To print article, click Print on your browser's File menu. Go back Posted 12/7/2000	sponsored by: The Speculator Why candlestick charts don't work Although we think it’s necessary to snuff out strategies based on candlestick and other technical charts, going against the trends scientifically can light a fire under your investment returns. By Victor Niederhoffer and Laurel Kenner As Leonardo da Vinci once wrote, no human investigation can be called real science if it cannot be demonstrated mathematically. Appearances and impressions don't fit the bill. Most of our friends and readers know we believe this, and so they wait to talk about candlestick charts until we're out of earshot. Money Plus. Easy online tools and free Bill Pay, too. Recently, however, one of our readers mistakenly sent us a private e-mail, meant for another one of our readers, extolling the merits of a new coffee-table book on candlestick analysis. This technique, for those unfamiliar with it, consists of the analysis of a series of bars that mark the open and close. From the top and bottom of the bars extend lines indicating the high and the low. The charts resemble a series of colored candlesticks, usually with wicks at both ends. "If you are not familiar with it already, you might enjoy 'Japanese Candlestick Charting Techniques,' by Steve Nison," the reader e-whispered to our resident behavioral finance expert, Brett Steenbarger, a gentleman known to be of broad mind and inclusive interests. "Candlestick charting dates from Tokugawa times -- late 16th, early 17th century. It predates the great Dawn-of-Modern-Capitalism blowouts that we all know and love -- tulip bulbs, Mississippi, South Seas. And it predates modern charting by several hundred years at least. "The book has pages and pages of descriptions of the psychology that the candlestick patterns illustrate -- e.g., head and shoulders, which the Japanese rice traders called ‘Three Buddha Top.' The one I like is ‘High Waves' -- makes me seasick just to think about it." After confiding his yen to explore such formations as "Doji Star," "Three White Soldiers Advance on the Dark Shadow" and "Candles Exhaust Themselves to Give Light to Men," the person concluded: "Much as I admire Vic and Laurel, for me it is going to take [a lot] to overturn 400 years of profitable tradition." The problem with charting is that 95% of it involves taking two points, drawing a straight line between them and then extending that line. If the point later in time is above the first, the line slopes up. If below, it slopes down. But where is the predictive value in such an extrapolation? Charting and romance A Japanese market philosopher, Shui Mitsuda, found an analogy between charting and romance: "It is quite easy to tell whether my girlfriend is angry or not. But it is impossible to tell when she will forgive me or if she will ever be crazy about me again, just from the past experience."
Sponsored Link Find "Japanese Candlestick Charting Techniques" at Barnes and Noble.	Mr. Mitsuda added: "By changing the range of variables in a chart (duration, price range), the graph often shows completely different faces. In the one-year range, Sony (SNE, news, msgs) is a strong buy. In the 10-year range, I get too scared. In the one-month range, I have no idea." Precisely. Subjectivity has its place in the selection of ties and deciding whether to decorate the home in French country or Barcelona modern. But, as reader John Lamberg writes, even when these subjectively interpreted visual patterns seem to work, they say nothing about the duration or magnitude of the subsequent move. Mr. Lamberg advised the reader not to part with more money for coffee-table books best used to level out uneven legs of the table. Camouflage The lowly puss moth caterpillar has in its arsenal dozens of methods of deception to escape predators. Its green color helps it hide among leaves. It can inflate a fold of skin around its head to make a scarlet "face" with two black spots resembling eyes. It can raise both ends and become boat-shaped. Two rear tails can protrude bright red whips that it lashes about, as it rocks its head from side to side. Why should not humans, who are at least as knowledgeable as caterpillars and have the benefit of reading the history of thousands of years of deception starting with the Trojan horse, be able at least to raise their level of trickery above that of extending a straight line? On Monday, the Nasdaq Composite ($COMPX) hovered near 2,500, down 50% from its yearly high and close to a two-year low. Before you could say "Jackie Robinson," it rose 10% on Tuesday in the biggest rally in its 29-year history. Extrapolations of straight lines somehow don't look so good when buying at the lows realizes such returns. If there is some predictive value to a chart, then the numbers that create the charts will contain that information. Science has always proceeded by trying to quantify rather than by relying on visual impressions and anecdotes, and we do believe that prices and volumes contain predictive significance. Rather than providing the valuable but destructive role of helping to bury charting analysis, we prefer to work with the highs, lows, opens and closes of the daily prices that go into the charts.
The way to make money in stocks is to do the opposite of what seems obvious -- and recently profitable. For instance, if selling rallies three days in a row works -- then it probably won't work the fourth day.	SATs, redux The approach we take is based upon a test-taking method pioneered by our friend Adam Robinson, co-founder of Princeton Review. It involves putting yourself in the mind of the men and women at the Educational Testing Service who generate the questions for college exams. They always begin the tests with easy questions where the obvious answers are correct, and move to tricky ones, where the obvious answers are wrong but close to the right answer. The way to handle those tricky questions is to avoid the obvious response. The way to make money in stocks is to do the opposite of what seems obvious -- and recently profitable. (For instance, if selling rallies three days in a row works -- then it probably won't work the fourth day.) It's at least as difficult to make money in the markets as it is to score high on the Educational Testing Service's exams. But instead of throwing you hard questions toward the end of an exam, the market likes to throw you tricky moves when you're off your guard in hopes that you will give the wrong answer and thereby provide contributions to the very high overhead of the market. On a long-term basis, the market loves to throw in a nice 10% up-move in the Nasdaq, for example, when everyone is bearish, as happened Tuesday. We are pleased by this 10% rally, as we have been predicting it for much too long a time. On a day-to-day basis, the market likes to throw moves at you that you are accustomed to answering the wrong way. Such trick questions can be tested, and in our own work we have tested about 1,500 of them, so as not to allow the market to make a monkey of us the same way it does so frequently to those who rely on visual impressions of charts. Take, for example, the 80 times in the last six years that the market has gone up 20 or more points in the S&P 500 ($INX). This corresponds to about a 2% rise. (We will use futures here in all our calculations because those represent prices that you can actually trade at and make or lose money at, rather than the slow-moving index which often lags behind real price moves.)
Recent Articles • A swingin' Gestalt theory to rock your stocks, 11/30/00 • Place a bet on strategic long-shot investing, 11/16/00 • Why high P/E stocks are good for you, 11/9/00 more...	On days after big rallies, slow-moving but wise-guy traders like to set themselves up for a decline. The reason is that there is a slight tendency for the market to decline on the days following big rises. Now, what the market likes to do on occasions like this is give these traders an easy profit the first time. Sort of like giving you a red herring. But after you've made money with that trade and you're ready to short the next time, the market likes to stick it to you. What's the answer? How to answer correctly in a situation like this? Try buying on the day after a big rally. Set your limit a point below the low of the previous day. Hold until the close. The expectation is the S&P will be up 1.5 points, which gives you a 60% chance of profiting. The odds that this would occur are only five times in 100. By applying such a procedure to your trades, you won't get rich. But at least you will be going against the natural tendency, that of making the mistakes that the market likes you to make. By using the highs, lows, opens and closes in such a way, you will be adapting a scientific approach that has the chance of getting you or your kids into graduate school. We realize that many people disagree with us about charts, and we invite comments so that we can learn from each other. We will respond to all e-mails. At the time of publication, Victor Niederhoffer and Laurel Kenner both owned candlesticks, but are ashamed to admit it. MSN Money's editorial goal is to provide a forum for personal finance and investment ideas. Our articles, columns, message board posts and other features should not be construed as investment advice, nor does their appearance imply an endorsement by Microsoft of any specific security or trading strategy. An investor's best course of action must be based on individual circumstances.