Expectancy
The average R-multiple per trade, the single most important number to evaluate a strategy on, and the metric the trading-marketing layer prefers not to lead with.
Expectancy is the average outcome of a trade in a strategy, expressed as an R-multiple. It is the single most important number to evaluate a strategy on, because it captures both the win rate and the relative size of wins and losses in one quantity.
The formula
expectancy = (win rate × average winning R) − (loss rate × average losing R)
A strategy that wins 50% of the time, with an average win of 1.5R and an average loss of 1R, has an expectancy of:
0.5 × 1.5 − 0.5 × 1 = +0.25R per trade
Over a hundred trades, the expected outcome is +25R. If each trade risks 1% of equity, that is +25% on the account, before friction and compounding effects. (See R-multiple for the underlying unit.)
Why it is the right metric
The win rate alone does not determine profitability. The average win size alone does not determine profitability. The combination of the two, with their relative frequencies, does. Expectancy is the smallest number that captures both.
A worked illustration of why this matters:
| Strategy | Win rate | Avg win | Avg loss | Expectancy |
|---|---|---|---|---|
| A | 70% | +1R | −3R | 0.7 × 1 − 0.3 × 3 = −0.2R |
| B | 50% | +1R | −1R | 0.5 × 1 − 0.5 × 1 = 0R |
| C | 40% | +3R | −1R | 0.4 × 3 − 0.6 × 1 = +0.6R |
| D | 30% | +5R | −1R | 0.3 × 5 − 0.7 × 1 = +0.8R |
Strategy A wins 70% of trades and loses money. Strategy D wins only 30% and makes the most. The pattern is general: many high-edge strategies are wrong more often than right, and many high-win-rate strategies have negative expectancy.
Positive expectancy is necessary but not sufficient
A positive-expectancy strategy will make money on average. It will not necessarily make money for you, because:
- Variance is real. A +0.5R-per-trade strategy still has losing streaks of 5-7 trades regularly. Without sizing for the streaks, the account does not survive to capture the long-run average. (See Risk Management Basics.)
- Costs reduce expectancy. The expectancy you calculate from backtests is usually gross of friction; the realised expectancy after spread, swap, and slippage is meaningfully lower. A strategy with +0.2R gross expectancy can be a 0R strategy net of costs.
- Behavioural drift turns positive expectancy negative. A positive-expectancy system run by a trader who abandons it during the inevitable losing streaks does not realise the positive expectancy. The execution is part of the strategy.
- Backtest expectancy overestimates live expectancy. Sample bias, parameter optimisation, and regime-specific behaviour all inflate the backtested number. The realised expectancy after a year of live trading is usually 30-60% lower than the backtest suggested. (See Why Most Retail Systems Fail.)
What to look for
A defensible expectancy claim has four properties:
- It is computed from a real trade log, not a hand-picked sample.
- It is net of all costs the trader actually paid.
- It is computed over a sample large enough to be statistically meaningful, typically 100 trades or more.
- It carries an honest variance estimate, not just the point statistic.
A claim of “70% win rate” or “5% per month” without an expectancy figure attached should be treated as marketing copy, not evidence.
In practice
Track every closed trade in R-multiples. Compute the average R across a meaningful sample. That number, more than any other, tells you whether what you are doing has an edge. It is the metric the trading-marketing layer prefers not to lead with, which is itself informative.
See R-multiple for the underlying unit, and Risk Management Basics for how expectancy combines with position sizing to produce real account growth.