Monte Carlo simulation

One backtest is one draw from an unknown distribution. Monte Carlo simulation re-generates that history thousands of times — by resampling returns (bootstrap, block bootstrap) or by fitting a stochastic model (GARCH-filtered, Student's t, skewed-t) — and reads the resulting fan of equity paths: percentile bands, probability of hitting a ruin threshold, the depth of typical and extreme drawdowns.

Its honest use is at the size you actually intend to trade: tails scale non-linearly with leverage, and a strategy that survives 1 contract can carry meaningful ruin probability at 3. Its dishonest use is as decoration — a fan chart drawn from an overfit backtest simulates the overfit, not the strategy.