Backtest without a stop loss.
Not trade without one. Backtest without one, and find out whether your signal was ever real.
This was Dave’s topic and it is one of the rare techniques I was surprised to learn most traders skip. I call it running it naked. Entry, timed exit, nothing else.
Two reasons. The first is the smoke test. Take a random entry, wrap tight risk management around it, and you can make almost anything look profitable. Good for you, but you have not found an edge, you have found a stop loss. Strip it out and see whether the signal predicts anything on its own.
The second is finding where the stop belongs. Maximum adverse excursion is the worst point of each trade, how far underwater you went before it resolved. Collect that across every trade, normalize by ATR, and you stop guessing at round numbers.
The trap: optimizing purely for profit usually says the best stop is no stop, which is untradeable. And tight stops look brilliant in a backtest right up until the bid-ask spread eats them. I have built the 5-cent-stop strategy that printed money in simulation and died on contact.
One more, because order matters. Choose the stop first, rerun, then choose the target. One backtest cannot tell you which came first.
If you want systems where the stop, the sizing, and the honest drawdowns are already worked out, that’s Stats Edge Pro. $149.99 a month, 30-day money-back guarantee.
Michael Nauss, CMT, CAIA, CDMS