Why win rate misleads

Win rate is the most quoted number in trading and one of the least useful on its own.

Here are two strategies. Which is better?

	Strategy A	Strategy B
Win rate	70%	35%
Average win	$50	$400
Average loss	$200	$100
Per 100 trades	−$2,500	+$7,500

Strategy A is right more than twice as often and loses money. Strategy B is wrong about two times in three and makes a great deal.

Nothing about A feels wrong while you’re trading it, which is the problem. Seven wins out of ten is lovely. The three losses each quietly undo four wins.

Why high win rates are seductive and dangerous

It’s very easy to build a strategy that wins 90% of the time: take a tiny profit whenever you’re ahead, and hold every loser until it comes back. That feels wonderful — a long run of green — right up until the one that never comes back.

That shape has a name in the trade: picking up pennies in front of a steamroller. The equity curve climbs steadily and then falls off a cliff, and everything up to the cliff looked like skill.

What to look at instead

Expectancy per trade — what one average trade is worth, wins and losses together. That single number tells you whether the strategy makes money.

The ratio of average win to average loss. A strategy winning 35% of the time needs its winners to be about twice its losers to break even. One winning 70% can afford winners half the size of its losers. Neither is better; they just need different shapes.

The worst stretch, not the average. See drawdown.

On the Review screen

BotStock’s trade summary shows you a win rate and a net profit or loss for the period. Read them together, always — and if they disagree, believe the second one. A 68% win rate sitting next to a negative number is the table above, happening to you.

The per-trade detail is on the same screen. If those two figures are pulling in opposite directions, sorting by size and looking at your three largest losses will usually explain it in about a minute.