Win rate vs expectancy: why a high win rate can lose money

Win rate is the share of trades a strategy wins, and expectancy is what it makes or loses per trade on average once the size of the wins and losses is counted. A strategy can win most of its trades and still lose money, so read expectancy first and the win rate next to the payoff.

The win rate is the number traders like to quote. Eight winners out of ten sounds like a strategy that works, and it’s easy to push higher: take profits sooner, move the target closer, and more trades end in the green. Whether the strategy makes money is a separate question, and the win rate can’t answer it on its own.

This guide explains what win rate and expectancy measure, using two strategies: one that wins often and loses money, and one that wins rarely and makes it. Then it covers the win rate you need to break even at a given payoff, what profit factor adds, and the losing streaks that come with a low win rate. It ends with the order in which to read these numbers in a backtest, and how EdgeLuma reports them.

What is win rate in trading?

Win rate is the share of closed trades that ended in profit: the number of winning trades divided by the number of trades. If a strategy closes 50 trades and 20 of them make money, it wins two trades in five.

It’s the easiest number in a report to understand, and it answers a fair question: how often is this strategy right? It doesn’t say how much each trade makes or loses. A win of a few dollars and a win of a few thousand count the same, and so do a small loss and a disaster.

Two details move the number. A trade that closes at exactly zero is neither a win nor a loss, and leaving those trades out of the count pushes the win rate up a little. And the win rate depends on where the target sits. Move the target closer and price reaches it more often, so the win rate goes up while each win gets smaller.

Why can a high win rate lose money?

Because the win rate counts trades, and the money depends on their size. Two strategies show it. Each one risks the same amount on every trade, which traders call 1R (R is the risk of one trade, so a trade that makes twice its risk is +2R), and each closes ten trades:

• Strategy A takes profits early. Its winners make +0.3R, its losers lose the full 1R, and it wins seven trades out of ten.
• Strategy B lets its winners run to three times the risk, +3R, and loses 1R on the others. It wins three trades out of ten.

After ten trades, A has won seven times and is down 0.9R. Its seven winners made 2.1R together, less than its three losers cost. B has lost seven times and is up 2R, because its three winners made 9R and its seven losers cost 7R.

The shape of strategy A is common. When the rules take small profits quickly but give losing trades plenty of room, they win often, and every so often a single loss takes back weeks of winners. Averaging down into a losing position has the same shape: most trades come back, and the few that don’t are very large.

What is expectancy in trading?

Expectancy is the average result of one trade, counting both how often a strategy wins and how much it wins or loses. It’s the expected value of a single trade, and it tells you whether the strategy made money over its trades. Measured in R, it works out like this:

Expectancy = (win rate × average win) − (loss rate × average loss)

For the two strategies above:

• A: (0.7 × 0.3R) − (0.3 × 1R) = 0.21R − 0.3R = −0.09R per trade.
• B: (0.3 × 3R) − (0.7 × 1R) = 0.9R − 0.7R = +0.2R per trade.

Since expectancy is an average per trade, it adds up with every trade you take. Over 100 trades, A should end somewhere around −9R and B around +20R, give or take the luck of the sample. The same averaging is why expectancy needs many trades before it means anything: over ten, a single big winner moves it a lot.

What win rate do you need to break even?

It depends on the payoff, the size of the average win compared with the average loss. At break-even the wins and the losses cancel out, so the win rate times the average win equals the loss rate times the average loss. Solved for the win rate:

Break-even win rate = 1 ÷ (1 + payoff)

With winners as big as losers, a payoff of 1, you need to win one trade in two. With winners three times the size of the losers, one in four is enough. With winners a third the size of the losers, close to strategy A’s, you need to win more than three trades in four.

Costs move the curve. Fees and slippage make every win a little smaller and every loss a little bigger, so the payoff after costs is lower than the one you planned, and the win rate you need goes up. The tighter the stop, the bigger the effect, as our guide to fees and slippage explains.

What is profit factor?

Profit factor is the total made by the winning trades divided by the total lost by the losing ones. Above 1, the strategy made more than it lost. Below 1, it lost more than it made. For strategy A it’s 2.1R ÷ 3R = 0.7, and for strategy B it’s 9R ÷ 7R, about 1.29.

Profit factor and expectancy always agree on the direction. When expectancy is above zero, profit factor is above 1, and the other way round. They answer different questions, though. Expectancy is in R, so it tells you what a trade is worth and you can multiply it by the number of trades you expect to take. Profit factor is a ratio, so it tells you how comfortably the winners cover the losers, whatever the size of the account.

Both have the same weakness: they come from the trades you have. A profit factor of 3 over 12 trades can be a single lucky winner, so read either number next to the number of trades.

Win rate, expectancy and profit factor compared

Each of the four numbers answers one question and leaves another open:

How long can a losing streak get?

Longer than most traders expect, and much longer with a low win rate. The arithmetic is the one Mark Schilling worked out for coin tosses in The Longest Run of Heads (The College Mathematics Journal, 1990), with a losing trade in place of heads.

We simulated 100 trades many times over for both strategies, with every trade independent of the one before. A strategy that wins three trades in ten typically loses about ten in a row somewhere in those 100 trades, and eight or more in most runs. One that wins seven trades in ten usually tops out at three losses in a row.

That’s what a low win rate costs. Strategy B makes money over 100 trades, and somewhere in them it will very likely lose ten trades in a row, a fall of 10R from a strategy that works. If you know that stretch is coming, you can size each trade so that ten losses in a row is something you would sit through.

A high win rate has the opposite problem. Its losing streaks are short, so it feels safe, and its rare losses are what sink it: each of strategy A’s losses wipes out more than three of its wins.

How to read these numbers in a backtest

Read them in this order, so that each number frames the next:

1. Start with the number of trades. Under a hundred or so, every other number is mostly luck, as our guide to backtest overfitting explains.
2. Read expectancy after costs. Above zero, the rules made money on average over those trades. Below zero, no win rate makes up for it.
3. Read the win rate next to the payoff. Together they tell you how the strategy earns its expectancy: often and small, or rarely and large.
4. Check the maximum drawdown and the longest losing streak. They are the stretches you would have to live through, and a low win rate makes both longer.
5. Open the trades. If the expectancy rests on two or three outsized winners, take them out and look at what’s left.

How EdgeLuma reports these numbers

EdgeLuma counts every result in R, after fees and slippage. 1R includes the costs of the round trip, so a trade stopped at its original stop is exactly −1R. The top of every backtest report shows four numbers:

• Expectancy, the average result per trade in R, after costs.
• Total result, every closed trade’s R added up, with the number of closed trades under it.
• Win rate, with the average win and the average loss in R under it.
• Max drawdown in R, with the longest losing streak under it.

The run’s details add the payoff, the average win divided by the average loss. A trade that closes at exactly zero stays in the count, so it lowers the win rate without being a loss. The report doesn’t print a profit factor, but the list of trades gives every result in R if you want to work it out.

To understand a result, attach the run in the Chat and ask the EdgeLuma AI assistant, for example, why the win rate is high and the expectancy negative. It answers from the run’s own trades. Like any backtest, a report replays the past: it doesn’t tell you what to trade, and nothing in a report is financial advice.