Win rate alone means nothing
A system that wins 70% of the time sounds excellent. A system that wins 35% of the time sounds terrible. Either can make money or lose money: it depends on how big the wins are compared with the losses, and on costs. To judge a win rate you need its partner, the risk-reward ratio, and you need both in R.
Risk-reward (RR) is the average win divided by the average loss. A target twice as far as the stop gives RR = 2: a win is +2R and a loss is -1R.
The break-even formula
The cost c is the spread plus commission (and swap, if you hold overnight) for one trade, measured in R. If your stop is 10 pips and the spread is 0.6 pips, c = 0.06R.
A table to keep
| Risk-reward | No costs | Cost 0.05R per trade | Cost 0.12R per trade |
|---|---|---|---|
| 0.5 (target half the stop) | 66.7% | 70.0% | 74.7% |
| 1 | 50.0% | 52.5% | 56.0% |
| 1.5 | 40.0% | 42.0% | 44.8% |
| 2 | 33.3% | 35.0% | 37.3% |
| 3 | 25.0% | 26.2% | 28.0% |
A cost of 0.12R is what a 0.6-pip spread costs on a 5-pip stop. Our 39-strategy assessment found a median spread cost of 0.128R per trade on the 15-minute chart and 0.057R on the one-hour chart, so these columns are realistic.
Read the table across a row to see what costs do: at a risk-reward of 1, a cost of 0.12R per trade lifts the break-even win rate from 50% to 56%. Read down a column to see why low-win-rate systems need large targets: at a risk-reward of 3, winning only about a quarter of the time can break even. Neither direction is better in itself; what matters is whether your real win rate sits clearly above the break-even line for your real risk-reward and costs, over enough trades to rule out luck.
Worked example: the high-win-rate scalper
Without the spread: win +3 pips, loss -6 pips, RR = 0.5
Break-even win rate = 1 / (1 + 0.5) = 66.7%
With the spread: win = 3 - 0.6 = 2.4 pips, loss = 6 + 0.6 = 6.6 pips
Break-even win rate = 6.6 / (2.4 + 6.6) = 73.3%
Same check in R: c = 0.6 / 6 = 0.1R, so (1 + 0.1) / (0.5 + 1) = 73.3%
A system that wins 70% looks excellent and still loses money
The spread raised the win rate needed just to break even by more than six percentage points. That is why systems that advertise only a win rate are showing you the number that hides the spread best.
Expectancy: the one number that combines them
Win rate and risk-reward are two halves of one number: the expectancy, or average result per trade. In R it is: expectancy = (win rate x average win) - (loss rate x average loss), with wins and losses measured after costs.
Win rate 45%, average win +1.8R, average loss -1.05R (stop plus costs)
Expectancy = 0.45 x 1.8 - 0.55 x 1.05
= 0.81 - 0.5775 = about +0.23R per trade
Break-even check: RR = 1.8 / 1.05 = about 1.71; break-even win rate = 1 / (1 + 1.71) = about 37%
45% is above 37%, consistent with a positive expectancy
A positive expectancy on paper is the start of the question, not the answer. The next questions are whether it holds on trades you did not use to design the rule, whether it beats random copies of the same trades, and whether it survives 1.5 x the spread. In our 39-strategy assessment, the one strategy that met all of those in its original test stopped passing once overnight financing was charged.
How to use this on your own trades
- From your journal or backtest, take the average win and the average loss in R, after costs.
- Compute RR = average win / average loss.
- Compute the break-even win rate: 1 / (1 + RR).
- Compare it with your actual win rate over at least a hundred trades.
- Better still, skip the ratio and compute the average result per trade in R after costs. If it is positive and stays positive at 1.5 x the spread, the win rate and RR are working together; if not, they are not.
Limits
Real results also have a spread of outcomes: trades exit early, stops slip past their level in fast markets, and targets are not always reached exactly. Use actual average wins and losses from your records, not the planned ones.
Key takeaways
- Break-even win rate = 1 / (1 + RR); with costs it is (1 + c) / (RR + 1).
- A 0.6-pip spread on a 6-pip stop with a 3-pip target raises the break-even from 66.7% to 73.3%.
- Judge win rate and risk-reward together, in R, after costs.
- The average result per trade after costs is the number that decides everything.
Education only, not investment advice. Results quoted from TradeLab research are historical simulations; trading can lose more than you expect.
