Win Rate vs Profitability: Why They're Not the Same

What is the difference between win rate and profitability in trading?

Win rate vs profitability trading questions cause persistent confusion because the two metrics can move in opposite directions: a strategy can win the majority of its trades and still lose money, or win less than half the time and still come out ahead. The reason is simple — win rate counts how often you win, but it says nothing about how much you win or lose each time. A high win rate is not evidence of profitability, and a low win rate is not evidence of failure.

The missing piece is size. A trader who wins 9 out of 10 trades might look impressive until you learn that each win is small and the single loss is large. Conversely, a trader who loses more often than they win can be solidly ahead over time if their wins are several times larger than their losses. Signal providers frequently advertise claimed win rates in their marketing because the number is easy to present and hard to contextualise. The figure that actually determines profitability is expectancy, which combines win rate with the size of wins and losses.

This is why evaluating any signal provider or strategy on win rate alone is unreliable. A headline percentage tells you nothing about the relationship between the typical winning trade and the typical losing trade. Those two numbers, combined with win rate, determine whether the approach tends to add or subtract value over time.

How expectancy ties win rate and trade size together — and what costs do to it

Expectancy is the average result per trade expected over a large number of trades. The standard formula is: expectancy = (win% × average win) − (loss% × average loss). A positive result means the approach tends to add value on average; a negative result means it tends to bleed money regardless of how any single week looks. Win% and loss% are written as decimals — a 40% win rate is 0.40 — where the corresponding losing rate is 0.60.

It is important to recognise that this formula is calculated before costs. Every trade involves spread, commission, and in volatile conditions, slippage. For a signal strategy that targets relatively small price moves, these costs can consume a significant portion of the expected gain on winning trades while adding to the loss on losing trades. A strategy with a raw expectancy of +$8 per trade might deliver real-world expectancy of +$2 per trade after costs — or negative expectancy if the per-trade cost exceeds the margin. Asking what costs look like, and how they affect the formula, is as important as knowing the raw win rate.

• win% and loss% are decimals: a 40% win rate is 0.40 — not automatically a losing rate — and the remaining 0.60 is the loss rate.
• Average win and average loss are the typical profit and loss per trade in your account currency or in R (multiples of amount risked).
• Expectancy is a long-run average, not a prediction for the next trade. Any single trade can still win or lose regardless of the formula.
• Costs reduce real-world expectancy below the raw formula. For high-frequency or small-target signals, costs deserve explicit attention.

A low win rate that is still profitable

Consider an illustrative example. Suppose a strategy wins only 40% of the time — not an impressive-sounding number — but the average winning trade makes $300 while the average losing trade costs $100, because winners are allowed to run and losses are cut quickly at a defined stop.

Plugging those numbers in: expectancy = (0.40 × $300) − (0.60 × $100) = $120 − $60 = +$60 per trade. Despite losing more often than it wins, this illustrative approach has positive expectancy. It works because each win is three times the size of each loss, so the minority of winning trades can still outweigh the majority of losing ones.

This is the engine behind many trend-following approaches. They accept frequent small losses in exchange for occasional larger wins. The emotional cost is genuine — long stretches of losing trades are part of the structure, not a sign something is wrong — but the mathematics can still work as long as wins genuinely outweigh losses and risk per trade is kept consistent. Without controlled sizing, even a positive-expectancy structure can fail in practice.

A high win rate that is not profitable

Now consider the opposite. Imagine a strategy that wins a claimed 90% of the time. On paper that looks nearly unbeatable, and it is exactly the kind of figure often used in marketing. But suppose each win banks a modest $20, while the rare loss runs to $300 because there is no firm stop and the position is held in the hope of recovery.

Expectancy = (0.90 × $20) − (0.10 × $300) = $18 − $30 = −12 per trade. Despite the impressive-sounding claimed 90% win rate, this approach loses money on average. A handful of large losses quietly erase dozens of small wins. This is the mathematical structure behind strategies that produce beautiful win-rate numbers for extended periods and then suffer a single catastrophic loss that undoes months of gains.

Strategies that produce very high win rates often achieve this by taking small profits quickly and refusing to accept losses, letting the occasional losing trade balloon. The win rate stays high right up until one outsized loss changes everything. A high hit rate is only an asset when losses remain controlled relative to wins — without that, win rate is a misleading number.

Why chasing win rate is the wrong goal

Once the expectancy formula is clear, the problem with treating win rate as the primary target becomes obvious. The easiest way to raise a win rate is to take profits early and give losses room to recover. Both habits shrink the average win and inflate the average loss — which can drag expectancy negative even as the percentage of winners climbs. Optimising for win rate can actively worsen the underlying math.

Win rate is also the metric most commonly used in marketing, precisely because it is easy to present attractively and the least informative on its own. A claimed 70% win rate with no information about average win versus average loss, sample size, or how losses are handled tells you almost nothing about whether an approach is sound. Any standalone win-rate figure without those context numbers should be treated with caution.

The more useful question is whether the full picture — win rate, average win, average loss, and costs — produces positive expectancy, and whether that margin is meaningful enough to survive the inevitable variation in results. Win rate is one input into that calculation, not the answer by itself.

Sample size: how many trades before expectancy tells you anything?

Expectancy is a long-run statistical average. That means it is only meaningful over a sufficiently large sample of trades. Over a small number of trades — say, 10 or 20 — random variation can produce results that look very different from the true long-run expectancy in either direction. A strategy with genuinely positive expectancy can still show a losing run of 10 trades. A lucky streak of 10 wins does not prove positive expectancy.

Signal providers sometimes publish track records of 10 to 30 trades and imply a settled, reliable edge. As a general educational benchmark, evaluating expectancy with any useful confidence typically requires at least 50 to 100 completed trades in comparable market conditions — same asset class, similar volatility environment, consistent position sizing. Even then, the estimate carries uncertainty. A provider's published results should be treated as a hypothesis about their edge, not a verified fact, and meaningful assessment requires asking how many trades the track record covers and under what conditions.

This is why running a few trades and concluding that a provider's signals always work or never work is statistically unreliable. The sample is too small for either conclusion to be stable. Building your own journal over time, tracking your actual results across a meaningful number of trades, is the only way to form a grounded view of whether a signal approach is producing positive expectancy for you specifically in your specific conditions.

Expectancy plus consistent risk sizing is what makes the edge real

Positive expectancy in theory only becomes positive expectancy in practice when paired with consistent risk sizing. The formula assumes a roughly predictable loss size on each trade — which is what a defined stop-loss provides — and a roughly consistent stake — which is what fixed-fractional position sizing provides. Without those two constraints, a single oversized loss can wipe out the statistical edge the formula suggested, because the real average loss no longer matches the assumption you used.

A common educational framework is to risk only a small, fixed fraction of the account on each trade — for example, 1% to 2% — so that no single outcome is catastrophic and the expectancy has room to play out across many trades. Pairing that discipline with a firm stop on every position keeps the average loss predictable, which keeps the expectancy calculation honest. This connects directly to risk-reward: the ratio between your average win and your average loss is one of the two levers, alongside win rate, that determine expectancy.

None of this removes the risk of loss. Expectancy is estimated from past results, and past performance does not guarantee future results. Even an approach with genuinely positive historical expectancy goes through losing periods that can be longer and deeper than expected. Results vary and losses are likely for many traders, and the patterns that produced positive expectancy in one market environment may not hold in another. Only risk what you can afford to lose, and treat every expectancy figure — however carefully calculated — as an estimate, not a guarantee.