Bootstrap Monte Carlo Tests for Trading Strategy Robustness

Summary

This article uses Monte Carlo resampling to examine how an Expert Advisor’s historical trade outcomes might vary in future sequences. Its main method is a bootstrap: sample historical relative trade returns with replacement, generate many sequences, and calculate outcomes such as final profit and drawdown. The article also discusses approximating the return distribution with an empirical distribution, a simple two-outcome model, or a smoothed continuous estimate.

It proposes several ways to judge robustness, including comparing average simulated profit with its scatter, limiting drawdown through simulated termination, and testing whether early and late trade samples appear consistent. The examples and code are intended to support optimization in MetaTrader, but the approach assumes that trades are independent and drawn from a sufficiently representative distribution. Resampling past trades cannot reproduce every dependence or price-regime change, and the document notes that generating randomized price paths could give a more complete, computationally demanding analysis.

Key ideas

• The bootstrap method builds simulated trade sequences by resampling historical trade returns with replacement.
• A distribution of simulated outcomes provides more information about possible profit and drawdown than one historical result.
• The article proposes profit relative to scatter, drawdown-constrained profit, and distribution stability as optimization criteria.
• The basic model assumes independent trades sampled from a historical distribution, which may not reflect future market conditions.
• Randomized price-series testing could capture more strategy behavior but requires additional computation and tooling.

Tags

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