New Techniques  |  JAN 2009

Join The Band by Marco Alves

Join The Band by Marco Alves Apply this method to moving average crossovers to get rid of the lag and the false signals. One of the first indicators that any technical analysis novice studies is the moving average crossover. Moving averages (MAs) smooth a price series by determining the average closing price for a determined period (the last n bars) and, as a result, are lagging indicators, more suited to trending markets rather than rangebound ones. If two MAs of different periods are used together, a simple trading system can be built around it easily. Every time the shorter (faster) moving average crosses above the longer (slower) one, a buy signal is generated; a sell signal is produced when the faster average crosses below the slower one. As any technical analyst knows, these crossovers are prone to whipsaws; the price moves just enough in one direction to trigger a signal, then quickly changes direction, triggering an opposite signal. This causes early entries and exits that jeopardize trade performance (Figure 1). Whipsaws are the result of the sensitivity of MAs to data fluctuations. The classical approach to this problem has been to increase the averaging period (Figure 2) at a cost of increased lag, which, if too pronounced, may render the indicator useless. In addition, whipsaws tend to affect different time frames in a similar manner. For example, a set of two MAs on a daily chart will probably incur a similar frequency of false signals during the course of seven months (154 bars) as an equally parameterized set of MAs will sustain on a three-year weekly chart (3 * 52 = 156 bars). The annoyance frequency just moves to a larger time frame. So the problem lies in the concept: simple line crossovers must be replaced as signal generators by a different kind of triggering system.

by Marco Alves