
to form, i.e., 0.5 L. At point B, about half the sample (50 percent) had
r e a c h e d their objective within twice t h e time that it took for the p a t t e r n to

develop (i.e., 2 L). By p o i n t C, for 3 L, approximately 60 p e r c e n t of all
patterns h a d achieved their objective. Finally, nearly 70 p e r c e n t of all patterns in bull markets h a d achieved their objective by 5 L (point D).
To p u t it a n o t h e r way, t h e r e was a 70 p e r c e n t c h a n c e that the price
objective would be achieved within five times t h e time n e e d e d to form the
p a t t e r n . For e x a m p l e , if a p a t t e r n took 30 days to c o m p l e t e , t h e r e was a
70 p e r c e n t c h a n c e that t h e objective would be achieved within 150 days.
Some objectives would be achieved earlier, say in 30 days, b u t in this example, 70 p e r c e n t of all p a t t e r n s would have attained their objectives by day
150. Arguably m o r e impressive is t h e fact that approximately 25 p e r c e n t
of all patterns (point E) achieved 400 p e r c e n t of their objective (four times
the objective) within five p a t t e r n lengths (5 L ) .
It's i m p o r t a n t to u n d e r s t a n d that from a practical p o i n t of view, some of
these results are n o t as good as they look. This is because the path to the
objective is rarely a straight line a n d can often be quite fickle. In o t h e r
words, a p a t t e r n may reach its objective or a multiple thereof, b u t its volatility before it gets there could easily frustrate the trader, resulting in liquidation

How Do Price Patterns Test?

343

Chart 18-3 Bullish breakout—good example.
Event ID: 250877

Event Duration: 24

VLO: NYSE (daily)
2000-01-18: Head and shoulders bottom

Chart 18-4 Bullish breakout—bad example.
Event ID: 165894
PTEN: NASDAQ (daily)
2000-06-29: Head and shoulders bottom

342

before t h e objective is r e a c h e d . T h a t is o n e reason why it does n o t make
sense to trade on price patterns alone. It is i m p o r t a n t to also check that
m o m e n t u m is n o t unduly o v e r e x t e n d e d at the time of the b r e a k o u t a n d so
forth. T h e word unduly has b e e n emphasized because it is very likely that
m o m e n t u m will be partially o v e r e x t e n d e d to some degree at the time of
most breakouts. It is often a g o o d idea to check the m o m e n t u m calculated
from a time frame h i g h e r t h a n the o n e that is b e i n g traded. For example,
anyone who is trading off the daily charts with a two- to three-week horizon s h o u l d check m o m e n t u m indicators based o n i n t e r m e d i a t e time
frames, a n d so o n .
Charts 18-3 a n d 18-4 illustrate the point m a d e earlier concerning volatility. Chart 18-3, for example, offers a classic result with a m o r e or less immediate move toward a multiple of the objective. On the other h a n d , Chart 18-4
shows a far m o r e difficult situation. T h e objective was easily obtained, but
not before a nerve-shattering decline well into the body of the formation.
Chart 18-5 shows that the results from tops in primary bear markets were
n o t as reliable as those from bottoms in bull markets, since it took 3 L before
half the sample reached its objective. This compares to only 2 L for the bottoms in bull markets. By the same token, p o i n t B shows that it took 4 L

Chart 18-5 Bearish breakouts in bear markets.
Event Duration: 25

344

Part IV: Miscellaneous Issues

before 60 p e r c e n t of tops r e a c h e d their target, whereas bottoms took only
3 L. Finally, 25 p e r c e n t of all patterns r e a c h e d approximately 270 p e r c e n t
of their objective by 5 L. This c o m p a r e d to a 400 p e r c e n t achievement for
25 p e r c e n t of bullish patterns in bull markets.

How Do Price Patterns Test?

345

Chart 18-7 Bearish breakouts in bull markets.

Contra-Trend Signals
Not surprisingly, contra-trend signals (e.g., bullish patterns in bear markets)
did n o t perform as well as breakouts that developed in the direction of the
primary trend. In Chart 18-6, we see that only a b o u t 55 p e r c e n t of all bullish breakouts in bear markets achieved their objective by 5L (point A). This
compares to 70 p e r c e n t for positive breakouts in bull markets. Moreover,
t h e m a g n i t u d e of most breakouts was far less. In the case of t h e pro-trend
moves, 25 p e r c e n t of patterns achieved 400 p e r c e n t of their target, b u t for
positive breakouts in bear markets, the c o r r e s p o n d i n g percentage sank to
just over half that level (point B).
In Chart 18-7, only about 43 percent of tops in bull markets reached their
objective within 5L, compared to around 62 percent in bear markets (point A).

Chart 18-6 Bullish breakouts in bear markets.

While we obviously c a n n o t say that all contra-trend breakouts will fail,
t h e r e s e a r c h definitely confirms that t h e o d d s of success are greatly
increased with signals that develop in t h e direction of t h e prevailing primary trend.

Conclusion
This research strongly suggests that price patterns, at least the head-andshoulders a n d double varieties, work. Pro-trend signals are m u c h m o r e likely
to result in positive returns than contra-trend signals. Unfortunately, a tight
publishing schedule did n o t p e r m i t a m o r e t h o r o u g h examination of the
results or the possibilities of testing for o t h e r ideas—for instance, objectives
based on p e r c e n t a g e m e a s u r e m e n t s as opposed to point or dollar moves,
or the ability to limit losses o n c e a price h a d seriously b r o k e n back into the
body of the formation. It would also have b e e n interesting to assess the
results by limiting the identified patterns to slightly m o r e stringent consensus ratings to see if the results would have c h a n g e d any.

Appendix
Individual Patterns
Summarized

Chart A-l

W

Chart A-2

o

Chart A-3

Chart A-4

Chart A-5

Index

Index
Accumulation, 63
And rectangles, 75-76, 85-86
Area gap, 224
Arithmetic scaling, 9, 19-21
Defined, 19
Head-and-shoulders measurement
implications, 100-101
Measuring implications for price
patterns, 81-82
And price patterns, 83-84
And trendlines, 41-45
Ascending right-angled triangle
Characteristics, 158
Failures, 161-164
Retracement moves, 158
Volume characteristics, 159
Psychology of failures, 165-167
Bar charts
And head-and-shoulders formations,
101-104
Vs. line or close only, 38
Black candle, 244
Breakaway gap, 224-228, 239
Breakouts
And divergences, 322-323
And volume, 323-324
Tips for spotting valid breakouts,
327-330
Broadening formation with a flat bottom
Characteristics, 171
As a continuation pattern, 179
Failures, 180-182
Measuring implications, 175-176
Underlying psychology, 174-175
Volume configuration, 171
Broadening formation with a flat top
Characteristics, 171
As a continuation pattern, 179

354

Failures, 180-182
Measuring implications, 175-176
Underlying psychology, 173-174
Volume configuration, 171
Broadening wedges, 182-185
Cancellations, 85-86
Churning, 63
Chinese double bottom, 135-138
Classic gaps, 234
Close-only charts
And head-and-shoulders formations,
101-104
Vs. bar charts, 38
Common gap, 224
Complex head-and-shoulders patterns, 105
Consolidation patterns, 7
Continuation gaps (see runaway gaps)
Continuation patterns, 7
Contra-trend signals, 339, 344-345
Countercyclical breakout, 10
Csidata.com, 335
Cup-and-handle pattern, 201-203
Cutting losses, 314, 318
Dark cloud cover, 257, 329
Descending right-angled triangle, 161
Failures, 161-164
Psychology of failures, 165-167
Diamonds top, 188-190
Distribution
And rectangles, 75-76, 82, 85-86
Divergences and breakouts, 322-323
Doji evening star, 309
Doji morning star, 309
Double bottom, 114, 128
Characteristics, 132
Chinese double bottom, 135-138
As consolidation patterns, 140

Failures, 140
"Lucky seven" double bottom,
142-144
Measuring implications, 32
Platform double bottom, 138-139
Underlying psychology, 133
Volume configuration, 132
Whipsaw double bottom, 142-144
Double top
Characteristics, 127-129
Measuring implications, 130
Underlying psychology, 129
Volume configuration, 127
Down trendlines
Defined, 35
Downward sloping head-and-shoulders
pattern, 104
Dragonfly doji, 286
Drew, Garfield, 3
Edwards and Magee, 127, 128, 129, 196,
225,232
Elliott, 6
Emotional points on charts as potential
support/resistance, 26-28
Engulfing patterns, 256
Escher, Rick, 337
Evening star, 309
Exhaustion, 225, 242, 245, 269, 280
And resistance, 327
Exhaustion bars, 275-278, 331
Characteristics, 275
Compared to one-day island reversal, 275
Compared to key reversal bar, 275
Underlying psychology, 276-277
Volume configuration, 277
Exhaustion break, 325
And breakouts, 87
And trendlines, 49-53
And volume, 61, 65
Exhaustion gap, 230-232
Failure and countercyclical breakout, 12
Fibonacci, Leonardo, 29
Fifty percent rule, 28, 291
Flag, 217
Characteristics, 211-213
At half mast, 213
Measuring implication, 213-214
Volume configuration, 211-212
Fundamentals, 3, 4

355
Gann, W.D., 6
Gaps, 222-240
Breakaway gaps, 224-228, 239
Common or area gaps, 224
Continuation gaps (see runaway gaps)
Defined, 222
Exhaustion gaps, 230-232
Filled, 223, 224
Intraday gaps, 234-235
Measuring gaps (see runaway gaps)
Runaway gaps, 228-230
Measuring implications, 228-229
As support/resistance zones, 27
Underlying psychology, 223
Wide gaps (see exhaustion gaps)
Gartley, H.M., 127, 231
Giant wedges
Characteristics, 191
Compared to symmetrical triangles,
191
Compared to trend channels, 191
Underlying psychology, 192
Gravestone doji, 286
Harami, 266-267
Hammer, 286
Head-and-shoulders bottoms, 110-117
Characteristics, 110
Complex, 110
As continuation patterns or
consolidation, 117-120
Downward sloping, 110
Failures, 124-125
Measuring implications, 111
Upward sloping, 112
Volume characteristics, 110
Head-and-shoulders tops, 96-109
Arithmetic vs. log scaling, 100-101
Characteristics, 96-97
Complex, 105
Downward sloping, 104
As continuation or consolidation
patterns, 117-120, 236
Failures, 120-124
Inbound trend, 337
Measuring implications, 99
Neckline, 96, 104, 188
Right-shoulder shakeout, 107
Underlying psychology, 98-99
Upward sloping, 104, 108, 327
Volume characteristics, 96—98, 99

356
Inside bars, 259-268
At bottoms, 295
Compared to a Japanese Harami,
266-267
Compared to outside bars, 259
Determining significance, 259-261
Underlying psychology, 259-261
Volume configuration, 261
Intrabar gaps, 234
Intraday trader, 10
Intraday trends, 12
And reliability, 12
Inverse head and shoulders (whead-andshoulders bottoms)
Inverted hammer, 286
Island reversal, 231, 236, 296, 327
Japanese candlesticks
Black candles, 244
Compared to Western one- and two-bar
patterns, 243-244
Importance of opening and closing
prices, 243-244
White candles, 244
Key reversal bars, 269-275
Characteristics, 269
Retracements, 271
Underlying psychology, 270
Volume configuration, 269, 272
KST, 6, 300, 334"
Logarithmic scaling, 9, 19-21
Defined, 20
Head-and-shoulders measurement
implications, 100-101
Measuring implications for price
patterns, 81-82
And price patterns, 83-84
And proportionate moves, 100-101
And trendlines, 41-45
Long-legged doji, 286
Long-term trader, 10
"Lucky seven" double bottom,
142-144
Measuring gaps (see runaway gaps)
Megaphone bottom, 170
Morning star, 309
Moving averages, as potential
support/resistance, 26

Index

News, price reaction to, 8, 271
Nifty Fifty, 3
Nison, Steve, 243
One-day island, 236
One- and two-bar patterns
And breakouts from classic patterns,
325-328
Characteristics, 241-243
Compared to Japanese candlesticks,
243-244
"Day" vs. "bar," 241
General interpretive principles, 242-243
One-decision stocks, 3
O'Neil, William, 201, 202
Opening gap, 234, 235
Orthodox broadening formation,
169-171
Measuring implications, 170
Oscillator, 98, 129, 232
Outside bar top, 245
Outside bar bottom, 245
Outside bars, 244-258, 263, 284, 295, 327,
329, 331
Compared to candles, 256-258
As consolidation pattern, 248
Defined, 244-255
As dominos or reverse dominos,
254-256
Factors determining significance,
245-249
Retracement moves, 250
Underlying psychology, 246-248
And volume, 247
Oversold, 133, 232
Overbought, 232
Parabolic rise, 59-60, 271
Peak-and-trough analysis, 9, 14-19
Full reversal signal, 14-15
Half reversal signal, 15-17
And double tops and bottoms, 129
And head-and-shoulders patterns, 116
And rectangle breakouts, 76
What constitutes legitimate peaks and
troughs, 17—19
Peltier, Bob, 335
Pennant
Characteristics, 214-219
Compared to triangle, 214
Piercing white line, 257

Index

Pinocchio bar, 278-286, 301, 302, 305, 309,
325,327
Characteristics, 278-279
As support/resistance level, 284
Underlying psychology, 279-281
Platform double bottom, 138-139
Positive divergence, 114
Prevailing trend assumed to be in
existence, 77
Previous high
As potential support/resistance zone,
24-25
Previous low
As potential support/resistance zone,
24-25
Price pattern cancellation, 85-86
Price pattern significance, 79-81
Primary trend, 10, 338, 339
Pring, Martin, 98, 243, 319
Pring.com, 300, 334, 337
Proportionate moves as potential
support/resistance, 28
And double tops, 130
One-third two-thirds retracement, 29
Pro trend signals, 339, 340-344
Quiet three-bar reversal, 305-306
Rate of change, 323
Ratio scale, 29
Real body, 243, 256
Recognia.com, 335, 336, 337, 338
Record volume, 64
Rectangle, 71-95, 111, 114
Characteristics, 74-76
Confirmation of valid breakout, 86-88
As continuation or consolidation
patterns, 76-77
Measuring implications, 81-82, 94
And resistance, 73
Retracements, 84—85
Reversal in peak/trough progression, 76
As reversal, 74-76
Significance of, 78, 79-81
Underlying psychology, 74-76
Volume configuration, 91-95
Record volume, 64, 275
Resistance, 173, 273, 284
Defined, 22
Determining significance, of, 31—34

357

Identifying potential zones, 23, 24-31
And rectangles, 73, 81, 87, 90-91
Reversal to support role, 24
Retracement moves, 84-85, 175, 250, 271
Return trendline, 49
Reverse head and shoulders (see head-andshoulders bottom)
Reversal patterns, defined, 7
Right-angled broadening formations (see
broadening formations with a flat
top or bottom)
Right-angled triangle, 158-168
Rounding bottoms, 196-201
Rounding tops, 196-201
Round numbers, as potential
support/resistance, 26
RSI, 6
Smoothed, 254-256, 319
Runaway gap, 228-230
And measuring implications, 228
Saucer bottoms, 196-201
Secular trend, 13
Selling climax, 61-62
Sentiment, 173, 247, 248, 295
Shakeout tops and bottoms, 203-206
Short positions, 90, 163
Sold out vs. oversold, 133
Stops (protective), 90, 108, 163, 272, 284,
293,320
Support, 273, 284
Defined, 22
Determining significance of, 31-34
Identifying potential zones, 23,
24-31
And rectangles, 81, 87
Reversal to resistance role, 24
Super cycle, 13
Symmetrical triangle, 111, 149-158
Characteristics, 149-151
Measuring implications alternative,
153-156
Measuring implications classic,
151-153
Underlying psychology, 151-152
Volume characteristics, 150-151, 152
Technical analysis, defined, 5
And probabilities vs. certainties, 29, 79,
114, 251, 317

358

Three-bar reversals, 301-309
Characteristics, 301-303
Compared to Japanese candlesticks,
309-311
Quiet three-bar reversal, 305-306
Three-bar-plus reversal, 306-308
Underlying psychology, 301-303
Three-bar-plus reversal, 306-308,
Throwback move, 41
Time frames, 9-13
Importance of monitoring several,
298-300
Importance of using perspective,
330-334
And price pattern significance, 79-80
Tower bottoms, 311
Tower tops, 311
Trading and mental rehearsal, 317
Transitional turning point, 72-73, 77
Trendlines
Angle of ascent or descent, 46
Continuation penetration, 38-40
Common sense in construction, 38
Correct construction, 35
Defined, 35
Determining significance of, 45-46
And exhaustion break, 49-53
Extended trendlines, 41
Length, 45
Measuring implications, 46-48
Number of times touched or
approached, 46
Primary trendline, 36
Return trendline, 49
Reversal penetration, 38-40
Secondary trendline, 37
As support or resistance level, 26, 46
Trend channels, 48-49
Throwback move, 41
And whipsaws, 49-53
Trends
Defined, 10
Intermediate, 10
Intraday, 12
Primary, 10
Secular, 13
Short-term, 11
Triangles, 149-168, 217, 334
Triple bottom, 145
Triple top, 144-145
Characteristics, 144

Index

Compared to head and shoulders and
rectangle, 144
Two-bar reversals, 239, 263, 289-301, 327,
329
Characteristics, 289-293
Compared to Japanese candlesticks, 309
And domino and reverse domino effect,
300-301
And fifty percent retracement rule, 291,
295
And support/resistance zones, 290, 291,
292
Underlying psychology, 293
Volume configuration, 293
Up trendlines
Defined, 35
Upward sloping head-and-shoulders
pattern, 104
Volatility and trend reversals, 260, 271
Volume, 54-67
Benefits of volume studies, 54
Differs in a bull and bear market, 59
Goes with the trend, 55
Independent from price, 54
Leads price, 56-57
Negative divergence between volume
and price, 56
Parabolic blow-off, 59-61
Principles of interpretation, 55-67
Record volume, 64, 275
Rising price and falling volume as a
negative factor, 57
Selling climax, 61-62
Wedge
Characteristics, 219
Compared to a pennant, 219, 220
In bear market, 219
In bull market, 219
Volume configuration, 219, 221
Weight of the evidence, 5, 6, 54, 102
White candle, 244
Wick, 243, 256
Wide gaps (see exhaustion gaps)
Whipsaws, 88, 251, 254, 281
Tips on spotting potential whipsaws, 322
And trendline violations, 49-53
When do they occur, 319
Wykoff, Richard, 6

About the Author
Martin J. Pring e n t e r e d the financial markets in 1969 a n d has grown to b e c o m e a
leader in the global investment community. In 1981, he f o u n d e d the I n t e r n a t i o n a l
Institute for E c o n o m i c Research a n d began providing research for financial institutions a n d individual investors a r o u n d the world. Since 1984, he has published a
m o n t h l y m a r k e t review offering a long-term synopsis of the world's major financial
markets, a n d since 1988 has b e e n actively involved in Pring T u r n e r Capital G r o u p ,
a m o n e y m a n a g e m e n t firm.
D e m a n d e d as a speaker worldwide, he is the a u t h o r of several o u t s t a n d i n g books,
including the classic Technical Analysis Explained, now in its fourth edition. Since this
u n i q u e b o o k first a p p e a r e d in 1979, Technical Analysis Explained has established itself
as the n u m b e r o n e guide of its kind, a n d it is used as a training tool by international
technical societies and many universities. It is o n e of the three main books for Level 1
CMT certification for the Market Technicians Association. Translated into m o r e than
eight languages, the b o o k is, as q u o t e d in Forbes, "widely r e g a r d e d as the standard
work for this g e n e r a t i o n of chartists." According to Futures Magazine "it is o n e of t h e
best books on technical analysis to c o m e o u t since Edwards & Magee's classic text
in 1948 ... belongs on the shelf of every serious trader a n d technical analyst." It has
since b e c o m e t h e text on which o t h e r works in t h e field have b e e n based.
Martin p i o n e e r e d the i n t r o d u c t i o n of videos as an educational tool for technical
analysis in 1987, a n d was the first to i n t r o d u c e educational, interactive CDs in this
field. His latest releases include Introduction to Technical Analysis, an eight-hour workb o o k CD-ROM course, Technician's Guide to Day Trading, the Introduction to Technical
Analysis, Breaking the Black Box, Introduction to Candlesticks, How to Select Stocks Using
Technical Analysis, a n d a two-volume course on m a r k e t m o m e n t u m .
Recognized by his peers as a technical leader a n d innovator, he was awarded the
A. J. Frost Memorial Award by the Canadian Technical Analysts Society in 2000, a n d
in 2004 was h o n o r e d with the Market Technician's Association (MTA) Annual Award.
Described by Barron's as a "technician's technician," Martin's articles have b e e n
featured in Barron's, a n d he has b e e n q u o t e d in the Wall Street Journal, the Financial
Times, the International Herald Tribune, the New York Post, a n d the Los Angeles Times
newspapers as well as the National Review.
Over the past 31 years, his research has led to t h e d e v e l o p m e n t of reliable financial a n d e c o n o m i c indicators for timely a n d effective forecasting. Martin's personal
Barometers for the Bond, Stock, a n d Commodity markets have identified major turning points since the 1950s on a timely basis a n d have o u t p e r f o r m e d the b u y / h o l d
a p p r o a c h by a wide margin.
For many years, Martin's primary interest has b e e n e d u c a t i n g students of technical analysis in the basic a n d finer points of this art. He enjoys m e n t o r i n g students
of technical analysis from the college level to professionals already in t h e field, sharing t h e wealth of knowledge he has g a i n e d t h r o u g h his own e x p e r i e n c e a n d
research. In this regard, he has spoken on technical analysis to the D a r d e n Business
School, G o l d e n Gate University, a n d University of R i c h m o n d , Virginia.



---

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.7
Linearized                      : Yes
Encryption                      : Standard V5.5 (256-bit)
User Access                     : Print, Print high-res
Create Date                     : 2009:11:22 09:36:16+03:00
Creator                         : Adobe Acrobat Pro Extended 9.0.0
Modify Date                     : 2009:11:22 18:04:35+03:00
XMP Toolkit                     : Adobe XMP Core 4.2.1-c041 52.342996, 2008/05/07-20:48:00
Metadata Date                   : 2009:11:22 18:04:35+03:00
Creator Tool                    : Adobe Acrobat Pro Extended 9.0.0
Format                          : application/pdf
Document ID                     : uuid:42d8de70-37f7-4e9a-9e90-bdf0bca8216f
Instance ID                     : uuid:1971c1d4-ea11-4d6f-b16a-d46e5bb45a5f
Producer                        : ABBYY FineReader 8.0 Professional Edition
Page Count                      : 185

EXIF Metadata provided by EXIF.tools