The win rate is the percentage of trades that close in profit. It is the first number most people look at, and it is the number most easily made to look good, because it says nothing about how much the winners won or the losers lost.
Why it misleads
A strategy that wins 90 per cent of its trades, making £50 each time, and loses 10 per cent, losing £600 each time, loses money: nine winners earn £450 and one loser costs £600. A strategy that wins 35 per cent, making £300, and loses 65 per cent, losing £100, makes money: 35 winners earn £10,500, 65 losers cost £6,500. The first strategy feels wonderful for weeks at a time. The second feels like failure most days. Only one of them is profitable.
How a high win rate is manufactured
The easiest way to produce a high win rate is to take profits quickly and let losses run. Small targets are hit often; wide stops, or no stops, are hit rarely. Grid and martingale systems take this to its conclusion: they never accept a loss on a single trade, so nearly every trade closes in profit, and the losses accumulate as open positions until they are closed all at once. A record with a 95 per cent win rate and no losing months is almost always one of these, and its equity curve is smooth right up to the day it is not.
Losing streaks follow from the win rate
The probability of a run of losses depends only on the win rate and the number of trades. With a 40 per cent win rate, the chance of a given trade beginning a run of six losses is about 4.7 per cent, so in 200 trades a run of six will almost certainly occur and a run of eight or nine is likely. With a 60 per cent win rate, six in a row is a 0.4 per cent event per trade, still likely at some point in a few hundred trades. Streaks are not evidence that a strategy has stopped working. They are a property of its win rate, and the risk setting has to be chosen with the longest plausible streak in mind.
An EA is advertised with a 92 per cent win rate over 500 trades. Its average winner is £18 and its average loser is £310. Expected result per trade: 0.92 multiplied by £18, minus 0.08 multiplied by £310, is £16.56 minus £24.80, a loss of £8.24 per trade, about £4,100 over the 500 trades. The 92 per cent was true. It was also the reason the strategy lost.
Never assess an EA by its win rate. Ask for the average winner and average loser alongside it, or better, for the live record, from which both can be read. A win rate above 80 per cent should prompt a specific question: how large is the average loser, and how long are losing trades held open? If the answers are "large" and "a long time", the EA is holding losses rather than taking them, whatever its description says.
Conversely, a win rate in the 30s or 40s is not a flaw. Most strategies with a genuine edge in trending or breakout behaviour win less than half their trades and make their money on the size of the winners. The record, not the percentage, is the judge.