article
Win Rate vs. Expectancy: Why More Winning Trades Can Still Lose Money
A high win rate feels reassuring, but the size of wins, size of losses, and trading costs determine whether the process has positive expectancy.
Win rate answers only one question
Win rate is the fraction of trades that finish positive. It is easy to understand and easy to market, which is exactly why it can become misleading. A strategy can win frequently and still lose money if its occasional losses are much larger than its routine gains. Another strategy can lose more often than it wins and still be profitable when its winners are sufficiently large.
The arithmetic starts with average outcome, not emotional comfort. Expected value combines the probability and average size of gains and losses. Costs then reduce that expectation further. This framing prevents the researcher from rewarding a strategy simply because it produces many small positive trades while hiding a tail of infrequent but destructive losses.
Pay attention to the shape of outcomes
Averages can also hide important structure. If most profits come from a few extreme winners, removing or shrinking those observations may change the conclusion. If losses have a long tail, the average loss may underestimate what happens in stressed conditions. Looking at the distribution of trade outcomes gives more information than comparing two single averages.
Sequence matters operationally even when it does not change the arithmetic mean. Long losing streaks can trigger risk limits, reduce usable capital, or cause a human operator to abandon a valid process. A strategy with acceptable long-run expectancy may still be impractical if its path demands more capital or patience than the operating plan can support.
Costs can flip a small edge
When expected profit per trade is small, fees and spread deserve the same attention as the signal. A one-line change in a cost assumption can move expectancy from positive to negative without changing the nominal win rate at all. This is especially relevant for high-turnover strategies, but it also matters whenever the strategy's expected edge is only a small fraction of the instrument's normal trading friction.
Researchers should therefore report expectancy after reasonable costs and stress those costs. If a strategy remains attractive only under the most favorable fill assumption, the evidence is fragile. A lower-frequency strategy with a smaller headline win rate can be economically stronger if its payoff distribution leaves more room for imperfect execution.
Choose metrics that match the decision
There is no single metric that replaces judgment. Win rate can be useful for understanding the cadence of feedback, expectancy for economic value, drawdown for path risk, turnover for implementation burden, and exposure for concentration. The right report shows enough of these dimensions to explain how the strategy makes and loses money.
The practical lesson is simple: never use win rate as a synonym for quality. Ask what an average winner looks like, what an average and extreme loser look like, how many observations support the estimate, and what remains after costs. A good metric should make the mechanism easier to inspect, not give the strategy a more flattering headline.