What I've Learned During 37 Years on Wall Street

Burrough's first computer - a Burroughs B205 (1960).

So, when I entered Ohio State University in 1965, I signed up for every course that in any way related to the stock market.

Eddie: Did you learn anything there that you would apply in your career?

Robert: Finance professor Leo Stone told my Securities Markets class, "Some people think technical analysis can give them an edge." That was all I needed to hear. Although he spoke these words with a somewhat skeptical tone (still favored by finance professors when they refer to technical analysis today), he offered it as a possible topic for the required term paper. I jumped at the chance.

Eddie: How did you go about it?

Robert: I spent about 2 months intensely studying Joseph E. Granville’s first book, A Daily Strategy For The Maximization Of Stock Market Profits, which was published in 1960, years before his later work on on-balance volume. Every day I mined The Wall Street Journal for my indicator data. All calculations had to be done by hand in those days.

Eddie: Did you actually attempt to apply the indicators to calling the daily market action?

Robert: Of course. You can’t track the indicators without calling the market! I calculated win/loss statistics for my daily predictions. It did not take long for me to start to get a feel for the day-to-day movements of the Dow Jones Industrial Average.

Eddie: What kind of indicators were used 36 years ago?

Robert: Many of the ones still used today, such as stock price indexes and comparisons between them, including divergences, volume, up-down volume, advance-decline line, overbought/oversold, new highs-new lows, moving averages, seasonal tendencies, sentiment, speculation, and various momentum oscillators. The 50-day and 200-day moving averages already were known and respected. Moving averages of different lengths indicate trends in multiple time frames, which is an important concept to understand for both investors and traders.

Eddie: Are there others that have fallen out of favor?

Robert: Yes, specialist, member, and odd-lot short sales ratios, margin debt, institutional cash levels, and Barron’s Confidence Index do not work the same way they used to due to structural changes in the market.

Eddie: Before personal computers, back testing must have been cumbersome.

Robert: Exactly. Back then, indicator research involved much hard and tedious work for very little return. Pencil and paper and punching keys on mechanical calculators were error prone, and it was hard to find your mistakes. In the 1980s, the PC revolutionized indicator testing and systems development.

Eddie: What work did you find immediately after studying finance in college?

Robert: I landed a rare position with a brokerage firm as a fundamental analyst trainee, which paid $600 a month. I could have made more in another industry, but I was focusing on my long-term career path. I crunched numbers in the back room for about a year before they allowed me to analyze 10-Ks, interview corporate management, and write original fundamental research reports.

Eddie: How long were you a fundamental analyst?

Robert: My career went very well for a few years, especially in the bullish stock market trend of 1971 into 1973, as the stocks I recommended went up nicely. But 1974 was THE defining moment.

Eddie: What happened then?

Robert: I had been assigned by my brokerage firm to fundamentally cover a mortgage REIT based in Houston, Texas, which was a building boomtown in 1973. Today, some might call it a bubble. This REIT had done very well, but as interest rates ran up steeply in the first half of 1974, I questioned the CFO about the spreads between his borrowing costs and loan returns. When the term structures of assets and liabilities are mismatched, that means risk The CFO assured me that his company was very safely structured and hedged. Unfortunately, his strategy turned out to be inadequate for the steep rise in interest rates that was to come. Every time the stock broke down below a chart support level, I phoned him. But it wasn’t until the stock had dropped in half that he finally admitted that management had lost control of their borrowing costs.

Eddie: Did that shake your confidence in talking to management and pouring over their reported numbers?

Robert: Clearly, corporate managers don’t always know what they are doing, they can’t predict the future, and they don’t always tell the unbiased truth. Neither do reported numbers, which are always late and often manipulated. From 1974 on, technical analysis was the ONLY unbiased truth for me. All my experience told me that it worked, while I could clearly see that fundamental analysis could run off the tracks in a major way.

Eddie: But weren't most investors still biased toward fundamental analysis?

Robert: They still are. Most investors feel that there is common sense logic to fundamentals. They think they understand buy low and sell high. In reality, most wind up buying high and selling low. And when they try to buy on the cheap, they find that the cheap get cheaper, because cheap stocks are down for good reasons. The valuation is not your friend. Rather, the trend is your friend.

Eddie: So in 1974, you made a radical and complete transition from fundamentalist to technician. How did you arrive at the level where you felt confident enough write such an authoritative book on technical analysis?

Robert: While working in Smith Barney’s technical research department, Alan Shaw, Ron Daino, and I were asked to review 30 technical analysis software programs for publication in PC Magazine in April 1986. That is how I met Tom Meyers, who was the consulting editor on the project. Tom already had written the first of its kind book about applying personal computers to technical analysis, so he knew his way around the world of publishing. Later, at a Computrac conference, Tom and I realized that we both wished we had one good book with all the indicators in it, complete with back tested proof of effectiveness, clearly showing which indicators were best and which ones didn't work. Nobody at that time had attempted such a big task. So, we published the first edition of The Encyclopedia of Technical Market Indicators in 1988.

Eddie: But Meyers did not work with you on the completely revised Second Edition, published in 2003.

Robert: By that time, more than a decade after the first one, Tom was too busy with other work, so I had to produce the new edition by myself. It was a lot of work, and this Second Edition is much more comprehensive than the first, with 820-pages of thorough indicator descriptions, formulas, and test results covering more than 100 indicators.

Eddie: Did you find indicators that worked well before the 1990s continued to perform well in 1990s and beyond?

Robert: Yes. For example, the very popular 200-day moving average continued to make money and is still relevant, whether you're a long- or short-term trader. As a general rule, your odds of success will be significantly greater if you only go long stocks that are above the 200-day moving average. That applies whether you are a swing trader or tend to hold stocks for many months. By doing this, you're keeping yourself on the side of the bigger trend, which often dominates and overwhelms trends in shorter time frames. I apply this myself in my own trading. I recall that Larry Connors in How Markets Really Work also found that there has been a real statistical edge to being a buyer when the market is above the 200-day moving average.

Eddie: Based on all that research you published in The Encyclopedia of Technical Indicators, Second Edition, what else can you tell us that will most help traders improve their trading performance today?

Robert: While I'm best known for my technical research and systems development, I'm also a trader and portfolio manager. I can tell you that, first and foremost, you've got to start with basic money management. I've boiled it down to a set of rules that I've found to be important based on practical experience. I believe that if you follow these rules, they'll save you from self-destruction and you will skew the odds in your favor. I posted the most important seven rules on my website, www.robertwcolby.com.

Robert: I've seen traders do a lot of wild things and then blame bad luck when they lose. You need to know that if you break these rules, you WILL pay the price, sooner or later. The first two rules, "preserving capital" and "trading in the direction of the larger trends", are the most important.

Eddie: These days, traders use software that contains well over a hundred indicators. It can be confusing. Which ones work and which ones don't?

Robert: In my book, on pages 32 to 35, there appears a list of 126 Comparably Measured Technical Market Indicators. On my website, I ranked them by “Annual Relative Advantage”, in order of their profitability, from best to worst. In order to measure profitability, I used the "signal event method", which identifies specific entry and exit signals for each indicator for different parameter settings.

Editor's note: According to Robert Colby, the numbers next to each indicator represent how the performance of the indicator compares with buy and hold. The calculation to derive this number is rather complex. But taking the first one as an example, Robert says that the Exponential Moving Average (5 days) is 766849 times better than buy and hold during the test period.

Copyright © 2000-2005 by www.robertwcolby.com. All rights reserved. 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.

Eddie: So, you used different look back periods for parameters. What about the default settings, such as the 14-period RSI that traders are accustomed to?

Robert: I took an entirely fresh look at each indicator, with no special consideration for the most popular default settings. I've always felt that default setting are like opinions and are subject to change. I back tested to find the parameters that worked best, and those parameters are the ones I published in my book.

Eddie: So, what you're giving traders is not only the best indicators but also the best settings to use for those indicators.

Robert: Right. In addition, I try to help traders understand of the strengths and weaknesses of different indicators. For example, oscillators such as RSI and stochastics were already oversold on the Friday before Black Monday, October 19, 1987. Anyone who bought that oversold condition lost big the very next trading day. I know traders who actually bought then based on some oscillator, and it was an extremely costly and painful experience for them.

Eddie: Now let's get into the top indicators list itself. After you did all that testing, which ones did you find worked the best?

Robert: Based on 101 years of daily closing prices for the Dow Industrials (DJX | Quote | Chart | News | PowerRating), short-term trend following maximized simulated profits. The top-performing indicator strategy, going both long and short, stop and reverse, that I found was a trend-following moving average crossover:

1. When the DJIA close today crosses above its 5-day Exponential Moving Average (EMA) calculated as of yesterday’s close, cover shorts and buy long at the current close today.

2. When the DJIA close today crosses below its 5-day Exponential Moving Average (EMA) calculated as of yesterday’s close, exit long and sell short at the current close today.

Eddie: That seems too easy.

Robert: Keep in mind that we are talking about my findings based on pure statistical research on historical data. There are other important considerations for actual trading. For example, if your trading costs are high or if you can’t execute an order precisely and consistently when the signal is given, then short-term systems are not practical for you. Also, just because short-term trend-following has worked best over the past 100 years, this does not mean that there have not been long periods of drawdown. In fact, the 5-day EMA crossover strategy has suffered a drawdown since 10/21/2002, shortly after the DJIA made a bear market low. Also, there was a 3-year drawdown from 3/22/91 to 3/24/94.

Eddie: Why do short-term systematic-trend following systems go "off the track" like that?

Robert: The cause is the lack of strength in the underlying trend. Trend following works well when prices are trending, either up or down. But choppy markets are never good for trend following. The whipsaws produce small losses that add up. For more than a year now, there has not been much of a trend for the major stock price indexes. Funds managed by some of the biggest systematic trend followers have lost more than 20% this year. Still, if you are nimble enough and can execute trades cheaply during the trading day, you can make a profit trading the short-term trends. Of course, that requires more effort and more finesse than the extremely simple end-of-day EMA crossover strategy we have been talking about.

Eddie: Are there any indicators among the ones you've ranked that have performed strongly in the past and have held up well in the present?

Robert: Short-term oscillators, such as CCI, RSI, and Stochastics, when used with longer time frame trend filters, such as a 200-day moving average, have continued to perform profitably over the most recent choppy years. Overbought and oversold works well most of the time in trading ranges.

Eddie: Can you give me a specific example, say on the very popular RSI?

Robert: I back tested RSI against the most recent past 23 years of S&P 500 stock index continuous futures contract end-of-day data, taking 70 as a sell and 30 as a buy, only taking signals in agreement with the 200-day simple moving average, and executing at today’s close with zero commission and slippage. All look back period lengths from 2 to 11 would have been profitable in my historical simulation. This suggests a robust indicator. But note that test results were poor when not filtered by the long-term trend (200-day simple moving average), so the filter is critical. A long-term trend permission filter eliminates many losing trades against the major trend.

Eddie: But in your book, don’t you caution traders about the dangers of using overbought/oversold oscillator signals?

Robert: Although OB/OS signals work well most of the time, they hurt traders at unusual times of big market moves, such as the Crash of 1987 and decline of 1998. Again, let me emphasize that acting on RSI OB/OS signals mechanically without filters can cause large drawdowns. I doubt that Welles Wilder (the developer of RSI) ever intended it to be used strictly mechanically.

Eddie: For your personal trading, what type of trader are you?

Robert: I do exactly what we have been talking about here. I am a systematic position trader of stocks. I don’t day trade. I often hold a position for weeks or months, trying to capture a big move. You need to ride some big moves to make trading pay. I trade filtered systems that buy when a stock dips to oversold and sell when it jumps up to overbought. I use trend filters, so I don’t take short-term positions that are against the main longer-term trend. I don’t try to finesse entries or exits when I get a signal. I don’t wait for short-term price confirmation when I enter or exit a trade on a signal. These bad habits can lead to hesitation.

Eddie: You have worked for some of the biggest names in the investment and trading business over the years. What services do you offer your clients?

Robert: I offer research consulting and systems development for professional investors and traders seeking objective, unbiased decision-making methods that have been thoroughly tested against actual historical market data. My company, Colby Research, has developed a multifactor model for stock selection and timing that generates unbiased and specific signals that can be applied to any universe of stocks or other financial instruments. This model objectively identifies trend changes as well as critical support and resistance levels where an instrument is likely to reverse. A completely independent and separate method ranks all financial instruments in any chosen universe of instruments by risk/reward probabilities, from best to worst. Also, I offer portfolio management to institutional and qualified investors. I invite readers to contact me through my website, www.robertwcolby.com, for details about my services.

Eddie: Sounds great. Well, Robert I want to thank you for joining us.

Robert: Thanks for having me.