Risk reward ratio vs win rate: break-even after costs
At a risk reward ratio of 1:3 a strategy breaks even when 25.0% of its trades win, but only before costs. Across 6,722 backtested trades of three 1:3 strategies on 20 assets, costs cut the average win to 2.95 R and raised the average loss to 1.05 R, which moves break-even to 26.3%. The actual win rate was 22.4%.
The formula and what it leaves out
The textbook rule says that a strategy whose target is RR times its stop breaks even at a win rate of 1/(1+RR): one trade in two at 1:1, one in three at 1:2, one in four at 1:3. The rule assumes that every winner earns exactly RR and every loser loses exactly one stop. Real trades do not behave like that. The spread, the commission and the swap come off every trade, whether it wins or loses, so winners earn a little less than planned and losers lose a little more.
Measured in units of risk (R, the amount lost at the stop), the cost of one trade looks small: a few hundredths of R. But it is paid on every trade, and at 1:3 roughly three trades out of four are losers. The costs therefore land mostly on the side of the equation that already dominates, and the win rate you need climbs.
Break-even win rate after costs, measured
We took every closed trade of five strategy templates, each run separately on 20 assets, and computed for each template the average net win, the average net loss and the win rate at which the two cancel out: average loss ÷ (average win + average loss). This is the same formula as 1/(1+RR), only with the realised numbers instead of the planned ones.
| Strategy | Target | Average win after costs | Break-even after costs | Actual win rate |
|---|---|---|---|---|
| Fair value gap | 1:3 | 2.87 R | 26.6% | 26.6% |
| Order block | 1:3 | 2.95 R | 26.2% | 23.1% |
| Level sweep | 1:3 | 2.96 R | 26.3% | 22.1% |
| Range deviation | Opposite side of the range | 4.33 R | 19.3% | 21.8% |
| Asian range | Opposite side of the range | 1.52 R | 40.7% | 39.7% |
For the three templates with a fixed 1:3 target, costs move break-even from 25.0% to about 26.3%. The fair value gap template sat almost exactly on its break-even line, which is why its small edge before costs disappeared after them. The order block and level sweep templates were below the line even before costs. Only 1 of the 5 templates had a win rate above its own break-even after costs.
Looking at single strategy-and-asset combinations of the 1:3 templates with at least 20 trades, 11 of 40 won more often than their own break-even rate after costs. With a few dozen trades per combination, a part of that count is chance, so a single winning combination is not evidence of an edge.
Targets that are not a fixed ratio
For the range deviation and Asian range templates the target is the opposite side of the range, so every trade has its own ratio. The formula still works if you plug in the average realised win. Range deviation won 4.33 R on average and needed only 19.3% winners; the Asian range won 1.52 R and needed 40.7%. A larger target lowers the required win rate and a smaller one raises it. Neither is better on its own: what decides the result is how the actual win rate compares with the rate the realised wins and losses require.
Why a high win rate is not the goal
Strategies are often advertised by their win rate alone. A win rate without the ratio is meaningless: a strategy that wins most of its trades with a target far smaller than its stop can lose money, and a strategy that wins one trade in five can make money if its winners are large enough. The useful question is always the gap between the actual win rate and the break-even rate for the realised wins and losses, after costs.
How to use this on your own strategy
- Measure the average win and the average loss in R after costs, not the planned ratio.
- Compute break-even as average loss ÷ (average win + average loss) and compare it with the actual win rate.
- Remember that a short stop raises the cost per trade in R and with it the break-even rate; a lower timeframe for the stop is not free.
- Judge the gap between the two rates together with the number of trades: a few percentage points over a few dozen trades is within chance.
How we measured
- Engine backtests on the production server, each of 20 assets separately, period 2022-01-01 to 2026-01-01, starting balance $10,000, fixed risk of 1% per trade.
- Commission and swap measured for one commission-charging broker; entries and exits at bid and ask prices, so the spread is already inside the prices.
- Win rate = take-profits ÷ (take-profits + stop-losses); exits at break-even or by time are left out, as in the engine's own win rate. Average win and loss = net result of take-profit and stop-loss trades in R, where 1 R is the fixed risk per trade.
Backtests run on October 30, 2026.
Questions
What win rate do I need with a 1:2 risk reward ratio?
One trade in three before costs, by the formula 1/(1+2). After costs it is higher; how much higher depends on the size of your costs relative to the stop.
Is a higher risk reward ratio always better?
No. A higher target lowers the win rate you need but also lowers the win rate you get. Only the two together, after costs, decide the result.
Why is the average loss larger than one R?
Because commission and swap are added to the stop. The stop itself costs exactly one R; the costs come on top of it, while on a winner they come off the profit.
Does the break-even shift matter if the strategy has an edge?
Yes. An edge smaller than the cost shift disappears after costs, which is what happened to the fair value gap template in this test.
For education, not investment advice. Past results, including backtests, do not guarantee future results.
← All articles