Trading Expectancy
A framework for combining win rate, average win and average loss to evaluate a trading process over many trades.
Expectancy asks what the process earns or loses on average
Trading expectancy combines how often trades win with the average size of wins and losses. It is a framework for evaluating a series of trades, not predicting the outcome of the next trade.
(Win rate × average win) − (Loss rate × average loss)
Example
If a hypothetical process wins 40% of the time, averages $200 on winners and loses $100 on losing trades, its arithmetic expectancy is $20 per trade before fees and execution effects: (0.40 × $200) − (0.60 × $100).
Why a high win rate can still lose
A strategy can win frequently but give back more on its occasional losses than it earns on winners. Conversely, a lower-win-rate approach can have positive expectancy if winners are sufficiently larger than losses.
Use meaningful samples
Expectancy estimates are unstable with very small samples and can change as market conditions change. Track actual results after costs, separate strategies where possible, and avoid treating historical expectancy as a guarantee.
Sample size and market regime
Ten trades are usually too little to say much about a noisy process. As the sample grows, estimates become more informative, but changes in volatility, competition or execution can still make old data less representative of the future.
Costs belong in the calculation
Commissions, exchange or regulatory fees, spread and slippage can turn a small gross edge into a negative net edge. Frequent strategies are especially sensitive to execution friction.
Positive expectancy can still lose for long periods
Expectancy is an average, not a schedule. A positive historical expectancy does not prevent losing streaks, drawdowns or future strategy degradation. Variance determines how unevenly outcomes arrive.
The core equation
A common simplified model is: expectancy = (win rate × average win) − (loss rate × average loss). The units can be dollars, percentage return or R-multiples. The calculation describes the historical average per trade under the measured assumptions.