Win Rate vs Expectancy

Why the number everyone quotes tells you the least about whether a system makes money.

Win rate is a common number most traders reach for and yet is one of the least useful on its own. It tells you how often you win, and nothing about how big those wins are, how big the losses are, or how often you get to trade.

Expectancy is what makes money

Expectancy is your win rate times your average win, minus your loss rate times your average loss.

expectancy = (win rate x average win) - (loss rate x average loss)

It accounts for the size of the wins and losses, not only how often they land. Work it in R-multiples, profit or loss as a multiple of the risk you put on, and it gets concrete.

Two systems, same question

System A wins 30% of the time. Sounds dreadful. Its winners are four times the size of its losers, so (0.3 x 4) - (0.7 x 1) gives +0.5R per trade. A strong, durable book. Most trend-following lives here: wrong well over half the time and printing money across a year.

System B wins 90% and hands almost all of it back on the losing 10%, because the losers are uncapped. Ninety percent at +1R against ten percent at -9R nets to zero, and the first time a loser runs to -10R the whole thing tips negative. A high win rate with no control on loser size is a blow-up on a timer.

The input win rate hides

Frequency. Expectancy per trade x the number of trades you take is what grows your account. A 0.2R edge taken 500 times a year buries a 1R edge you only get to express twice. Win rate is silent on this axis entirely.

So win rate is one input out of three: how often you win, how big the wins are against the losses, and how often you play. It happens to be the least informative of the three, and yet the one that people get obsessed with.

Why win rate still feels so important

Being right is satisfying in a way that grinding out a profit while losing most days never is. "I lose 65% of my trades and I still make money" is true, durable, and yet is a hopeless thing to put on a sales funnel page. Win rate scratches that itch for you and generally makes gurus sound better than they are at trading.

What to track instead

Expectancy is what actually pays, even on the days it makes you look wrong.

Expectancy in R-multiples is the first thing I want to see on anyone's record. Then the full distribution of your R outcomes, not just the average, because the tails are where the account-enders live. Then whether both survive out of sample, because an edge that only exists in the backtest was well... never an edge. Get those three honest and win rate drops to a footnote.

None of this makes a high win rate bad. A high win rate with controlled losers and decent frequency is a lovely thing to trade. The number is not the problem. Reading it in isolation is.