IMO, the most robust way to backtest a systematic strategy is to do the following. I will give a simple example later to clarify.
1) use a historical data set that is long enough to cover multiple periods of market regimes including bullish, bearish and ranging.
2) select "Training Period...