The one-sentence definition
Expectancy is the average result per trade: how much you'd expect to make (or lose) on a typical trade, combining your win rate with the average size of your wins and losses.
How it's calculated
Expectancy = (win rate × average win) - (loss rate × average loss).
Multiply the chance of winning by the average winning amount, subtract the chance of losing multiplied by the average losing amount. A positive result means the strategy makes money on average per trade.
A worked example
Take a strategy that wins just 40% of the time. Its average win is $300; its average loss is $100.
Expectancy = (0.40 × $300) - (0.60 × $100) = $120 - $60 = $60 per trade.
Despite losing 60% of its trades, it makes $60 on average every time it trades, because the wins are three times the size of the losses. Now flip it: a strategy that wins 70% of the time but whose average loss ($300) dwarfs its average win ($100) has expectancy of (0.70 × $100) - (0.30 × $300) = $70 - $90 = -$20 per trade - a loser, despite the high win rate.
Why it matters for backtesting
Historical expectancy summarises the average result across the trades in your sample. A small sample or a few unusually large wins can make that average unstable. Read it alongside trade count, costs, and the distribution of individual results.
Win rate alone leaves out the size of gains and losses. Expectancy combines both, so it can reveal a strategy that wins often but loses more on its losing trades than it earns overall.
Common mistakes
Where expectancy gets misread:
- Chasing win rate instead. A 90% win rate with occasional catastrophic losses can still have negative expectancy.
- Computing it from too few trades. Expectancy is an average; it only means something across a decent sample.
- Ignoring costs. Fees and slippage come straight out of expectancy and can flip a marginal strategy negative.
- Assuming it's stable. Expectancy measured in one market regime can change in another - test across varied conditions.