Professional Documents
Culture Documents
Journal of TA #67 (2013)
Journal of TA #67 (2013)
Of Technical Analysis
Issue 67
(Published 2013)
Journal of Technical Analysis • 2013 • Issue 67
Table of Contents
Editorial Board 2
Wayne Whaley
Publisher
Market Technicians Association, Inc.
61 Broadway, Suite 514
New York, New York 10006
646-652-3300
www.mta.org
Journal of Technical Analysis is published by the Market Technicians Association, Inc., (MTA) 61 Broadway, Suite 514, New York, NY 10006. Its purpose is to
promote the investigation and analysis of the price and volume activities of the world’s financial markets. Journal of Technical Analysis is distributed to individuals
(both academic and practitioner) and libraries in the United States, Canada and several other countries in Europe and Asia. Journal of Technical Analysis is copyrighted
by the Market Technicians Association and registered with the Library of Congress. All rights are reserved.
The articles in this issue highlight the diversity of topics which are of interest to the technical analyst. George
Schade provides a history of the development of the advance-decline indicators. The MTA is a leader in
providing a record of these historical developments for future generations of technical analysts. George’s
article also reminds us that observing the market and asking simple questions can lead to timeless indicators.
Gregory Kuhlemeyer and Robert Kunkel demonstrate how the classic notion of momentum can be used as a
trading tool in a specific market. The article by Camillo Lento highlights how the application of other fields,
such as fractal geometry, can increase our understanding of the market.
In addition to these three new articles, we are printing the 2010 and 2011 Charles H. Dow Award winning
papers in this issue. Wayne Whaley’s paper, “Planes, Trains, and Automobiles: A Study of Various Market
Thrust Measures,” analyzes the use of thrust and capitulation measures as tools for gauging the potential for
sizable intermediate market moves. “Analyzing Gaps for Profitable Trading Strategies,” written by Richard
Bauer and me, provides insight into the price movement that tends to occur after a stock gaps up or down.
The production of an issue is a complex process, requiring the assistance of many individuals. Of course, a
journal begins with authors who desire to share their knowledge, expertise, and experience with the broader
community of technical analysts. But, that is just the beginning. Once papers are submitted for consideration,
they are subjected to a double-blind review process. In this process, papers are reviewed by at least two
experts who provide feedback regarding the accuracy of the work as well as gauge the suitability of the piece
for publication in the Journal of Technical Analysis. This is known as a double-blind review process because
the reviewers do not know the identity of the author and the author does not know the identity of the reviewers.
In addition, the staff at the MTA office provides significant support in the production and distribution process.
I would like to thank all of those who contributed to this process. If you are considering submitting an article
for a future issue, or if you would like to serve as a reviewer, let me know.
Literature Review
Breakaway gaps occur when price suddenly breaks through a formation boundary and signal the beginning
of a trend. These are thought to be the most profitable gaps. In fact, David Landry (2003) provides a method
for mechanizing trading of breakaway gaps known as the “explosion gap pivot.” Runaway gaps, also known
as measuring gaps, occur during a trend, often in the middle of a price run. These gaps are traded in the
direction of the gap to profit from the directional trend. According to Bulkowski (2010), an upward runaway
gap occurs on average 43% of the distance from the beginning of the trend to the eventual peak, and a runaway
gap down occurs at 57% of the distance on average.
However, a third type of gap, the exhaustion gap, occurs at the end of a strong trend; because price may
reverse immediately or remain in a congestion area for some time, trading these gaps should be avoided.
In hindsight it is easy to recognize an exhaustion gap from the profitable breakaway and runaway gaps,
but as they are occurring, the gaps can have similar characteristics. In his book The Master Swing Trader,
Alan Farley (2000) extends Edwards and Magee’s discussion of gaps to include a “hole-in-the-wall” strategy.
Farley gives extended examples of situations where an exhaustion gap occurs in the opposite direction from
what would be expected to occur.
Few of the detailed studies of gaps have systematically considered gaps occurring in the stocks of publicly
traded companies. Instead, most have dealt with index futures contracts or tracking stocks such as SPY. For
example, Weintraub (2007) claims that the tendency for a gap to be closed is indirectly proportional to the
size of the gap; he attempts to distinguish between common gaps and breakaway gaps by considering the
magnitude of the gap in the mini index futures contracts. Bukey (2008) studies double gaps, defined as two
gaps occurring within ten days of each other, in SPY. He finds that double gaps are unremarkable unless they
are divided into two categories: filled and unfilled. If SPY gaps up twice within a 10-day period and the first
gap was not filled, the market is more likely to fall the next day and then trade sideways. If the first gap is
filled, then the SPY drop is often delayed. Also, if the SPY double gaps down and the first gap was filled, then
the market is more likely to rebound within four days.
Data and Methodology
To study more closely the gaps for individual stocks, we consider stocks included in the Russell 3000
between January 1, 2006 and December 31, 2010.1 During this time period, 20,611 gap ups occurred and
17,435 gap downs occurred. With 1,259 trading days in the sample, this is an average of about 16.4 stocks
gapping up and 13.8 stocks gapping down each trading day. Although some days, such as April 1, 2009,
which had 375 gap up stocks and February 17, 2009, which had 409 gap down stocks, have a much higher
observation of gaps, a typical day is characterized by at least a few gaps. Gap ups occurred on 1153, or 91.6%,
of the trading days. Gap downs occurred on 1033, or 88%, of the days. A gap of one variety or the other
occurred on 1164, or 92.5%, of the days.
The gapping stocks represented a wide range of companies. One-thousand-one-hundred-and-thirty-three
of the stocks in our sample experienced at least one gap up, and 1,135 experienced at least one gap down.
Throughout this study, we use “Day 0” to represent the day a gap occurs. For example, consider a gap
up. The day before the gap is Day -1 and the stock’s high on Day -1 is the beginning of the gap. On the next
day (Day 0), the stock’s low exceeds the high on Day -1. We base our return calculations from the open at
the next day (Day 1) to the close on Day 1 to calculate a 1-day return. To calculate longer returns, the return
1
To be included in this sample, a stock had to have a trading volume of over 1 million shares on the gap day and the four prior trading days to
ensure that decent liquidity existed.
Investing in SPY instead of the gap up stocks presented in Table 1 would have resulted in an average loss
of 0.06% on these days. Thus, the stocks that gapped up performed much better the day after the gap than did
the average stock in the market. If the gapping stock is held for 5 or 20 days after the gap, on average, the
return will be positive and higher than the market return. These results suggest that stocks that gap up do, on
average, outperform the market over the next several weeks.
A closer look at the data, however, reveals that these gains come from a subset of the stocks—those that
are characterized by a white candle on the gap day (such as Gap A in Figure 1). The results suggest that when
a stock gaps up and closes higher than it opens, this upward price trend will continue for the next few trading
days, leading to a profitable trading strategy. However, if the price gaps up, but the close is lower than the
open, even though the gap remains unfilled, don’t expect the upward price movement to continue. Stocks
exhibiting these black candlesticks on the day the gap occurs tend to have negative returns, and underperform
the market over the next several days.
Looking at the price movement on the day of the gap appears to help identify profitable trading opportunities.
What if this analysis is extended to looking at the price movement the day before the gap occurs? Table 2
presents returns broken down by Day -1 candle color. This table shows that a black candle on Day -1 followed
by a white candle on Day 0 is associated with above market returns.
2
Numbers in all tables throughout the paper are percentage returns. Thus “-0.056” in the table represents a 0.056% decline.
Table 6 shows the results for gaps occurring above and below the 10-day moving average of price. Gap
ups that occur below the 10-day moving average of price have positive market adjusted returns for the one-,
three-, five-, and 20-day time periods. This suggests that gaps occurring below a 10-day moving average are
breakaway gaps, beginning an upward trend; this is especially true for gaps that have a white candle on the
day the gap occurs. Gaps occurring above the 10-day moving average tend to have a below market return,
suggesting that these are exhaustion gaps.
These results suggest going short the day after a gap down, whether the candle is black or white, but only
for the next few days. The positive 5-day and 20-day price movements for the gap down stocks suggests that
the downward stock price movement is short lived, and being long these stocks several days after their gap
down is profitable.
Table 10 looks at this trending question a little more closely by considering the color of the candle the day
before the gap occurs as well as the day of the gap. These results suggest that the shorting strategy is most
profitable when a white candle on Day -1 is followed by a gap down. Surprisingly the strongest downward move
occurs when a white candle occurs on Day -1 and Day 0. In this case, a short strategy is profitable out to Day 5.
Tables 15 and 16 consider longer moving averages. These two tables suggest that the stocks that are already
relatively low in price (trading below their 30-day and 90-day moving averages) are the most profitable stocks
to short on a gap down for the one-day and three-day time periods. However, these stocks tend to reverse
direction and outperform the market at the five-day and 20-day horizons. Interestingly, stocks that experience
a gap down when trading above their 30-day or 90-day moving average tend to outperform the market by over
1.3% over the next 20 days.
Julie. R. Dahlquist, Ph.D., CMT Richard J. Bauer, Jr., Ph.D., CFA, CMT
Senior Lecturer Professor of Finance
University of Texas at San Antonio Bill Greehey School of Business
2
An Analysis with TIAA-CREF
Greg Kuhlemeyer, Ph.D., Carroll University
Robert Kunkel, Ph.D., University of Wisconsin Oshkosh
Abstract
This study evaluates whether there is a momentum trading effect in real estate funds by examining the
TIAA Real Estate Account. We employ three return moving averages to develop a technical signal to identify:
(i) momentum trading days when investors are invested in the real estate fund and (ii) non-momentum trading
days when investors are not invested in the real estate fund. Our tests show a strong momentum trading
effect where the mean daily return for momentum trading days is 0.269% and significantly greater than the
-0.465% mean daily return for non-momentum trading days. Conversely, when equities and bonds are tested,
we find no momentum trading effect. We believe the momentum trading effect in the real estate fund is a
function of its underlying assets of real estate properties. The market values of the properties are determined
at four discrete times a year by independent third party appraisers rather than daily market prices. Because
the underlying assets, real estate properties, are reevaluated much more slowly than publicly traded stocks and
bonds, the real estate fund is able to generate momentum trading benefits.
Momentum trading is typically perceived and based on price movements where the goal becomes “buy and
sell higher” versus the traditional investment adage of “buy low and sell high”. Stocks with momentum are
expected to exceed their moving averages and this is most commonly interpreted, in the most basic design,
as a “buy” signal. Momentum investors use a variety of different technical analysis measures for moving
averages based on their perception of the correct approach for the situation. There are two concerns with price
moving averages. The first concern is that price moving averages are not directly related to returns. This study
takes on a unique approach in the literature by utilizing return moving averages that more closely align with
traditional fundamentalist views of the risk-return relationship. The second concern is that price, and hence
price moving averages, do not convey the same relative information as returns and their associated return
moving averages. A narrowly moving stock price can move above or below the price moving average by $1
and it may convey little new information if the stock is trading at $400 or much more if the stock is trading at
$4. If any momentum statistic moves either above or below the momentum statistic moving average, then it
conveys the relative importance of that day’s new information. In the previous example, clearly a 0.25% move
for the $400 stock has much less informational content than a 25% move for the $4 stock. It is this study’s
contention that assets with return momentum will signal opportunities to generate returns that will exceed a
buy-and-hold strategy. The ability to find a momentum trading effect in a real estate fund would also reject the
Efficient Markets Hypothesis (Fama 1970).
I. Literature
Technical and fundamental analyses have been part of the practitioner environment with fundamental
analysis generally supported by scientific modeling. Levy (1967) showed that a portfolio of well-performing
stocks with relative price strength would continue to perform well whereas a portfolio of poor-performing
stocks would continue to perform poorly. This study showed that technical analysis generated benefits.
Fama (1970) helped push technical analysis into the background of the academe with his Efficient Markets
Hypothesis in its three forms: weak, semi-strong, and strong. The implication is that if the Efficient Markets
Hypothesis holds in all forms, then neither fundamental nor technical analysis is valuable. Yet, there continues
B. Moving Averages
While previous researchers have used many moving average lengths, most moving averages have fallen in
the 5-day to 180-day range with a variety of weighting methodologies from equally to exponentially weighted.
We use three unique return moving averages of 5-, 30-, and 180-days. The 5-day window gives us roughly
a one week window into the very short-term momentum of the fund; the intermediate length 30-day return
moving average represents a one month window that captures a longer directional trend; and the 180-day
period captures two full appraisal cycles and is analogous to Levy’s (1967) 26-week period. Following Balsara,
Chen, Zheng (2009) we use the longer 180-day return moving average to capture long-term information and
informational content similar to a quarterly moving average, but believe that the short and intermediate terms
also convey information that can be readily captured in the period return moving averages.
Daily returns for each calendar day were calculated from TIAA’s daily unit prices with non-trading day unit
prices set equal to the last trading day’s unit price. To generate moving averages, we calculate the geometric
mean daily return for 5 calendar days (5GR), 30 calendar days (30GR), and 180 calendar days (180GR).
We then create a historical benchmark comparison utilizing 1,300 calendar days or approximately 43
months. The period length is long enough to generate a benchmark that would change very slowly and would
match a typical business cycle. Watson (1994) indicates that postwar business cycles have averaged just less
than 50 months versus slightly more than the average prewar business cycle of 25 months. Our period of 43
months was chosen as an approximate interim length leaning slightly towards recent trends. In addition, this
3 ½-year window provides a strong baseline moving average estimate that is not overly influenced by shorter
informational impacts and reduce trading. We then calculate return moving averages for 5 days, 30 days and
180 days as follows:
Where 5MA is the 5-day return moving average and 5GRt is the geometric mean daily return for 5 days.
Where 30MA is the 30-day return moving average and 30GRt is the geometric mean daily return for 30 days.
Where MS5 is the 5-day momentum statistic with a value of one when the geometric mean daily return
for the past 5 days (5GR) is greater than the 5-day return moving average (5MA), and otherwise the value is
zero. A positive indicator on the real estate account, for example, implies that the real estate fund is seeing
acceleration in the positive information from the appraisals of the underlying properties. Unfortunately, this
is too short to be relied on to make a trading decision when so many more appraisals are forthcoming and the
investor is limited in trading activity.
Where MS30 is the 30-day momentum statistic with a value of one when the geometric mean daily return
for the past 30 days (30GR) is greater than the 30-day return moving average (30MA), and otherwise the value
is zero. A positive indicator on the real estate account with this statistic indicates that over the past month the
appraisals have been coming in at a higher rate than expected with about one-third of appraisals completed. A
reasonable person might expect that this trend would continue.
Where MS180 is the 180-day momentum statistic with a value of one when the geometric mean daily return
for the past 180 days (180GR) is greater than the 180-day return moving average (180MA), and otherwise the
value is zero. In this situation we have gone through two complete cycles of appraisals within the real estate
account. The long-term trend is clearly very positive that appraisals are continuing to represent a return that
is above the historical norm.
The sum of the momentum statistics (5-days, 30-days, and 180-days) equals the total momentum statistic
as follows:
Where TMS is the total momentum statistic and may equal values of 3, 2, 1, or 0. When the TMS equals
3, this is an initial BUY signal where the investor would move funds into the corresponding equity, bond, or
real estate accounts. When the TMS equals 0, this is an initial SELL signal where the investor would move
funds into the money market account. When the TMS equals 1 or 2, this represents a HOLD signal where the
investor moves no funds.
Let us briefly think about what a BUY signal means in regards to the real estate account. The investor in
this situation is seeing that returns have accelerated above their historical moving average in the short-term,
intermediate and longer-terms. The implication is that the information content here is that over the last 6
Where Rt is the return on the account for day t, α is the intercept representing the arithmetic mean return
for the non-momentum trading days, β is the difference between the mean return for momentum trading days
and mean return for non-momentum trading days, DTMP is a dummy variable (1 = momentum trading day, 0 =
non-momentum trading day) and εt is the error term. The F-value tests whether the difference between mean
return for momentum trading days and mean return for non-momentum trading days is significantly different
from zero.
The intuition behind the regression model is very straight forward. A negative α coefficient would indicate
the model has identified non-momentum trading days that, on average, generate a negative return. A significant
t-statistic associated with the α coefficient indicates the negative mean return for the non-momentum trading
days in the real estate account is significantly different than a mean expected return of zero. A positive β
coefficient would indicate the model has identified the real estate account’s momentum trading days generate
a greater return than the non-momentum trading days. A significant t-statistic on the β coefficient would
indicate the mean return for momentum trading days is significantly greater than the mean return of the non-
momentum trading days. If the F-test is significant, then this indicates the regression model is statistically
robust and highly unlikely to have occurred by chance.
To test for a momentum trading effect, we estimate the regression model in Equation 8 to test whether the
difference between mean momentum trading return and mean momentum non-trading return is significantly
different from zero. The results of the regression are reported in Table III. When we analyze the Real Estate
Account we find a significant difference between the mean daily return for the momentum trading days and
non-momentum trading days. We find the difference between the mean daily returns of the momentum trading
days and the non-momentum trading days to be 0.073% which is significant at the one percent level. To the
average investor, this means that with minimal trading you can increase your daily return by 4.65 basis points
by being out of the account when one expects it to decline in value. Over the course of an entire trading year
you can improve your performance, on average, by over 2% without incurring any additional trading expenses.
Note that our evaluation window covers over a decade and this can have a significant impact on your overall
performance as an investor. You may also note that the equity account shows a similar benefit but it is not
considered significant. Equities have a much more volatile pricing regime meaning that the returns have a much
greater standard deviation, almost eight times greater, and that the result in equities could have reasonably
occurred by chance. With the real estate appraisal process the pricing changes occur much slower resulting in
a much smaller underlying volatility and making it highly improbable that these results could occur by chance.
As such, the regression model clearly shows there is a momentum trading effect in the Real Estate Account.
As expected, the Equity Index and Bond Market Accounts show no momentum trading effects as there are no
significant differences between the mean daily returns of the momentum and non-momentum trading days. In
other words, the β coefficients are not statistically different enough from 0 to be considered significant.
Endnotes
According to the prospectus for the TIAA Real Estate Account, TIAA limits trading to a single move out of
the account once per quarter via a telephone conversation with a firm advisor. According to the prospectus for
the CREF accounts, with the exception of money market, CREF limits the participants from going out-in-out
of the fund within a 60-day window. If the participants violate this restriction, then they are precluded from
making moves for 90 days. Note that this excludes transfers via mail. Therefore, mail requests are honored
without restriction as indicated above.
References
Appel, Gerald, 1979, The Moving Average Convergence-Divergence Method, Signalert Corporation, Great
Neck, NY.
Balsara, Nauzer, Jason Chen, and Lin Zheng, 2009, Profiting From A Contrarian Application of Technical Trading
Rules in the U.S. Stock Market, Journal of Asset Management 10, 97-123.
Bettman, Jenni L., Stephen J. Sault, and Emma L. Schultz, 2009, Fundamental and Technical Analysis: Substitutes
or Complements?, Accounting and Finance 49, 21-36.
Brinson, Gary P., L. Randolph Hood, and Gilbert L. Beebower, 1995, Determinants of Portfolio Performance,
Financial Analysts Journal 51, 133-138.
Fama, Eugene F., 1970, Efficient Capital Markets: A Review of Theory and Empirical Work,
Journal of Finance 25, 383-417.
Gordon, Myron J. and Eli Shapiro, 1956, Capital Equipment Analysis: The Required Rate of Profit,
Management Science 3, 102-110.
Graham, Benjamin and David L. Dodd, 1934, Security Analysis, McGraw-Hill Book Company, New York, NY.
Grinblatt, Mark, Sheridan Titman, and Russ Wermers, 1995, Momentum Investment Strategies, Portfolio
Performance, and Herding: A Study of Mutual Fund Behavior, American Economic Review 85, 1088-1105.
Jegadeesh, Narasimhan and Sheridan Titman, 1993, Returns to Buying Winners and Selling Losers,
Journal of Finance 48, 65-91.
Kirkpatrick, Charles. D. and Julie R. Dahlquist, 2010, Technical Analysis: The Complete Resource for Financial
Market Technicans, FT Press, Upper Saddle River, NJ.
3
Evidence from the Components of the
Dow Jones Industrial Average
Camillo Lento, PhD, CA, CFE
Abstract
This study develops new insights into the profitability of trading rules through a synthesis of fractal geometry
and technical analysis. The Hurst exponent (H) emerged from fractal geometry as a means of detecting long-
term dependencies in a time series, the same dependencies that technical analysis should be able to identify and
exploit to earn profits. Two tests of this synthesis are conducted. First, financial time series are classified into
three groups based on their H to determine if a higher (lower) H results in higher returns to trending (contrarian)
trading rules. Second, the relationship between H and profits from technical analysis are estimated through OLS
regression. Both tests suggest that the fractal nature of a time series explains a significant portion of the profits
from technical analysis.
I. Introduction
Technical analysis is a broad discipline that analyzes past price and volume data to identify patterns that
predict future price movements. Identified patterns provide the basis for technical trading rules, which generate
buy and sell signals. The efficacy of technical analysis has been researched extensively. The research results are
mixed, providing support for (e.g., Brock, Lakonishok and LeBaron (1992), and Gençay (1999)), and against
(e.g. Allen and Karjalainen (1999), Lo, Mamaysky and Wang (2000), and Bokhardi et al. (2005)) technical
analysis’ ability to forecast security returns.
Recently, Hurst’s exponent (H) (Hurst, 1951) has emerged from fractal geometry into economics research
as a means of classifying a time series based on its long-term dependencies (Peters, 1991 and Peters, 1994). A
value of H of 0.50 indicates that a series exhibits Brownian motion.1 0<H<0.5 indicates an anti-persistent series,
suggesting that the data set exhibits mean-reverting tendencies. 0.5<H<1 indicates a persistent series, suggesting
the data is trend reinforcing. The strength of the trend increases as H approaches 1. The H thus provides a method
of classifying time series, which may be beneficial for the discipline of technical analysis.
The purpose of this study is to develop new insights into the discipline of technical analysis through a synthesis
with fractal geometry. The synthesis posits the following: fractal geometry provides a technique (H) that detects
long-term dependencies (reinforcing or reverting trends) in the historical price data of a time series; these are the
same trends that technical analysis purports to identify and utilize to predict future price movements. Therefore,
trending trading rules should be more effective on trend-reinforcing time series, while contrarian trading rules
should be more effective in anti-persistent, or mean-reverting, markets.
It is important to note that the discipline of technical analysis includes various technical trading rules. In
addition, the technical trading rules can be specified differently, leading to a large number of trading rule variants.
Accordingly, popular trending and contrarian rules will be used to test the synthesis. Specifically, trending
Acknowledgements: The author would like to thank the anonymous reviewers for their comments that helped improve
this manuscript.
1
Brownian motion is a mathematical concept used to describe a random process.
A synthesis of technical analysis and fractal geometry can provide fertile ground for new theory development,
along with novel empirical tests, to enhance our understanding of the profitability of technical trading rules. The
following literature review develops this synthesis. Section A discusses the technical analysis literature, Section
B discusses fractal geometry, and Section C develops the synthesis and Hypotheses.
A.Technical Trading Rules
Early studies on technical analysis conducted by Alexander (1961 and 1964) and Fama and Blume (1966)
suggested that excess returns cannot be realized by making investment decisions based on filter rules. However,
Sweeney (1988) re-examined the data used by Fama and Blume (1966) and found that filter rules applied to
15 of the 30 Dow Jones stocks earned excess returns over buy-and-hold alternatives. Technical trading rules
have also been extensively tested in the foreign exchange market (Dooley and Shafer (1983), Sweeney (1988),
and Schulmeister (1988)).
The number of studies on technical trading rules significantly increased during the 1990s. Some of the most
influential studies that provide indirect support for trading rules include Jegadeesh and Titman (1993), Blume,
Easley, and O’Hara (1994), Chan, Jagadeesh, and Lakonishok (1996), Lo and MacKinlay (1997), Grundy and
Martin (1998), and Rouwenhorst (1998). Stronger evidence can be found in the research of Neftci (1991), Neely,
2
Brownian motion is a well-known paradigm in finance that can be described as a white-noise process for which the
independent increments are identically normally distributed. A white-noise process is the statistical paradigm against
which the sequence of increments from a chaotic dynamical process is typically contrasted. White noise traditionally
refers to a sequence whose increments are independently and identically distributed with zero mean and finite variance.
Brownian motion underlies most of modern finance theory’s most important contributions.
3
Mandelbrot (2004) provides a detailed history and discussion of the Hurst exponent, the Joseph effect, and Noah effect.
4
Additionally, the power spectrum of the increments of fractional Brownian motion is proportional to fβ,
where β= 2H-1, so that Brownian motion as a white-noise paradigm has a flat spectrum (Feder, 1988).
H and the profits from the trading rules are calculated for all 30 DJIA stocks (as of July 2008) for the 10-year
period of July 1998 to July 2008. The time period was selected as it provides enough data to calculate both the
H and the technical trading rules and ends just prior to the 2008 credit crisis. As the data set includes many
differing events (e.g., Long-Term Capital Management issues and the Asian crisis in the late 90s, dot-com bubble
in early 2000s, etc.), sub-period analysis is conducted to determine the sensitivity of the results to the time period
selected.
Trading rules can be calculated at various data frequencies. The data frequency selected depends on different
factors and preferences. This study utilizes daily and weekly data. Daily data is used because a typical off
floor trader will most likely use daily data (Kaastra and Boyd, 1996). Furthermore, intraday time series can be
extremely noisy. Along these lines, weekly data is also used as it is readily available to all traders.
The use of raw daily price data in the stock market has many problems, as movements are generally non-
stationary (Mehta, 1995), which interferes with the estimation of the H. The market-index series are transformed
into rates of return to overcome these problems. Given the price level P1, P2, … Pt, the rate of return at time t
is transformed by:
Table 1 - Sample data and descriptive statistics (July 22, 1998 to July 22, 2008)
IV. Methodology
The individual and average profits from 12 trading rules, along with H, are calculated for all 30 stocks.
Profitability is defined as the returns from the trading rules less the buy-and-hold strategy returns, adjusted for
transaction costs. Therefore, by definition, profits can also be negative.
A.Trading rules
The trending trading rules are the moving-average crossover rule (MACO), the filter rule, and the trading
range break-out rule (TRBO). A MACO rule attempts to identify a trend by comparing a short moving average
to a long moving average. The MACO generates a buy (sell) signal whenever the short moving average is above
(below) the long moving average. This study tests the MACO rule based on the following signals:
where Pt is the stock price at time t. Again, these are the same TRBO rules tested by Brock, Lakonishok, and
LeBaron (1992). In order to maintain consistency, all of the contrarian trading rules are defined in the exact same
fashion as in Brock, Lakonishok, and LeBaron (1992).
The contrarian trading rule will be represented by a variant of the Bollinger Bands (BB). Bollinger Bands
are a tool that can be used in a variety of trading rules, not necessarily contrarian in nature. However, this study
will employ a contrarian strategy using Bollinger Bands in an attempt to profit from anti-persistent trends in
financial time-series. Bollinger Bands require two parameters: the moving average and the standard deviation.
Traditionally a 20-period moving average is used. The BB strategy tested will generate a sell signal when the
price of the security exceeds the 20-day moving average plus two standard deviations (i.e., the market is said to
be overbought). A buy signal is generated when the price of the security is less than the 20-day moving average
minus two standard deviations (i.e., the market is said to be oversold). This strategy is denoted by BB(20,
2). In addition to using the traditional parameters, two variants are tested: BB(30,2) uses a 30-day average
to determine whether information from longer time frame can generate more informative signals. BB(20,1)
uses the traditional 20-day moving average, but uses only +/-1 to generate signals to determine whether a
narrower band can generate more precise signals. These are the same BB parameters tested in prior literature
(e.g. Lento, Gradeojevic and Wright, 2007).
The statistical significance of the trading rules is determined through a bootstrapping methodology as developed
by Levich and Thomas (1993). The bootstrap approach does not make any assumptions regarding the distribution
of the generating function. Rather, the distribution of the generating function is determined empirically through
numerical simulations. The data sets of raw closing prices, with the length N + 1, correspond to a set of log price
changes of length N. M = N! separate sequences can be arranged from the log price changes with a length of
N. Each element of the sequence (m = 1, …, M) will correspond to a unique profit measure (X [m, r]) for each
variant trading rule (r for r = 1, … , R.) used in this study. Therefore, a new series can be generated by randomly
rearranging the log price changes of the original data set.
This simulation can generate one of the various notional paths that the security could have taken from time
t (original level) until time t + n (ending day). The simulation process of randomly mixing the log price returns
with H [0, 1]. The Brownian motion is then the particular case where H = 0.5. The exponent H is called the
Hurst exponent. The H measures dependencies in time series’ non-stochastic motion and is calculated through
rescaled range analysis (R/S analysis). For a time series where X = X1, X2, …, Xn, R/S analysis can be calculated
by first determining the mean value m, followed by the mean adjusted series Y:
The H can be estimated through ordinary least squares regression. Between Log(N) and Log (R/S).
Figure 1 graphically presents the rescaled range analysis that is used to estimate the Hurst exponent on the Dow
Jones Industrial Average time series.
The H has been estimated on all thirty Dow components. Table 2 presents the descriptive statistics for the H
calculated on all thirty time series.
There are various scholars who rebut the ability of H to identify long-term dependencies by arguing that the
rescaled range analysis is skewed. Specifically, issues with the sensitivity of the H to short-term memory, the
effects of pre-asymptotic behavior on the significance of the H estimate and the problems with structural changes
(e.g., Ambrose et al. (1993), Chueng (1993), Jacobsen (1996)) have been raised. The most significant rebuttal
was offered by Lo (1991), who reports that the rescaled range analysis could confuse long-term memory with
the effects of short-term memory. However, since Lo’s publication, many economists have reported that his
tests were potentially flawed (Mandelbrot, 2004). Additionally, a number of new contributions have suggested
alternative techniques of estimating a pathwise version of H to eliminate the lack of reliability in the H (Bianci,
2005 and Carbone et al., 2004). Recently, Qian and Rasheed (2004) concluded that the H provides a measure for
predictability, especially for H that are larger than 0.50.
where Profitsi represents the returns in excess of the buy-and-hold trading strategy for DJIA component i, and Hi
represents the Hurst exponent for DJIA component i. The intercept is expected to be positive for Hypothesis 1
and negative for Hypothesis 2.
V. Results
A. Profits from Technical Analysis on the DJIA components
The profits from the technical trading rules and the H for each DJIA component are presented in Table 3 with
daily data and Table 4 with weekly data. Calculated with daily data, H ranges from a low of 0.452 to a high of
0.587. Weekly data result in a wider range of H (0.431 to 0.629).
The trending trading rules were profitable for 7 of the 30 components when calculated with daily data and on
13 of the 30 components with weekly data. It is interesting to note that the MACO and TRBO were much more
effective than the filter rules. The filter rules generated an average negative profit of 10.59% (daily) and 22.6%
(weekly). The filter rules’ poor performance is consistent with prior studies (Szakmary, Davidson, and Schwarz,
1999; Wong, C., 1997; Nelly and Weller, 1998).
The contrarian trading rules (BBs) generated average profits of 2.19%, 3.37%, and 2.52% for all 30 stocks
with daily data. The BBs generated profits on 23 of the 30 stocks with daily data. The results with weekly data
are more volatile, with average profits of 15.5%, 15.2%, and 17.0% from the trading rules, with only 21 of the
30 stocks being profitable.
B. Hurst Exponent and Profits from Trending Trading Rules (Hypotheses 1)
Hypothesis 1 postulates that stocks with higher H should yield higher profits from trending trading rules. To
test this relation, each DJIA component was grouped according to its H value. The first group includes stocks
with values of H less than 0.5. The second group includes stocks with a value of H that is greater than 0.5 but
less than 0.55. The final group is for stocks with a H greater than 0.55.
Table 4: Hurst exponent and technical trading rule profits with weekly data (percentage)
The results provide strong evidence that profits from trending trading rules are partially explained by the long-
term dependencies, as identified by the H estimation. All three panels reveal that profits are lowest for stocks
with values of H that are less than 0.5 and increase in H. The trading rules earned returns of 11.5% in excess of
the buy-and-hold strategy for stocks with a value of H greater than 0.55, while technical analysis underperformed
by 16.7% for stocks with values of H less than 0.5.
The robustness of the results is tested though sub-period analysis. The H is estimated over three sub-periods
(n.b.: the sub-periods divide the data set into three equal periods) and the profits from the technical trading
rules are also calculated on the same sub-periods (untabulated). The results of the sub-period analysis for the
association between H and the profits generated by technical analysis are presented in Table 6. The sub-period
analysis confirms the robustness of the results in Table 5, as profits are higher for stocks that exhibit long-term
dependencies as identified by the H. Sub-periods 1 and 3 exhibit consistent patterns of increasing returns in
conjunction with increasing H; however, the weekly data on sub-period 2 do not reflect the synthesis.
These results are consistent with the synthesis and tend to corroborate the proposition that a value of H less
than 0.5 exhibits anti-persistent trends that limit trending trading rules’ ability to identify patterns, whereas time
series with a value of H greater than 0.55 exhibit persistent trends that are identified by the trending trading rules
to earn profits. However, the results partially explain, as opposed to completely explaining, the profits from the
trading rules because there are anomalies. For example, Pfizer’s weekly time series exhibited an anti-persistent
nature (H of 0.431), yet technical analysis was able to earn an average profit of 14.2%.
In addition to the sub-period analysis, additional sensitivity testing is conducted on the categories selected
for the analysis. The tables have been recalculated with only two categories: 0.5 >H and 0.5 <H. The results
(not presented) are the same as the main results in Table 5. In addition, the tables have been recalculated with
the following three categories: (1) 0.5 >H; (2)0.5 <H< 0.54; and (3) 0.54<H. This categorization results in 10
observations, or one-third, in each category. Again, the results (not presented) are the same as the main results
in Table 5. The results also hold with the following classification: 1) 0.5 >H; (2)0.5 <H< 0.52; and (3) 0.52<H.
Accordingly, the results are robust to the boundaries selected for the categories. Note that there are not enough
The results are consistent with and corroborate the results presented in Tables 4 and 6, and further support the
synthesis in Hypothesis 1: H is able to identify long-term dependencies in time series data that are exploited by
technical trading rules to generate profits. The estimation has an R2 of 31%, with virtually all of the explanatory
power resulting from the H variable. Panel B presents additional insight, revealing the resulting R2 of 37% when
using the MACO and TRBO proxies for profits. The estimation with the filter rule as a proxy for profits does not
yield strong results. This is a function of the aforementioned lack of profitability for filter rules. The sub-period
analysis conducted in Section 5.3 provides further data to test the robustness of the estimation in Equation 12.
The results of the estimation with the sub-period data are presented in Table 8.
Table 8 - Estimation results from regression of profits from trending trading rules on H with
sub-period data
The results provide evidence that profits from contrarian trading rules are partially explained by the long-term
anti-dependencies in a time series. Panel A reveals that profits are highest for stocks with values of H that are less
than 0.5 and decrease in H. Contrarian trading rules were able to earn returns of 11.3% in excess of the buy-and-
hold trading strategy for stocks with values of H less than 0.5. The results from Panel B (daily data) and Panel
C (weekly data) are not as consistent.
Again, regression analysis between H and profits from contrarian rules is conducted. The results of the
regression of Equation 12 are presented in Table 10. The results are consistent with Table 9, as the regression
provides some evidence of the synthesis in Hypothesis 2. The estimation produces a low R2 of 5%. However, the
intercept for the H variable is negative and significant at the 10% level.
Overall, the empirical evidence provide some support for the acceptance of Hypothesis 2, as contrarian trading
rules appear to be more profitable on stocks with lower values of H.
Table 11 - Estimation results from regression of profits from technical analysis on lagged H
Table 12 reveals that the trading rules are more profitable when calculated with weekly data, as 201 of the 360
variants (30 stocks x 4 trading rules x 3 variants of each rule) are profitable, or 55.8%, while only 161, or 44.7%,
variants are profitable with daily data. As discussed earlier, all stocks that result in profits from technical analysis
when calculated with daily data are also profitable with weekly data. However, the weekly data result in much
more variation in the profits, made evident by wider ranges and higher standard deviations.
References
Alexander, S., 1961, Price Movements in Speculative Markets: Trends or Random Walks, Industrial Management
Review 2, 7-26
Alexander, S., 1964, Price Movements in Speculative Markets: Trends or Random Walks, No. 2, in P. Cootner(ed.),
The Random Character of Stock Market Prices (MIT Press, Cambridge, MA)
Allen, F., and R. Karjalainen, 1999, Using Genetic Algorithms to Find Technical Trading Rules, Journal of
Financial Economics 51, 245-271
Camillo Lento earned a Ph.D. degree in accounting from the University of Southern
Queensland, Queensland, Australia, in 2012. Currently, he is an Assistant Professor of
Accounting at the Faculty of Business Administration at Lakehead University, Thunder Bay,
Ontario, Canada. Camillo also holds a MSc. and HBComm. Degree, along with being a
Chartered Professional Accountant (Canada), a Chartered Accountant (Canada), and a
Certified Fraud Examiner.
4
From Observations, Logic, and Perseverance
Abstract
In September 1927, the Cleveland Trust Company published the first Advance-Decline Line originated by
Leonard P. Ayres and his assistant James F. Hughes. For the next three decades, Hughes applied and expanded
the count of the market while adding the Advance-Decline Ratio to the group of market breadth indicators.
After 1958, Richard Russell popularized the innovative work of Ayres and Hughes. Today, these indicators
have wide following.
I. Introduction
In September 1927, the following chart appeared on the last page of the Cleveland Trust Company’s
Business Bulletin:
The Bulletin explained that the line in the top section “does not show the average prices of a group of
stocks, but rather for each day the preponderance of advances over declines, or of declines over advances,
among all the issues dealt in during that market session. It shows the changing daily trend of the market as a
whole.” (Business Bulletin, 1927) The chart showed the trend between January and August 1927. The other
two lines, correctly named, were a “six-day centered moving average” of the number of shares and issues
traded. (Business Bulletin, 1927)
Ayres was born in Niantic, Connecticut, on September 15, 1879.2 He attended public schools in Newton,
Massachusetts, and in 1902 graduated from Boston University. He taught English in Puerto Rico; in 1906 he
became the superintendent of the island’s school system. In 1910, Ayres received a Ph.D. degree from Boston
University.
1
Portrait is by permission of www.asapresidentialpapers.info hosted by National Opinion Research Center, University of Chicago, for the
American Statistical Association (www.amstat.org).
2
The biography is compiled from www.arligntoncemetery.net and http//ech.cwru.edu (The Encyclopedia of Cleveland History) (visited on Nov.
6, 2011).
3
Ayres’ headstone in Arlington National Cemetery reads “He Sought The Truth That It Might Make Men Free.”
recovered to September, fallen to November, and then risen until the end of the year “many individual stocks
did not follow this general trend.” (Business Bulletin, Jan. 15, 1924) Figure 4 recreates the table published in
the Bulletin. Figure 4 shows that among the 629 NYSE issues tabulated, 203 reached their highest prices in
the first quarter and their lowest prices in the fourth quarter. In the fourth quarter, 26 stocks made their highs
and also their lows for the year.
Figure 4 – Stocks Making Their Highs and Lows, Quarterly, 1923
4
Photograph is courtesy of Cleveland State University Libraries, The Cleveland Memory Project, http://clevelandmemory.org. The domed build-
ing was completed in 1908.
Adding the percentages of advances and declines revealed a deterioration that the individual percentages
did not show. During most of 1929 a “creeping bear market” had been hidden by certain stocks advancing
so much they carried the stock market averages up to new levels prior to the October Crash. In spite of new
high records for volume and the market averages (the Dow Industrials closed at its then highest of 381.17 on
September 3, 1929), a bear market had been in progress.
I have not found documents showing that in those early years Ayres and Hughes either cumulated advances
and declines or used ratios. The earliest reference to a cumulative advance-decline line I have found is 1948.
(Mindell, 1948) Ratios came later as a result of Hughes’ innovation. They used daily data. Their analysis
covered the broad market traded on the NYSE.
C. Charles H. Dow (1851-1902)
Dow Theory scholar late Professor George W. Bishop, Jr. noted that Charles H. Dow penned a Wall Street
Journal editorial on January 23, 1900, in which Dow observed, “Take last week for instance: There were
dealings in 174 stocks.… Of these 174 stocks, 107 advanced, 47 declined, and 20 stood still.” (Bishop, 1964)
Bishop did not give any other reference Dow made to advances, declines, and unchanged. Bishop was
noting that Dow discussed stock market “techniques credited to others at a much later date.” He cautioned
that by quoting from this editorial “we do not mean to imply that General Ayres did not arrive at the ‘count of
the market’ independently.” (Bishop, 1964)
My opinion is Dow made a commenter’s observation but cannot be credited with originating the advance
and decline indicators.
5
Hugh D. Auchincloss, Jr., (1897-1976), the stepfather of the late Jacqueline Kennedy Onassis, started the brokerage firm.
6
Colby (2003) calls a smoothed version of this indicator the Advance-Decline Non-Cumulative: Hughes Breadth-Momentum Oscillator.
7
In 1948, Hughes was recognized for “his original studies of Climaxes.” (Mindell, 1948)
V. Conclusion
Russell took the work of Ayres and Hughes to the front stage. Russell is the closing actor in this wonderful
story.
Leonard P. Ayres and James F. Hughes originated the Advance-Decline Line and published its first chart
in September 1927. Ayres had noted that the stock market is selective as market averages can trend in one
direction while the movement of some stocks will diverge. Not all stocks move together.
This observation led to the innovation that a technical indicator based on advances, declines, and unchanged
could more accurately time major turning points. Economic relationships in business activities can tell us a
reversal is ahead, but the Advance-Decline Line can tell us when the reversal will occur.
For the next three decades, Hughes continuously applied and expanded the use of advances, declines, and
unchanged while developing the Advance-Decline Ratio. Hughes believed that a breadth and price index have
so much in common that it is highly abnormal for them to show a protracted divergence in trend. An important
divergence between breadth and price results in a reversal of the market trend that created the abnormal
divergence.
These are the concepts underpinning the Advance-Decline Line and Advance-Decline Ratio that Ayres
and Hughes originated. Derived from statistics of market action, with their logic grounded on observations of
market behavior, these indicators continue to excel.
George A. Schade, Jr., CMT, has extensively researched the history of technical indicators
and written about their origins and development. He has sought to answer long standing
questions of attribution and continuity. He is a trial court judicial officer.
5
A Study of Various Market Thrust Measures
Wayne Whaley
Background
This research can be summarized in the five stages below.
1. Revisit the significance of Breadth (Adv/Dec) Thrust in launching major market moves.
2. Evaluate the utility of evaluating market momentum via measures other than Breadth.
3. Study the statistical significance of Reverse Thrust, commonly referred to as capitulation.
4. Study the significance of lack of Thrust signal sightings over extended periods.
5. Combine all the above research into an Intermediate time frame trading model .
My approach was to evaluate several measures of market momentum from several angles of tape activity
over time periods from 1 to 100 days and identify those that were highly reliable in signaling intermediate
(6-12 month) moves in the S&P 500. I studied Breadth (Advances vs. Declines), Up vs. Down Volume, Price
Change, Trin (Volume in Advancing vs Declining issues), Number of Issues making New 12 Month Highs and
12 Month Lows. Provided in this paper is a summary of what I consider the most interesting and statistically
Many market technicians use the ratio of Advances to Declines, but I prefer this percentage calculation
because it has some “Normal” distribution properties that can be useful in statistical analysis.
Table 1 below contains a summary of the average 252 day (appx 12 mt) S&P 500 move across the range of
levels of the Five Day ADT reading for data for the past 40 years (1970 through 2009).
The average annual return for the S&P 500 Index over the time period evaluated was 8.0%. ADT readings
in the middle of the range (25-70) appear to be statistical noise, but the further you get from the mean, the
more statistically bullish the measures become. Measures between 70 and 75 have a statistically significant
bullish bias (18.63%/year) above the norm, with readings above 75 having a perfect record for forecasting
higher equity prices 252 trading days ahead.
One’s first instinct might be to expect extremely low readings to have the exact opposite effect of extremely
high readings, but market bottoms generally occur after what is referred to as a selling capitulation or selling
climax. As you can observe, Five Day ADT readings between 20-25, appear to have a statistically significant
bullish bias (15.59%) as well, with extremely low readings of less than 20 leading to higher prices 12 months
later 100% of the time over the data set studied.
Rather than simply using one year as my exit rule and measure of performance, I discovered I had a
much more reliable short trade by going short the S&P on MHL measures above 1.80 and incorporating the
previously defined bullish thrust signals for trade exits. Table 8 shows the results of selling short the S&P 500
on a MHL reading of 1.80 or higher and holding the position for a minimum of two years (504 trading days)
or until a Bullish Thrust reading, whichever came first. The fourth signal on 800211 was the only signal to be
terminated via two year time expiration. The remaining six were terminated via one of our previously defined
initial 12 thrust signals.*
1 Thrust signals are a very important tool for gauging the potential for sizable intermediate market moves.
2. The well documented NYSE 10 day Advance/Decline Thrust indicator is still very reliable when triggered,
but did not give any signals from 1994 to 2008.
3. Other tape measures, besides breadth, can also yield additional insight into Market Thrust potential. In
particular, Up Volume vs Down Volume and simple Price movement.
4. All three measures (Breadth, Up vs Down Volume and Price) had a strong positive correlation with forward
intermediate market moves when observed at the 99.8% occurrence level.
5. Extreme occurrences of Reverse Thrust (capitulation) are very constructive for the market as well, and also
had a strong positive correlation with forward intermediate market moves when observed at the 0.1% level.
References