the hAnDbook of technIcAl AnAlySIS8.3 Cycle Period, Multiple Timeframes, and Lagging Indicators
Oscillator Optimization via the Dominant half‐Cycle period
Oscillators and most overlays will normally tend to generate more effective buy
and sell signals if they are tuned to the dominant half‐cycle period. This will result
in the oscillator issuing buy and sell signals based on the wave cycle of interest
and not merely via its default lookback period, which may produce too few or too
many signals with respect to the wave cycle traded.
There are three ways to calculate the half‐cycle lookback period via any of the
following formulas, where N is the dominant cycle’s period (refer to Chapter 12
for examples of calculations for each formula):■ (^) (N+1)/2, and round up if N is even;
■ (^) (N/2) + 1, and round down if N is odd; and
■ (^) (2N+3)/4, round to closest integer.
For example, assume that we are interested in using the stochastic to generate
buy and sell signals. From Figure 8.16 we observe that the average dominant cycle
period on the daily GBPUSD chart, from 2005 to 2007, was on average, 75 days
(where a new cycle trough was formed at the same wave degree). The half cycle
would then be, using the second formula above, (75/2) + 1 = 38.5 days. We round
this down since N is odd (75), making the half cycle 38 days. We therefore tune
our stochastic’s lookback period to 38 periods. We see that it tracked and fore-
casted the cycle troughs perfectly, including troughs during the period after the
sample period was taken (out of sample data). Note that the standard 14‐period
fIgure 8.16 Cycle‐Tuned Stochastic Tracking the Cycle Troughs Effectively on the
Daily GBPUSD Chart.
Source: MetaTrader 4