A 70% win rate can produce consistent losses, and a 30% win rate can produce exceptional returns. Win rate alone tells you almost nothing about whether your strategy actually works.
It is not complicated. But it is almost never taught clearly, almost never calculated correctly, and almost never used as the primary performance metric by retail traders — which is precisely why most retail traders cannot tell you whether their strategy actually has an edge.
The formula
Expectancy = (Win Rate × Average Win) − (Loss Rate × Average Loss)
The result is expressed as the expected profit or loss per trade, in R (where R = your average risk per trade in dollar terms).
A positive expectancy means that on average, across a large sample of trades, the strategy will generate more than it loses. A negative expectancy means the opposite — regardless of how the wins feel, the losses feel, or what any single month's P&L shows.
Why win rate is not the story
Example A: Win rate 70%, Average win 0.8R, Average loss 2.0R
Expectancy = (0.70 × 0.8) − (0.30 × 2.0) = 0.56 − 0.60 = −0.04R per trade
This strategy loses money. Slowly. Every week.
Example B: Win rate 40%, Average win 2.5R, Average loss 1.0R
Expectancy = (0.40 × 2.5) − (0.60 × 1.0) = 1.0 − 0.60 = +0.4R per trade
This strategy makes money. A trader running this strategy will feel like they're losing most of the time — because they are, 60% of the time — and will frequently abandon it for something that "feels" better, i.e., something with a higher win rate.
This is the primary mechanism by which retail traders destroy edges that were working: they abandon strategies with favorable expectancy because the win rate feels too low, and replace them with strategies that have high win rates and catastrophically unfavorable average win/loss ratios.
What corrupts expectancy calculations
Not including all costs. The average win must be net of commissions, spreads, and platform fees. The "raw" win ignoring costs consistently overstates expectancy — sometimes enough to convert a genuinely positive strategy into a net-negative one when costs are included.
Small sample sizes. An expectancy calculated from 20 trades is worthless. At that sample size, variance completely dominates the result. The minimum meaningful sample is 100 trades on one strategy. Anything below that is noise with a formula applied to it.
Including off-plan trades. If your sample includes both on-plan and off-plan trades mixed together, the expectancy is not telling you about your strategy. It is telling you about your trading behavior, which is a different and more disturbing number. Calculate on-plan trades only.
Excluding "exceptional" losses. Every trader has experienced losses they classify as "not representative" — the news spike, the platform issue, the trade they "shouldn't have been in." Excluding unfavorable outcomes from the expectancy calculation is not analysis. It is selection bias with arithmetic applied.
1. Minimum 100 live forward-tested trades
2. A single, consistently applied strategy
3. Full cost accounting on every trade
4. Inclusion of all losses
5. Separation of on-plan from off-plan trades
When you have this, the expectancy calculation gives you an answer you can trust. A positive expectancy of 0.3R or above across a 100-trade sample is evidence of a real edge. A negative expectancy at 100 trades is evidence that the strategy, as currently applied, does not work.
Let the system do the math.
Edge Builder automatically calculates your true expectancy, separating your on-plan baseline from your off-plan variance, accounting for all commissions and fees. Know definitively whether you have an edge.