Most traders skip backtesting exotics because the data costs money and the spread noise can ruin naive performance metrics. But Python has s

Most traders skip backtesting exotics because the data costs money and the spread noise can ruin naive performance metrics. But Python has solved both problems if you approach them systematically. You should always use a spread-adjusted bar model and a walk-forward loop, otherwise the equity curve will lie to you.The template in this post uses vectorised events for speed, but drops into an event-driven loop for execution simulation. That hybrid lets you test stop-hunt behaviour and spread spikes that appear in USDTRY but never in EURUSD. It also keeps the code readable enough to extend without hiring a quant developer.Practical...