How a system winning 90% can lose money
Win rate counts trades without weighing them. A system that wins $1 on 90% of trades and loses $20 on the other 10% loses money: (0.9 × 1) − (0.1 × 20) = −$1.10 per trade.
A system winning only 30% but making five times what it loses gains: (0.3 × 5) − (0.7 × 1) = +0.8. This is why win rate is never read without the risk-to-reward ratio, and why a system is judged only by the expectancy that combines them.
The shared reference record
The same record used across the other performance pages: 100 trades, 40 winners averaging +2R and 60 losers averaging −1R, for an expectancy of +0.20R per trade.
Reusing one record across pages is deliberate: you can compare the measures directly instead of comparing different examples, and it becomes visible that each one views the same thing from an angle, and that none is sufficient alone.
The formula
Win rate = winning trades ÷ total trades × 100
A worked example
In the reference record: 40 winners out of 100, a win rate of 40%. Intuitively the figure looks weak, yet the system is profitable because the average win is twice the average loss.
The threshold separating profit from loss at 1:2 is 33.3%, so 40% clears it comfortably. Drop to 30% and the system loses money despite the figure "being close". That comparison — win rate against the breakeven its ratio implies — is the only reason the number is useful at all.
Common mistakes with this term
- Selecting strategies by win rate alone, favouring systems that win often and lose more.
- Comparing win rates between two methods without comparing their reward-to-risk alongside.
- Computing it from a small sample; ten trades cannot separate a real edge from luck.