A 91% win rate can lose money
A win rate says how often a strategy wins, not how much. With small wins and large losses, 91 wins in 100 trades can still lose money.
The win rate is the share of trades that closed in profit. It is the first number on most signal pages, and a high one feels safe. But it says only how often a strategy wins. It says nothing about how much it wins, or how much it loses when it does not.
The second half of the story is the payoff ratio: the average win divided by the average loss. A payoff of 0.5 means a typical win is half the size of a typical loss. You need both numbers to know whether a strategy makes money.
The arithmetic
Take 100 trades at a 91% win rate, with an average win worth 0.08 of an average loss. The 91 wins bring in 7.28 average losses’ worth of money. The 9 losses cost 9 average losses. The 100 trades end −1.72 average losses down.
That result per trade is the expectancy. It is negative here, even though 91% of the trades won. At this payoff, the strategy needs to win 92.6% of its trades just to break even.
Move the sliders below. The red curve is the win rate each payoff needs to break even. Your win rate is the dashed line. Where the dashed line sits under the curve, the strategy loses money, however high it is.
- Win rate needed to break even
- Your win rate
- Result of 100 trades, in average losses
- −1.7
- Win rate this payoff needs to break even
- 92.6%
- Wins it takes to pay for one loss
- 12.5
Notice how steep the curve gets on the left. When wins are small, every lost trade takes many wins to repay. A strategy there can win 91% of the time and still sit under the line.
A real signal at 99%
One anonymized live signal in our fixtures won 99.1% of its 552 trades. Only 5 of them did not win. It was profitable, and the engine graded it B.
- trades won
- 99.1%
- average win ÷ average loss
- 0.063
- win rate needed to break even
- 94.1%
- above break-even
- 5 points
Its average win was 0.063 of its average loss, so one loss erased about 16 wins. Its largest win was 422; its largest loss was 1,500. The win rate it needed to break even was 94.1%, which is 5 points under what it won.
5 points sounds like room. In trades, it is this: had 28 of its 552 trades lost instead of won, at its average sizes, the whole record would have made nothing. A strategy with few losses also has few chances to show how large a loss can get, so its average loss is the least reliable number on the page.
Across the corpus
We grouped 1,774 live forex signal histories with at least 30 trades by their win rate, and took the median payoff ratio in each group.
High win rates and small wins come together. That is not chance: a strategy that takes profit quickly and lets a loss run will win often and win small. Grid trading and martingale systems work this way by design.
Of the 98 signals that won 90% and over of their trades, 2 lost money on average. That is fewer than in any other group. But 28 of them, 29%, won less than 5 points above their break-even win rate. Monetio’s own flag for this pattern, a win rate of 85% or more with a payoff of 0.33 or less, fired on 124 of the 1,774 histories.
What to do with your own report
- Write the win rate next to the payoff ratio. A win rate on its own answers half a question.
- Work out the break-even win rate: 1 ÷ (1 + payoff). Subtract it from the actual win rate. A few points of margin on a 90% win rate is thin.
- Look at the largest loss next to the largest win. If one loss can erase dozens of wins, the record has not yet been tested by its worst day.
Monetio shows the win rate, the payoff ratio and the expectancy side by side, and flags a high win rate that rests on small wins.