Expectancy
A system with 70% win rate and 1R wins and 3R losses has expectancy of 0.7 × 1 − 0.3 × 3 = -0.2R per trade. It loses 0.2R on average despite winning 70% of trades. A system with 30% win rate and 4R wins and 1R losses has expectancy of 0.3 × 4 − 0.7 × 1 = 0.5R per trade — much better.
Win rate alone is misleading. A high win rate combined with poor reward-to-risk produces negative expectancy; a low win rate combined with strong reward-to-risk produces positive expectancy. Trend-following systems often run win rates of 30-40% but cumulative positive expectancy because winners are large.
Expectancy must be measured by playbook, not aggregated across all trades. A trader might have +0.3R aggregate expectancy that masks a +0.5R from one playbook and -0.2R from another. Killing the negative-expectancy playbook is often the highest-leverage improvement available.
How PerpLog uses Expectancy
PerpLog computes expectancy per playbook, per session, per day-of-week, and per streak state. The Adaptive Sizing engine sizes each trade based on the segmented expectancy of the matching context, not aggregate stats.
Related reading
- Kelly Criterion for Crypto Trading: A Practical Approach — Blog
- Why Every Trader Needs a Trading Journal (And How to Keep One) — Blog
Browse all terms in the trading glossary.