Friday, April 25, 2008

Strategy Test: Stochastics

In last week's strategy article I did some basic analysis on the MACD; a trend indicator. In this week's article I took a look at stochastics; a momentum indicator. However, unlike the traditional overbought/oversold triggers associated with using stochastics I have opted to use a cross of the mid-line (50) with a base fast stochastic setting of [39,1,1]. I couldn't test for the optimum [39,1,3] using the %K smooth as a trade trigger, so the generated signals were not as good as the basic slow stochastic signal I normally use (see chart for differences in the quality of the signals), but it did give an indication of the efficacy of this strategy.

Test period: A complete bull-bear cycle defined by the S&P (March 20th 2000 to October 8th 2007).


Number of shares: 100

Commission: $9.95 (included in the loss calculation)

Trades: Round-trip only; partial trades were excluded.

The basic settings on commission returned the following:

Total Profit: -$30,939
Winners: 256
Losers: 1011
Win percentage: 20%
Profit Factor: 0.74

Unfortunately, using the fast stochastic as a base for generating the signals produced a large number of whipsaw trades. Compared to the MACD strategy there were almost three times as many trades, which translated into three times the cost when commissions were factored. The huge loss would quickly break a trade account, so by itself the stochastic mid-line crossover system is not a recommended strategy.

The use of a protective stop failed to ease the pain; using a 4% stop reduced the loss to a (still) depressing -$23,968 off 26% winners, compared to the 20% winners of the no-stop strategy. Other tested stops didn't change the return over the non-stop strategy to any large degree.

Taking commissions out of the equation ($0) improved things a little, but there wasn't the swing to profitability as might have been hoped:

No stop: TP = -$5,276 on 23% winners
3% stop: TP = -$4,174 on 31% winners
4% stop: TP = $1,543 on 31% winners
5% stop: TP = -$7,122 on 29% winners
6% stop: TP = -$9,904 on 27% winners
7% stop: TP = -$9,430 on 26% winners
8% stop: TP = -$10,366 on 24% winners

Only the 4% stop strategy returned any profit (on 100 share lots with an average cost-per-share of $39.48). Stops above 5% gave less of a return than using no stop at all.

How does the strategy test for three random dates?

A start date was randomized and 1 year of data from each point was used. The randomization dats covered 2000/01, 2004/05 and 2006/07; a relatively good mix of bullish and bearish periods. The system was an across the board disaster in 2000/01 with 1-year of trading producing staggering losses ranging from -$7,612 (4% stop) to -$22,975 (no stop). The 2006/07 period produced the best returns from +$1,738 (8% stop) to +$6,036 (4% stop). The average return is given in the chart below:

Not surprising, the 4% stop loss proved to be the most effective strategy to employ - although none were ideal. No stop strategy brought the win percentage to the 40%+ area needed for most profitable systems.

Dr. Declan Fallon, Senior Market Technician, the free stock alerts, market alerts, and stock charts website