Win rate tells you how often you are right. Profit factor tells you how your total wins compare to your total losses. R-multiple tells you how each trade compares to the risk you took. Each answers part of the question every trader actually cares about: if I keep trading this way, do I make money?
Expectancy is the number that answers it directly. It is the average amount a strategy makes or loses per trade, once winners and losers are both accounted for. A positive expectancy means that, over enough trades, the strategy should earn money. A negative one means that no amount of discipline, confidence, or good days will fix it.
It is also one of the most commonly miscalculated numbers in trading. Traders compute it on a few dozen trades, in dollars while their position size keeps changing, before costs, and across every trade they have ever taken — including the ones that had nothing to do with their strategy. Each of those mistakes produces a number that looks precise and says very little.
This article covers how to calculate expectancy properly, how to express it in the unit that actually makes it useful, how to tell whether your sample is large enough to trust it, and how to separate the expectancy of your strategy from the expectancy of your behaviour.
What Expectancy Actually Measures
Expectancy is an average, but it is a specific kind of average: it combines how often you win with how much you win and lose. That combination is exactly what win rate and payoff ratio fail to capture on their own.
A strategy that wins 70% of the time can lose money if its losses are much larger than its wins. A strategy that wins 30% of the time can be highly profitable if its winners are several times larger than its losers. Neither number alone tells you which situation you are in. Expectancy does, because it weighs each outcome by how often it happens.
Two things follow from that:
- Expectancy is a property of a process, not of a trade. Any single trade can win or lose. Expectancy describes what the process produces on average across many repetitions.
- It only means something for a consistent process. If the rules, the markets, or the risk change halfway through your sample, you are averaging two different strategies together, and the result describes neither.
The Expectancy Formula
The standard formula, in money:
Expectancy = (Win rate × Average win) − (Loss rate × Average loss)
Where the loss rate is simply 1 minus the win rate, and the average loss is written as a positive number.
A worked example. Over 100 trades, a trader has 40 winners averaging $450 and 60 losers averaging $200.
- Win rate: 40%, loss rate: 60%
- Expectancy = (0.40 × $450) − (0.60 × $200) = $180 − $120 = $60 per trade
Over the next 100 trades taken the same way, that strategy would be expected to make roughly $6,000 — with the important caveat, covered below, that "roughly" can be very wide.
An even simpler way to get the same number: add up the net result of every trade and divide by the number of trades. The formula above is just that average, broken into its parts so you can see why it is positive or negative.
Expectancy in R: The Version Worth Tracking
Dollar expectancy has a problem: it changes whenever your position size changes, even if your trading does not. A trader who doubles their size will see their dollar expectancy roughly double without becoming a better trader in any way.
The fix is to measure every trade in R — its result divided by the amount you risked on it. The formula keeps the same shape:
Expectancy (R) = (Win rate × Average win in R) − (Loss rate × Average loss in R)
Using the same trader, who risked $200 per trade: the average win of $450 is 2.25R, and the average loss of $200 is 1R.
- Expectancy = (0.40 × 2.25R) − (0.60 × 1R) = 0.90R − 0.60R = +0.30R per trade
In practice, expectancy in R is simply your average R per trade. It says: for every unit of risk this strategy takes, it gives back 0.30 units on average. That number is comparable across account sizes, across markets, and across periods when your size changed — which is exactly what a dollar figure cannot do.
How dollar expectancy hides a decaying edge
Consider a trader who takes 100 trades over six months. For the first 50 they risk $100 per trade and average +0.40R. Pleased with the results, they raise their risk to $300 per trade for the next 50, and average +0.10R.
| Period | Risk per trade | Expectancy in R | Expectancy in $ |
|---|---|---|---|
| First 50 trades | $100 | +0.40R | $40 |
| Next 50 trades | $300 | +0.10R | $30 |
| All 100 trades | mixed | +0.25R | $35 |
In dollars, the second half looks only slightly worse than the first: $30 per trade instead of $40. In R, the edge fell by three quarters. The dollar figure made a serious deterioration look like a minor dip, because the larger size was masking it. The R figure shows the problem the moment it appears.
What a Good Expectancy Looks Like
There is no universal benchmark, because what counts as "good" depends on how the edge is structured and how often you can use it. What matters is how win rate and payoff combine. Here is what that combination requires, assuming a full 1R average loss:
| Win rate | Average win needed to break even | Average win needed for +0.25R |
|---|---|---|
| 30% | 2.33R | 3.17R |
| 40% | 1.50R | 2.13R |
| 50% | 1.00R | 1.50R |
| 60% | 0.67R | 1.08R |
| 70% | 0.43R | 0.79R |
Every row has the same expectancy in the last column. A trend-following strategy that wins 30% of the time and a mean-reversion strategy that wins 70% of the time can be equally good — they just get there differently. It is also why a win rate, taken by itself, tells you very little about a strategy.
Three practical tests are more useful than any benchmark:
- It must be positive after costs. Commissions, fees, spread, and slippage are paid on every trade, winners and losers alike. They come straight off expectancy (more on this below).
- It must be positive on the low end of its uncertainty range, not just on average. A strategy measured at +0.30R that could plausibly be anywhere between −0.20R and +0.80R has not proven anything yet.
- It must be worth your time at the frequency you can trade it. Expectancy per trade multiplied by trades per month gives expected R per month. A +0.20R strategy traded 40 times a month produces about 8R. A +0.50R strategy that only sets up 6 times a month produces about 3R. The higher expectancy is not automatically the better strategy.
A very high expectancy on a small sample should be treated with the same suspicion as an unusually high profit factor: usually it means a few outsized winners have not yet been balanced by the losses that will eventually come.
How Many Trades Before Expectancy Means Anything?
This is where most expectancy figures fall apart. Trading results are noisy, and an average taken over a small number of noisy results can land a long way from the true value.
Take a strategy that genuinely wins 45% of the time, with winners of +2R and losers of −1R. Its true expectancy is (0.45 × 2) − (0.55 × 1) = +0.35R. Here is roughly where a trader's measured expectancy is likely to land, depending only on how many trades they have taken:
| Trades measured | Measured expectancy will usually fall between |
|---|---|
| 30 | about −0.20R and +0.90R |
| 100 | about +0.05R and +0.65R |
| 300 | about +0.18R and +0.52R |
| 1,000 | about +0.26R and +0.44R |
These ranges cover about 95% of outcomes — roughly two standard errors either side of the true value — and they come straight from the spread of the strategy's own results.
Simulating this strategy thousands of times shows what that means in practice. After 30 trades, roughly one sample in fifteen shows a negative expectancy, and close to half land more than 0.2R away from the true +0.35R — looking either like a weak edge or like a much stronger one than it really is. A trader judging the strategy at this point is mostly reading noise, and a trader who abandons it here throws away a real edge and starts measuring a new one from zero. After 100 trades the range is still wide enough that "barely profitable" and "very profitable" are both plausible. Only in the hundreds does the number start to settle.
You can estimate the same range for your own results with two spreadsheet columns:
- Put every trade's result in R in one column.
- Calculate the average (your expectancy) and the standard deviation of that column.
- Divide the standard deviation by the square root of the number of trades. That is the standard error.
- Your expectancy is plausibly anywhere within about two standard errors either side of the average.
If that range includes zero, you do not yet know whether you have an edge. That is not a reason to stop trading the strategy — it is a reason not to change it, scale it up, or abandon it on the strength of the number alone. It is the same reasoning that applies to losing streaks: small samples produce extreme-looking results as a matter of routine.
Four Ways Traders Measure Expectancy Wrong
1. Measuring it in dollars while size keeps changing
Covered above: dollar expectancy mixes the quality of the trading with the size of the trading. If your risk per trade has changed at any point in your sample, measure in R.
2. Leaving costs out
Costs matter far more to expectancy than most traders assume, because they are fixed per trade while the edge per trade is often small. Expressing costs in R makes the effect obvious.
Suppose round-trip commissions and typical slippage come to about $12.50 per trade. For a trader risking $250 per trade, that is 0.05R. A strategy with a gross expectancy of +0.15R keeps +0.10R — a third of the edge gone to costs.
Now take a short-term trader with tight stops risking $50 per trade, paying the same $12.50. The cost is 0.25R per trade. A gross expectancy of +0.20R becomes −0.05R: a strategy that looks profitable on paper and loses money in practice. The tighter the stop relative to trading costs, the more expectancy has to cover.
3. Averaging different strategies together
A breakout strategy and a mean-reversion strategy traded in the same account produce one blended number that describes neither. If one is strongly positive and the other slightly negative, the blend looks mediocre and hides both facts. Tag every trade with the setup it came from, and calculate expectancy per setup.
4. Mixing trades that followed the plan with trades that didn't
This is the most important one, and the least discussed.
Every account's results contain two kinds of trades: the ones taken according to the written rules, and the ones that weren't — early entries, moved stops, oversized positions, setups that didn't really qualify, trades taken out of frustration. The overall expectancy is a blend of both. It measures the strategy and the trader's adherence to it at the same time, with no way to tell which is responsible for what.
An example. Over 100 trades, 70 followed the plan and averaged +0.45R. The other 30 broke at least one rule and averaged −0.40R.
- Blended expectancy = (70 × 0.45 + 30 × −0.40) ÷ 100 = (31.5 − 12) ÷ 100 = about +0.20R
The trader looking only at the overall figure sees a weak edge of +0.20R and starts looking for a better strategy. But the strategy, as written, is producing +0.45R. If all 100 trades had been taken the way the first 70 were, the result would have been around +45R instead of +19.5R. The difference — roughly 25R, or about a quarter of the starting account at 1% risk per trade before compounding — was not lost to the market. It was lost to execution.
The reverse happens too. If the rule-following trades are the ones with negative expectancy, the strategy itself is the problem, and no amount of discipline will fix it. The split is what tells you which problem you have, and the two need opposite responses. This is the practical side of why traders break their own rules: the cost of each rule violation is invisible until it is measured separately.
Expectancy by Segment: Where Your Edge Actually Lives
Once you measure expectancy in R, per setup, on rule-compliant trades, a second layer becomes available: splitting it by the conditions around the trade — instrument, session, day of the week, market regime.
An illustrative example of what that can reveal for a single setup:
| Segment | Trades | Expectancy |
|---|---|---|
| Breakout, London session | 64 | +0.38R |
| Breakout, New York afternoon | 41 | −0.12R |
| Pullback, all sessions | 38 | +0.21R |
| All trades | 143 | +0.19R |
The overall +0.19R hides one clearly productive segment, one that may be costing money, and one in between. Removing the weak segment could raise the overall expectancy considerably without changing a single entry rule.
Two cautions apply, and both come from the sample-size section above. Every segment you split off has fewer trades than the whole, so its range of uncertainty is wider — 41 trades is not enough to condemn a segment on its own. And the more segments you slice, the more likely it becomes that at least one looks bad purely by chance. Treat a weak segment as a hypothesis to watch over the next batch of trades, not as a verdict.
Positive Expectancy Does Not Protect You From Drawdowns
A positive expectancy describes the long-run average. It says nothing about the path to get there. A +0.30R strategy with a 40% win rate will still produce long runs of losses — the losing streaks article shows how long they typically get — and those runs arrive on their own schedule, not in proportion to the average.
This is why expectancy and risk per trade have to be decided together. Expectancy tells you whether a strategy is worth trading. Position sizing decides whether you survive long enough for that expectancy to show up. A strategy with a real edge, traded at a size its normal drawdowns can break, is still a losing strategy for the trader running it.
How to Calculate Your Own Expectancy, Step by Step
- Collect a consistent sample. One strategy or setup, one set of rules. If the rules changed, start a new sample from the change.
- Record the planned risk on every trade. Entry, stop, and position size at the moment of entry — without these, R cannot be calculated. This is one of the fields that belongs in every trading journal.
- Convert every result to R. Net result divided by the amount risked.
- Subtract costs. Either use net results, or subtract your average cost per trade in R from the final figure.
- Calculate the average R. That is your expectancy.
- Calculate the uncertainty. Standard deviation divided by the square root of the number of trades; about two of those either side of the average is your plausible range.
- Split rule-compliant trades from the rest. Calculate expectancy for each group separately. The compliant figure is your strategy. The gap between the two is the cost of your execution.
- Recalculate on a schedule, not after every trade. Every 25 or 50 trades is enough. Checking after each trade mostly measures noise and invites exactly the reactive changes that damage a working strategy.
The Short Version
- Expectancy is the average result per trade: (win rate × average win) − (loss rate × average loss). It is the most direct answer to whether a strategy makes money.
- Measure it in R, not dollars. Dollar expectancy changes whenever position size changes and can hide a deteriorating edge.
- There is no universal "good" number. The same expectancy can come from a low win rate with large winners or a high win rate with small ones. What matters is that it is positive after costs, at a frequency that makes it worth trading.
- Small samples are extremely noisy. After 30 trades, a strategy with a true +0.35R edge measures more than 0.2R away from that value almost half the time, and occasionally shows a negative expectancy. Calculate the uncertainty range before drawing conclusions.
- Costs are paid on every trade and hit tight-stop strategies hardest. Express them in R.
- Calculate expectancy separately for trades that followed your rules and trades that didn't. The first number measures your strategy; the gap measures your execution.
- A positive expectancy does not prevent drawdowns. Size your positions so you survive long enough for the edge to play out.
Want to see the expectancy of your strategy, separate from the expectancy of your mistakes? TradeProof AI tracks expectancy and average R across your trades, and grades each one against your own written strategy — so you can see which results came from following your plan and which came from breaking it. Log your first trade free.