Definition and mechanics of Expectancy
Expectancy is a performance concept that expresses the average result per trade (or decision) given an underlying distribution of outcomes. In its simplest form, it combines outcome size with the chance of each outcome: if you can label outcomes (for example, “win” and “loss”), and you know the average gain when a win happens and the average loss when a loss happens, then expectancy is computed from those averages weighted by their probabilities.
Two mechanics matter most.
-
What probabilities are used? Expectancy needs probabilities for each outcome category. Those probabilities must come from somewhere—often historical backtests or past trade logs.
-
What payoff is used? Payoff should reflect the real net outcome, not just the gross price movement. That means the expectancy calculation must align with how gains and losses are actually realized.
How Expectancy works in practice (and where it can mislead)
Expectancy is helpful when it matches the situation you are evaluating: similar decision rules, similar market conditions, and similar cost and execution behavior. But even when the arithmetic is correct, the limitations come from what is assumed.
A common failure mode is probability drift: the probability of winning or losing changes when volatility, liquidity, spreads, or volatility regimes change. If you estimate win/loss frequencies from historical data, but the future behaves differently, the weighted average no longer represents what will happen.
Another failure mode is payoff mismatch. Many expectancy calculations use simplified assumptions (such as ideal fills and ignoring slippage). In real trading, execution quality varies. That can change average win size, average loss size, and therefore the computed expectancy.
Evidence and example: small assumption changes can flip expectancy
Consider a simplified two-outcome model with two assumptions:
- You estimate the probability of an outcome (win) as a fixed number.
- You estimate average payoff for that outcome from historical results.
If the win probability estimate is off by a small amount, the weighted average can move noticeably, especially when win and loss magnitudes are not symmetric. If the average loss is larger in the realized environment than in your assumed payoff model (for example, because actual exits differ from the plan), then the expectancy calculation can become optimistic.
Key point: expectancy is not “wrong” because the math fails; it becomes less meaningful because the inputs no longer describe the future.
Limitations, risks, and failure modes to verify
Expectancy has limitations that are important to check independently:
-
Historical relationships do not establish future results. Even a stable-looking historical expectancy can deteriorate when market conditions shift.
-
Costs and execution uncertainty are easy to overlook. If spreads, commissions, slippage, or delays are not included consistently, the expectancy may describe a theoretical world rather than realized outcomes.
-
Data and categorization choices change expectancy. How you define “win,” “loss,” and the period of measurement can alter probabilities and average payoffs. Selecting a time window that happens to be favorable can inflate expectancy.
-
Non-stationarity and regime changes. Expectancy assumes the environment generating outcomes is stable enough for estimated probabilities and payoffs to remain relevant.
-
Overconfidence in a single number. A point estimate of expectancy hides variability. Two strategies with the same expectancy can have very different distributions of results (for example, frequent small wins versus rare large wins), affecting drawdowns and survivability.
Verification and next question to reduce uncertainty
To use expectancy responsibly as an analysis tool, verify what would have to remain true for the estimate to be meaningful:
- Probability verification: Do the win/loss frequencies hold under new, unseen data?
- Payoff verification: Are gains and losses computed using realized net results (including costs and realistic execution assumptions)?
- Stability check: Does expectancy remain similar across different market regimes or time periods?
If these checks fail, expectancy becomes a less reliable summary of future performance, and its limitation is not conceptual—it is empirical and input-dependent.