After a few hundred trades, compute for each grade: share of all trades, expectancy in R, and number of trades.
What a typical result might look like, as an illustration:
- A trades: 38% of the total, expectancy +0.3R.
- B trades: 44%, around +0.05R.
- C trades: 18%, clearly negative.
Two things keep the test honest:
- Tag at entry, never after. Grading after the result is useless.
- Watch for bias on borderline cases. Even tagging at entry, it's easy to rate a trade higher when you really want to take it. Having a script compute the checklist removes that.