A backtest will not yield a representative result that generalizes to future periods if the time period used does not reflect the current environment well, lacks relevant market regime aspects, and does not include enough data points or captures extreme historical events that are unlikely to repeat.

The solution involves using sample periods that include important market phenomena, or generate synthetic data that reflect relevant market characteristics (see the Resources section for guidance on implementation).