Testing Trading Strategies with Permuted Tick Price Series

Summary

The article describes a permutation test for assessing whether an Expert Advisor’s performance may arise by chance. It proposes testing the EA on an original price sample, then repeating the test on many altered versions and comparing a selected performance measure, such as net profit or Sharpe ratio. The fraction of permuted results that meet or exceed the original is used as an empirical p-value. The sample should preferably be out of sample, and the author recommends many repetitions for more stable comparisons.

To preserve the broad endpoint trend while changing the path, the method transforms tick values logarithmically, permutes price changes, and reconstructs the series. It focuses on bid, ask, and volume while retaining tick timing and flags; nonpositive values are handled to permit logarithms. The article outlines MQL5 tools for generating custom symbols and preparing repeated tests. This procedure can expose fragile performance, but it is computationally and storage intensive. Its conclusions depend on the permutation scheme and chosen statistic, and shuffled price paths may not retain all market structure relevant to actual trading.

Key ideas

• The test compares an EA’s original performance with results from repeatedly permuted price series.
• An empirical p-value is estimated from the share of permuted runs that equal or exceed the original result.
• The described shuffle permutes logarithmic price changes rather than raw price levels.
• Tick times and flags are retained while selected price and volume fields are reconstructed.
• Permutation testing requires substantial computing and storage resources, and its validity depends on the shuffle design.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.