How do I know if my trading strategy has an edge?

An edge is a positive expected value per trade, kept over enough trades that luck stops being the explanation. Both halves matter, and traders usually test only the first.

The number: expectancy

Expectancy is average win times win rate, minus average loss times loss rate. A system that wins $180 on 45% of trades and loses $110 on the other 55% expects about $20.50 a trade before costs. Positive expectancy is an edge; a high win rate alone is not, because one oversized loser can spend ten small winners. Compute it from written records, not memory. Memory keeps the wins.

The sample: about 25 trades before it means much

Small samples lie in both directions. A profitable week proves as little as a losing one. Around 25 trades of the same strategy, a real pattern starts to separate from noise, and the further past that you get, the harder the number becomes to argue with. The full arithmetic is in the noise-floor piece linked below. If your strategy has not produced 25 comparable trades yet, the honest answer to the edge question is: not knowable yet, keep logging.

The edge you have vs the edge you keep

The quieter failure is a real edge spent in execution. Break your own rules once a week, size up after losses, cut the winners your plan said to hold, and the expectancy you measured stops being the expectancy you collect. Test the gap directly: compute expectancy on all trades, then again on only the trades that followed your rules. The difference is what your execution costs. Travise computes that gap automatically as the Execution Leak, by grading every trade against the rules you committed to, but a spreadsheet and honesty get you the same number slower.