Why Your Backtest Is Lying to You (And How to Fix It)

Repaint is the most famous backtest lie, but it’s rarely the most expensive.

Look-ahead bias creeps in through fundamental data, alternative data, and split-adjusted prices that bake in future events.

Slippage modeling has gotten better in Pine v6, but the default is still optimistic for anything below 1m.

Survivorship bias — your universe of symbols today is not the universe you’d have traded five years ago. Use a delisted-aware data source.

I run every backtest with three slippage profiles and three commission profiles. If the strategy only works in the best case, it doesn’t work.