What Is Expectancy in Trading and How to Calculate It
Expectancy is the average dollar gain per trade across your trade log. Learn how to calculate expectancy from your own history and what the number means.
By Imperial Analytics
Expectancy is the single number that tells you whether your trading process makes money on average. It collapses win rate, average winner, and average loser into one dollar figure per trade. This post explains the formula, walks through a worked example, and shows where expectancy can mislead you if the sample is too thin.
What expectancy actually measures
Trading expectancy is the average dollar amount you can expect to win or lose per trade, computed across your full trade log. A positive expectancy means the process makes money on average. A negative expectancy means it bleeds. Expectancy says nothing about any single trade. It describes what the process produces over many trades.
Expectancy is a process metric, not a forecast. It does not predict the next trade. It describes the average outcome of the trades you have already taken, on the assumption that the conditions producing those trades continue to hold.
A trader with a positive expectancy of +$42 per trade and a closed-trade sample of 600 trades has a process that produced roughly +$25,200 in gross expected value across that history. The same trader on a 12-trade sample has noise dressed as a metric.
The formulation Imperial Analytics uses for expectancy in this post follows the form popularized in trading literature in the 1990s and 2000s.1
The expectancy formula
Expectancy equals win rate times average winner, minus loss rate times average loser. Win rate is the share of trades that closed in profit. Average winner is the mean net dollar profit on those winning trades. Loss rate is the share of trades that closed in loss. Average loser is the mean net dollar loss on those losing trades, entered as a positive number.
The formula in plain form:
Expectancy = (Win rate × Avg winner) − (Loss rate × Avg loser)
Two notes on this formula. First, scratches, meaning zero-net-P&L trades after fees, get classified as either wins or losses depending on your convention. A common convention is to count a zero-P&L close as a loss, because it consumed risk capital without producing a winner. Second, the dollar amounts already include commissions, fees, and slippage. Expectancy on gross P&L is a fiction. Expectancy on net P&L is the number that pays for groceries.
How to calculate expectancy from your trade log
Pull every closed trade from your journal. Separate winners from losers. Compute win rate, loss rate, average winning dollar amount, and average losing dollar amount. Apply the formula. The output is the average net dollar value of one trade in your sample. The arithmetic is small. The discipline is in defining what counts as a trade, applying fees to every line, and refusing to compute expectancy on a sample too small to be informative.
A worked example clarifies the steps. The figures below are illustrative.
Data note
The following numbers are illustrative. They are constructed to show the arithmetic, not to represent the performance of any real trader or account. Imperial Analytics does not publish account performance.
Assume a MES day trader closes 100 trades over a quarter with the following net distribution.
- label: Closed trades value: 100 semantic: neutral
- label: Win rate value: 42% semantic: neutral
- label: Avg winner value: +$185 semantic: profit
- label: Avg loser value: -$95 semantic: loss
| Metric | Value |
|---|---|
| Total closed trades | 100 |
| Winning trades | 42 |
| Losing trades | 58 |
| Average winning trade | +$185 |
| Average losing trade | -$95 |
Plug those into the formula:
Win rate = 42 / 100 = 0.42
Loss rate = 58 / 100 = 0.58
Avg winner = $185
Avg loser = $95
Expectancy = (0.42 × $185) − (0.58 × $95)
= $77.70 − $55.10
= +$22.60 per trade
The illustrative process generates an expected +$22.60 per trade over the sampled history. At the sampled pace of 100 trades per quarter, that is about +$2,260 per quarter in gross expected value, before any decision to size up or down.
Common mistakes when computing expectancy from your log
The arithmetic of expectancy is simple. The errors are mostly in the inputs. The three most common are computing on gross instead of net P&L, mixing closed and open positions in the same sample, and treating average winner and average loser as if they were medians. Each error pushes expectancy away from the number the trade log actually produced.
A few specific failure modes to watch for in your own log.
- Computing on gross P&L. Pre-fee expectancy is always more flattering than post-fee expectancy. Apply your real per-side commission, exchange fees, and any platform fees to each trade before the formula runs.
- Including open positions. Open positions have a paper P&L that can swing meaningfully before close. Expectancy is a closed-trade metric. Exclude open positions until they close.
- Reading the average as a median. The average winner is sensitive to a single outsized trade. If one outlier winner is pulling the average up, expectancy will overstate what a typical trade contributes. Inspect the distribution before trusting the number.
- Mixing strategies. Two strategies with very different win rates and average winners produce a blended expectancy that describes neither. Compute expectancy per strategy first, then per account.
- Recomputing expectancy after every trade. Expectancy is a process metric over a sample, not a real-time gauge. Recomputing nightly on a small window invites overreaction.
What different expectancy values mean
Positive expectancy means the process produces money on average across the sampled trades. Negative expectancy means the process loses money on average across that sample. Expectancy near zero, even slightly positive, means commissions and slippage can easily push the process underwater as conditions change.
A few rules of thumb that traders use when interpreting expectancy.
- Expectancy near zero is fragile. A small uptick in fees, a slightly wider average loser, or a slight drop in win rate flips the sign.
- Expectancy is per-trade, not per-day. A trader with +$5 per trade who takes 30 trades a session is in a very different place than a trader with +$5 per trade who takes 2 trades a session.
- Expectancy does not encode risk. Two processes can show the same expectancy and very different drawdown profiles. Pair expectancy with maximum drawdown and worst-day loss before drawing conclusions about whether the process is tolerable to run.
↳ Note
Expectancy is the average dollar value of one trade in your sample. It pays for groceries. It says nothing about the next trade.
Why expectancy can mislead without sample size
Expectancy computed from a small sample is dominated by noise. A handful of outsized winners can inflate a small-sample expectancy that vanishes once a larger sample is drawn. Sample-size discipline is not a stylistic preference. It is a structural feature of any inference drawn from a finite history.
A sample of 20 trades can produce wildly different expectancy values depending on which 20 trades land in the window. One outlier winner of +$2,400 on a 20-trade sample lifts the average winner enough to swing expectancy into clear positive territory, even if the underlying process is breakeven.
A practical floor for expectancy as a process verdict is several hundred trades in the same strategy, taken under broadly similar conditions. A 30-trade sample is a snapshot. A 300-trade sample is evidence. A 3000-trade sample is a verdict, until conditions change.
How expectancy and profit factor describe different things
Profit factor is the ratio of gross winning dollars to gross losing dollars. Expectancy is the average dollar outcome per trade. Profit factor tells you how efficient your winners are at covering your losers. Expectancy tells you whether the average trade puts money on the table. The two metrics can agree, and they can disagree at the edges. The disagreement is where the lesson is.
A process with a profit factor of 1.5 and an expectancy of +$30 per trade is straightforward. The winners outpace the losers in aggregate, and the average trade is positive.
A process with a profit factor of 1.8 driven by one giant winner and many small losers can still show modest expectancy, because the median trade is a small loser and the giant winner is what carries the ratio. Looking only at profit factor would hide that asymmetry. Looking only at expectancy would hide the dependence on the outlier.
Track both. Decide what to change in the process based on which one is moving and which one is steady.
Frequently asked questions
Frequently asked questions
- q: What is a good expectancy value for futures trading? a: Good is relative to your cost structure and sizing. A process that produces +$10 per trade after fees on consistent volume is preferable to a process that produces +$50 per trade only when conditions are perfect. The healthier read is direction and stability over a meaningful sample, not a fixed target number.
- q: How many trades do I need before expectancy is meaningful? a: Several hundred closed trades in the same strategy under broadly similar conditions is a working floor. A 30-trade sample is a snapshot, not a verdict. The right minimum depends on how skewed your winner distribution is. The more your expectancy leans on a few large winners, the larger the sample needed to trust it.
- q: Does expectancy include commissions and fees? a: It has to. Expectancy computed on gross P&L overstates what the process actually produces. Apply your real per-side commission, exchange fees, and any platform fees to each trade before computing the average winner and average loser.
- q: Is expectancy the same as average trade? a: Yes. The expectancy formula resolves to the simple average net P&L per closed trade. The formula is useful because it separates the inputs of win rate, average winner, and average loser, so you can see which input is driving the result.
- q: How does expectancy relate to R-multiples? a: R-multiple expectancy expresses the same idea in units of risk rather than dollars. If you define R as the dollar amount risked on each trade, expectancy in R is the average R-multiple per trade. Both forms answer the same question. R-multiples make it easier to compare expectancy across instruments and account sizes.
If you want expectancy and the inputs behind it computed from your own trade history with explicit sample-size discipline, Imperial Analytics is in early access. We surface the metrics from your trades, label sample sizes, and refuse to call a 30-trade window proof of anything.
Sources
Footnotes
-
Van K. Tharp. Trade Your Way to Financial Freedom. McGraw-Hill, 2007. ↩