Somebody sends you a track record: 68 percent of trades closed in profit. It sounds like a good system. It is not information. A method that wins two trades out of three and gives back three times as much on the third is losing money steadily while the win rate on the marketing page stays flattering. Expectancy is the number that catches this.
The definition is plain. Expectancy is the average result of one trade, taken over a large number of trades:
Expectancy = (win rate × average win) - (loss rate × average loss)
That is it. No parameters to tune, nothing to interpret. If the result is positive, repeating the process has a positive expected value. If it is negative, repeating it faster only loses money faster.
Express everything in R and the number becomes portable
Raw currency amounts make expectancy hard to compare, because a trade risking 40 dollars and a trade risking 400 dollars sit in the same average. The fix is R: define R as the amount risked on a trade, then measure every outcome as a multiple of it. A trade that hits a target twice the size of its stop is +2R. A trade stopped out is -1R. Half a stop given back on a manual exit is -0.5R.
Once outcomes are in R, expectancy reads as an average R per trade. A system at +0.25R means the average trade, over many trades, returns a quarter of what it risks. That figure is comparable across gold and index CFDs, across a 2,000 dollar account and a 200,000 dollar one, and across two traders using the same method at different position sizes. It is also the only sane way to compare your live record with your backtest, since the position sizes almost never match.
Win rate and risk to reward are one dial, not two
Every win rate has a matching risk to reward ratio at which the system breaks even, and the relationship is simple: breakeven win rate equals 1 divided by (1 + R), where R is the reward multiple. That arithmetic produces the table people should look at before they argue about which style is better.
| Average reward per 1R risked | Win rate needed to break even | Expectancy at a 50% win rate |
|---|---|---|
| 1.0R | 50.0% | 0.00R |
| 1.5R | 40.0% | +0.25R |
| 2.0R | 33.3% | +0.50R |
| 3.0R | 25.0% | +1.00R |
| 0.5R | 66.7% | -0.25R |
The last row is where most struggling traders live. Taking profit at half the stop distance produces a lovely win rate and a negative expectancy, and it feels correct every single day until the monthly statement arrives. This is the practical case for thinking in risk to reward ratios rather than in hit rates.
Costs are inside the number, not next to it
Expectancy calculated from theoretical levels flatters every system. The realised figure has to be built from fills, not from the chart.
Take a gold trade with a 20 pip stop and a 40 pip target, nominally 2R. Add a 2 pip spread on entry and the real risk becomes 22 pips while the reward becomes 38. The ratio drops from 2.00 to about 1.73. Add commission and the swap on a position held overnight and it falls again. Nothing about the strategy changed. The expectancy did, and at a 40 percent win rate that difference is the distance between a system that works and one that does not. If you have not read how the spread is charged on entry, that arithmetic is worth doing by hand once.
Expectancy is an average, not a prediction. A positive figure says nothing about the next ten trades, and every system with an edge still produces losing runs long enough to feel terminal. Leveraged trading carries a high risk of loss, and a positive historical expectancy is not a promise that the future resembles the past.
Expectancy per trade is only half the picture
Two systems both at +0.3R are not equally useful if one produces four opportunities a month and the other produces sixty. The quantity that matters for a real account is expectancy multiplied by the number of trades taken in the period, usually written as expectancy per month in R.
This is also where traders talk themselves into trouble. Increasing trade count multiplies whatever expectancy you have. If the figure is +0.2R, more opportunities help. If it is -0.1R, more opportunities are the fastest possible way to a large drawdown. Frequency is an amplifier with no opinion about direction.
The other constraint is that expectancy per trade tends to fall as frequency rises, because the extra trades are the marginal ones: worse locations, wider spreads, thinner sessions. A trader who doubles activity and watches expectancy halve has gained nothing and doubled the cost paid to the broker.
Sample size, and the trap of an early number
Expectancy from thirty trades is a number, not evidence. It is worst for low win rate systems, where the whole result depends on a few large winners: run the same thirty trades in a different order and the equity curve looks like a different strategy. A system with a one in three hit rate will produce runs of ten consecutive losses now and then, and those runs are ordinary rather than diagnostic.
Three habits make the estimate trustworthy. Record every trade in R at the moment it closes, including the ones you exited early, since those are exactly the ones memory edits. Recalculate expectancy on a rolling window as well as on the full history, because a drop in the rolling figure is the earliest sign a market condition has changed. And keep backtest and live expectancy in separate columns forever. When live sits far below the test, the usual explanation is not bad luck but an overfitted test, or costs that were never modelled.
When the number comes out negative
There are three real responses, and trading smaller is not one of them. Reduce the cost per trade, by holding fewer positions overnight or trading instruments with tighter spreads at the hours you actually trade. Change the exit, since in most journals I have looked at the losing side is fine and the profit taking is what destroys the ratio. Or stop running the method and go back to testing.
What does not work is cutting size and continuing. Halving risk halves the loss rate in currency terms while leaving the sign of the expectancy exactly where it was. A negative edge traded politely is still a negative edge, and the account gets there anyway, just later.
"People ask me what my win rate is. It is the wrong question, and my answer would not help them anyway, because I have no idea what their average loss looks like."
— Alex Onta, Executive Director, SINGUARD
Key Takeaways
- Expectancy is win rate times average win minus loss rate times average loss. Measured in R multiples it becomes comparable across instruments, accounts and traders.
- Breakeven win rate equals 1 divided by (1 + reward multiple), so a 2R system needs 33.3 percent and a 0.5R system needs 66.7 percent.
- Spread, commission and swap belong inside the calculation. A nominal 2R trade with a 2 pip spread on a 20 pip stop is closer to 1.7R in practice.
- Trade frequency multiplies expectancy without changing its sign, and reducing position size on a negative edge slows the outcome rather than fixing it.
Frequently Asked Questions
What is the formula for trading expectancy?
Expectancy equals the win rate multiplied by the average win, minus the loss rate multiplied by the average loss. When every trade risks the same amount, expressing wins and losses as multiples of that risk turns the result into an average number of R per trade, which can be compared across instruments and account sizes.
Is a high win rate a sign of a good system?
No. Win rate is only half of the calculation. A system that wins nine times out of ten and loses fifteen times its average win on the tenth trade has negative expectancy. The win rate is only meaningful once you know the average size of the wins and the losses.
How many trades do I need before expectancy means anything?
More than most traders assume. Thirty trades gives a number, not evidence, especially for systems with a low win rate where the result depends on a handful of large winners. Treat early expectancy as a rough sign of direction and keep recording, because the estimate becomes stable slowly and never becomes a forecast.