Process Over P&L: Execution Discipline

P&L on any single trade is noise. It's the output of one roll of a dice you've weighted slightly in your favor — not a verdict on whether your process is sound. Judging a rule-based system by its last trade is the fastest way to abandon a good system right before it pays off.

The fix is treating a batch of trades — twenty is a workable minimum — as a single data point, not twenty separate report cards. Inside that trial, every trade is executed exactly the same way: same setup criteria, same position size, same stop. No adjusting risk after a loss. No doubling after a win. You're not trading, you're running a sampling procedure.

Only at the end of the trial do you look at the aggregate: win rate, average R, expectancy. That's the only number that tells you anything. And critically — it's the only point where you're allowed to change the rules.

This is what removes revenge trading. Revenge trading is a reaction to treating one trade as meaningful. If the only meaningful unit is the twenty-trade block, one bad trade is just an input, not an injury.