Win Rate in Investing: Why a High Hit Rate Can Still Lose Money

Win rate is the share of your positions that ended in a profit. Close twenty trades, and if twelve of them made money, the win rate is 60 percent. It is the most intuitive number in any strategy report and one of the least informative on its own, because being right often and making money are not the same thing.

The reason is simple arithmetic. Win rate counts outcomes. It does not weigh them. A strategy that is right nine times out of ten can still lose money if the tenth outcome wipes out the other nine, and a strategy that is right three times in ten can do well if those three are large enough. Until you know the sizes, the frequency tells you very little.

What it measures

The definition in words: take the number of closed positions that ended above their entry cost, divide by the total number of closed positions, and express it as a percentage. Losing positions are the remainder. Positions that closed exactly flat are usually counted as neither, or bundled with losses, and the convention varies between reports.

That is the whole calculation. Note what it deliberately excludes:

• The size of each result. A one rupee gain and a hundred rupee gain are the same single win.
• The capital committed. A win on a small position counts as much as a win on a large one.
• Time. A position held for a day and one held for three years are one win each.
• Open positions. Win rate is computed on closed trades, so anything still running is invisible.

How to read it

Win rate becomes useful the moment you pair it with the payoff ratio, which is the average size of a winning trade divided by the average size of a losing trade. Together the two describe the economics of the approach. Separately, neither does.

The relationship is easy to reason about with a hypothetical. Suppose an approach wins on 40 percent of its positions, and the average winner is three times the average loser. Over ten trades, that is four wins of three units each, so twelve units gained, against six losses of one unit each, so six units lost. Net six units, from an approach that was wrong more often than it was right. Now flip it. Suppose an approach wins on 80 percent of its positions, but the average loser is six times the average winner. Over ten trades, eight wins of one unit is eight units, against two losses of six units, which is twelve. Net negative four units, from an approach that was right most of the time. Both examples are illustrative arithmetic, not results from any real strategy, but they show the point cleanly: the frequency and the magnitude have to be read together.

This is why win rate and profit factor belong side by side. Profit factor already combines frequency and size into one figure. Win rate tells you which of the two is producing it.

A second reading habit is to ask what the win rate is a consequence of. Win rate is not really a choice; it is a by-product of the exit rules. An approach that closes positions at a small predetermined gain and holds losers hoping for recovery will mechanically produce a high win rate and a poor payoff ratio. An approach that cuts losses quickly and lets winners run will produce the opposite. Neither the high number nor the low one is evidence of skill. They are evidence of a design decision.

Third, check the sample. A win rate from thirty positions carries very little information, because the natural variation in outcomes is wide at that sample size. Over hundreds of positions the number stabilises and starts to mean something. A win rate quoted without a trade count is not a statistic, it is a headline.

Fourth, look at the win rate across sub-periods rather than for the whole history. Many approaches have win rates that swing dramatically with market conditions, high in trending stretches and low in choppy ones. A single average across a decade can conceal two completely different behaviours.

The most useful question about a win rate is not whether it is high. It is whether it is consistent with the payoff structure the approach was designed to produce, and whether it holds up outside the period the rules were built on.

What it does not tell you

It does not tell you whether the approach made money. This is the headline limitation and it deserves stating flatly. Win rate and profitability are separate facts. You cannot infer one from the other without knowing the sizes.

It does not tell you about the worst outcome. Win rate is blind to the tail. An approach can have a very high win rate and still contain the possibility of a single loss large enough to matter permanently. For that you need the loss distribution, the largest single loss, and maximum drawdown.

It says nothing about sequence. Ten losses spread evenly through a decade feel very different from ten losses in a row. Win rate compresses the ordering out of existence, and ordering is exactly what determines whether a drawdown is survivable in practice.

It can be inflated by holding losers open. Because the measure only counts closed positions, an approach that refuses to realise a loss can report an excellent win rate while carrying unrealised damage. Any statistic computed on closed trades alone is vulnerable to this, whether the behaviour is deliberate or not.

It is heavily affected by costs at small trade sizes. Brokerage, securities transaction tax, stamp duty and slippage turn marginal winners into small losers. A gross win rate can be noticeably higher than the same approach measured net, and the gap widens with turnover. This is one more reason to check whether a report shows gross or net trade results.

It does not survive translation between styles. Comparing the win rate of a short-holding-period approach with that of a multi-year one is not a meaningful comparison, because a trade means a different thing in each. Win rates are only comparable within similar designs.

It is not a forecast. A win rate measured on a past sample describes that sample. Rules tuned until the win rate looked good on historical data are a textbook case of overfitting, and the number that results tends not to travel.

The honest summary is that win rate is a description of style, not of quality. It tells you what kind of experience an approach produces: frequent small confirmations, or long stretches of being wrong punctuated by a few large payoffs. That is genuinely worth knowing, because people abandon approaches whose rhythm they cannot tolerate. Just do not mistake it for evidence that the approach works.

Related reading

• Portfolio and Backtest Metrics, Explained: the hub guide to the numbers in a strategy report and what each one leaves out.
• What Is Profit Factor: the ratio that combines how often you win with how much you win.
• Average Trade Profit, Explained: why the average result per position can be misleading when a few outcomes dominate.
• Gross Profit and Loss, Explained: the costs that sit between a gross result and a net one.
• Common Backtesting Mistakes: the recurring errors that make historical statistics look better than reality.

This article is educational. Altys Labs is not a registered research analyst or investment adviser, and nothing here is investment advice or a recommendation to buy, sell, or hold any security.