Walk-Forward Validator

Whether your strategy's optimization process generalizes — i.e., does optimizing on one period produce parameters that work on the next period? The validator runs an automated walk-forward and reports per-window stability metrics.

The visualizer shows window-by-window detail interactively. The validator runs the full walk-forward and produces a pass/fail report. Use the validator for go/no-go decisions; use the visualizer for diagnosis when validation fails.

Three thresholds documented on the methodology page: (1) median OOS Sharpe ≥ 50% of in-sample Sharpe, (2) parameter stability (no parameter has more than 50% relative variation across windows), (3) no single test window destroys the lifetime P&L. All three must pass.

Minimum 5; ideally 8-10. With 5 windows you get a basic stability picture. 10+ gives confidence intervals on the OOS Sharpe estimate. The methodology page warns at 4 or fewer windows — too few to call the result robust.

If you tune until validation passes, you've used the validator as just another in-sample fit, which defeats the purpose. The validator should be a final gate, not an iterative optimizer. Use a held-out 'final' window that's never seen by the validator until the strategy is otherwise locked.