TRADING TECHNIQUES Whipsaws, Linear Trends
And Simple Moving Averages by Robert Wayman
How does whipsaw magnitude affect a simple moving average system? Here's a common solution for the problems of the SMA.
Previously, we looked at cycles and found that there was a noise level-based spectral cutoff point, above which were tradable cycles. We also found that selecting a cycle to trade established the averaging period we need to use in any simple moving average (SMA) crossover trading system. This time, let's look at whipsaws and how whipsaw magnitude actually affects the profitability of a simple moving average system. Let's also take a look at a common fix for the problems of the SMA and try to draw some conclusions about the direction of effective system development.
ONE MORE HURDLE
Once we've selected the averaging period for the moving average crossover system, we only have one more hurdle to cross in our quest for profitability, but it's a big one: whipsaws. What level of whipsaw cycle magnitude will render an SMA crossover trading system useless? My moving average crossover system is based on a simple moving average, but the basic thrust of the derivation applies to any moving average system with a price path crossover characteristic. To see how whipsaws affect a simple moving average crossover system, we'll need a model
for the system. The place to start is by defining a model of the simple moving average driven by a sinewave style of
I won't list the formulas and derivation since they're somewhat complicated, but for those who are interested, the
derivation is composed of several steps. First, a simple moving average (SMA) formula is generated by convolving†
a sinusoidal price path with a rectangular window. The trading model is built by taking the difference between the
SMA output and the price path. A buy signal would occur when the SMA crosses below the price path and a sell
signal would occur when the SMA crosses above the price path. Buy when: