Trading systems are marketed on win rates because a high one is easy to understand and sounds like competence. It is half of a number. A method winning ninety percent of the time can lose money steadily, and one winning thirty percent can compound, and knowing which requires the other half.
The relationship in one line
- Expectancy = (win rate × average win) − (loss rate × average loss).
- A positive expectancy is the whole requirement; the split between its terms is a style choice.
- High win rate usually means small wins and occasional large losses.
- Low win rate usually means frequent small losses and rare large wins.
The arithmetic, worked
Take a method winning ninety percent of trades, with an average win of ten units and an average loss of a hundred. Over a hundred trades: ninety wins of ten is nine hundred; ten losses of a hundred is a thousand. The method loses a hundred units per hundred trades while being right nine times out of ten.
Now take one winning thirty percent, with an average win of two hundred and an average loss of fifty. Over a hundred trades: thirty wins of two hundred is six thousand; seventy losses of fifty is three thousand five hundred. It makes two and a half thousand while being wrong more than twice as often as it is right.
Neither is a trick. Both describe real, common strategy shapes, and both illustrate the same point: the win rate on its own carries no information about profitability.
The trade-off is structural
You cannot generally raise both terms. Taking profits earlier raises the win rate and lowers the average win; letting winners run raises the average win and lowers the win rate, because more trades come back through the entry before reaching a distant target. Widening the stop raises the win rate and the average loss together.
Every parameter change moves both numbers, usually in opposite directions. This is why optimising a strategy on win rate produces worse results than optimising on expectancy, and why a trader who tightens targets to feel better about their hit rate frequently converts a profitable method into a losing one without changing anything else.
Why the high-win-rate shape is psychologically dangerous
A method with a ninety percent hit rate and a large average loss is comfortable almost all of the time. The rare loss arrives after a long run of wins, when confidence and often position size are at their highest, and it is large enough to remove several months of accumulated gains.
Grid and martingale approaches are the extreme case: they can show a hit rate near a hundred percent for extended periods because losing positions are held rather than closed, and the loss when it arrives is the whole account. The equity curve looks like a straight line right up until it does not.
A published win rate without an average win-to-loss ratio is not a claim
Any system, signal service or strategy advertised on hit rate alone has omitted the number that would let you evaluate it. Ask for average win, average loss, and largest loss. If those are not available, the win rate is not evidence of anything.
Variance and how long you need
Expectancy is a long-run average, and the shorter the sample the less it tells you. A method with a thirty percent win rate will produce runs of eight or ten consecutive losses regularly — not as a malfunction but as ordinary probability. A trader who abandons the method during such a run has made a decision on noise.
This is why the low-win-rate shape demands more capital and more discipline than its expectancy alone suggests. The arithmetic is better; the experience is worse, and the experience is what people act on.
Measuring your own numbers honestly
- Take at least a hundred closed trades from your actual record, not a backtest.
- Compute the win rate, the average win, and the average loss separately.
- Compute expectancy per trade and multiply by your trade frequency.
- Record the longest losing streak in the sample.
- Set position size so that streak, plus a margin, is survivable.
- Recompute quarterly rather than after every trade.
Which shape to aim for
Neither is correct in general, and the choice should follow temperament as much as arithmetic. A trader who cannot sit through a ten-loss streak should not run a low-win-rate method regardless of its expectancy, because they will abandon it during the streak and take the losses without the recovery.
What is not optional is knowing both numbers. A trader who cannot state their average win, average loss and win rate from their own record does not know whether their method makes money, and no amount of confidence in individual trades substitutes for that.
Common questions
What is a good win rate?
There is no such thing in isolation. A 30% win rate with wins four times the size of losses is excellent; a 90% win rate with losses ten times the size of wins is a losing system.
Why do high win rate systems blow up?
Because the shape usually pairs frequent small wins with rare large losses, and the large loss arrives after a long comfortable run when position sizes have often been increased.
How many trades before my numbers mean something?
At least a hundred closed trades, and more for a low win rate, where streaks are long and a small sample can be dominated by them. Fewer than fifty tells you almost nothing.
Should I aim for a fixed risk-reward ratio?
A fixed minimum ratio is a useful filter, but forcing every trade to a ratio the market does not offer produces targets that are never reached. The ratio should come from where the structure allows the exit, not from a rule applied first.
