In a sideways period when both trend following and swing-trading techniques generate mostly losses, the best strategy is to stay out of the market altogether and wait for some kind of consistent move to develop. The Euro went through such a period in summer 2004. After making a low in May around 1.1750, the Euro made a high in July at 1.2460. The 20-day moving average shows there was an underlying upward bias, but to trade directly on the 20-day moving average would have resulted in overtrading and losses — notice how many times the prices crossed the moving average. If this is a trend, it’s not a tradable one using a trend-following technique.
Although you could have used a swing-trading tool such as the RSI and not done too badly during this period, gains were not large and the risk was very high. There are no fewer than 53 changes in direction between the end of May low (where the horizontal line begins) and the upside breakout point over the topmost horizontal line in October. Even if you came out ahead, your nerves were shot In it’s top window is a nifty indicator invented by Welles Wilder called the average directional movement index (ADX). The purpose of the ADX was to measure the strength of a trend, but it just as easily indicates the absence of a trend when its readings are very low — in this case, readings under 15.
The ADX is a real pain to calculate, and fortunately most charting software packages do it for you. The principle is to calculate an “up directional movement index,” called +DI, which is a moving average of closes on days when the high is higher than yesterday’s high. The default number of days in the moving average is 14, but you can vary this. It’s labeled +DI because it’s a positive number, capturing the amount of the bar that is above the trading range of the day before. This is consistent with the idea that a series of higher highs is an uptrend; +DI simply measures the force of the uptrend.
You also calculate a second directional movement index, the down version, or -DI, which is a moving average of days on which the low was lower than the day before. Again, a series of lower lows is a downtrend, and –DI measures the force of the trend. Obviously, every day’s bar can be only one or the other. When there’s an inside day — a bar with a lower high and a higher low than the day before there is no directional movement. An outside day, which has a higher high and lower low than the preceding day, also has zero directional movement.
It shows the ADX in the top portion of the chart. When it is rising, upward directional movement is overwhelming downward directional movement, and vice versa. Because ADX is a true index, it varies between zero and 100.
Research by Chuck LeBeau of www.traderclub.com shows that when the ADX rises above 15 from a flat spot (a “basing pattern”), a trend is being built. The ADX usually rises to no more than about 70, whereupon the slope of the index starts flattening out or falling and you consider the trend has stopped accelerating.
(This doesn’t necessarily mean you automatically exit, but it does mean you should raise your stop closer to the last close, since deceleration of the trend may foreshadow a reversal of the trend.) When the ADX is flat, you have no directional movement, either up or down. Higher highs are being washed out by lower lows in the moving average period. It’s no coincidence a flat ADX is seen when the average highlow range is contracting. For example, if in a trending period the average high-low range is 100 points and then contracts to 30 to 50 points, trendedness is decelerating. Conversely, when the average daily high-low range is rising from 35 to 50 points to 100 points, trendedness is accelerating.
The 21-day ADX approaches the 15 level in July, but then falls back. The indicator stays flat for the rest of the period and doesn’t rise above 15 until the breakout in October that takes the price over the July high. (You might also notice the flat spot features a head-and-shoulders pattern that was not a true headand- shoulders — the lowest low at the end of the pattern in August is not lower than the lowest low in the entire pattern, and therefore the pattern is not confirmed. If it had been confirmed, we would have expected the Euro to fall outside the bottom of the horizontal range, since the head-and shoulders pattern is a bearish pattern.) You would have done well to stay out of the market over this entire period whether you were using trend-following or swing-trading techniques.
You might have eked out a small profit, but in the absence of a clear direction, you were taking an unnecessary risk because you were trading all the time without a clear expectation of a positive result — something statisticians refer to as “positive expectancy.” You should never place a bet unless you have a better than 50-50 probability you will win.
In the absence of directional movement, how do you estimate the probability of any bet being a winning bet? In poker, you have to know when to fold ‘em, and in trading, sometimes the optimum thing is to do nothing. A flat ADX under 15 represents one of those times.
Subscribe to:
Post Comments (Atom)
Post a Comment