I am not sure i don't oppose that - given the tremendous influence of indicators people never even bother to properly understand, mathematically. But simulating proper fills is really hard - it is a lot less so if you run a feed replay (something ninja can not properly do to start with) of all book changes in proper order. But if you backtest based on basically - bars.... sitting duck, waiting for a disaster. That said, replay is computational intensive. And storage intensive. Blows the budgets of most retail traders, especially most that just spend 0 for ninja