Run variants side by side.

Compare up to five strategy variants on the same universe, period, and benchmark. The point is robustness — if one parameter swing destroys returns, you don't have edge, you have an artefact.