Show how sensitivity analysis is applied

Sensitivity analysis checks how a backtest or simulation result changes when inputs or assumptions change. Analysts vary one choice at a time, such as the look-back window, rebalancing frequency, transaction cost level, factor definition or distribution, and see whether conclusions hold. Net return after costs, for example, falls by turnover times cost per unit traded, so doubling the cost assumption for a high-turnover strategy can erase its edge. In simulation, replacing a normal distribution with a fat-tailed one shows how much tail risk depends on the model. A robust strategy performs similarly across reasonable alternatives; one whose results collapse with small changes is probably overfitted. The trap is reporting only the most favourable parameter set.