T-test evaluation of simulated trade returns

The simulated sample and its costs

The evaluation used 24 trades generated in a simulation of the September 1994 Deutsche Mark futures contract. Cost assumptions included an estimated 20-dollar round-turn commission and a stop-order slippage penalty equal to 10 percent of the daily range.

Net return is the dollar result of each completed trade after those estimated costs. The t-test was applied to each trade's net dollar return, beginning with the sample average return and the sample standard deviation.

How the t-score was judged

Hypothesis testing is a formal check of whether a sample statistic is large enough, relative to a tabled threshold, to reject a stated null in favor of an alternate claim. The null hypothesis is the default claim under test, typically that average net trade return is not distinguishable from no edge.

The t-score is a standardized measure of how far the average trade result sits from the null, using the sample's own dispersion and trade count. The sample standard deviation of those trade dollar returns was 1027.28. Sample standard deviation measures how widely individual trade dollar results spread around their average.

A statistical table supplied a critical value of 1.71 that the t-score had to exceed in order to reject the null hypothesis. A critical value is the tabled cutoff a t-score must exceed before the null is rejected at the chosen significance level. The computed t-score was 1.88, which exceeded 1.71, so the evaluation rejected the null hypothesis in favor of the alternate hypothesis.