Trading performance metrics explained: expectancy, win rate, payoff ratio and more

The trading metrics that matter most are expectancy (your average result per trade, measured in R), profit factor (gross wins divided by gross losses) and maximum drawdown (the largest fall from a peak). Win rate on its own is misleading, because a strategy that wins 70% of the time can still lose money if its losses are much larger than its wins. And every metric depends on sample size: a few dozen trades say far less about a strategy than most traders assume.

This guide defines each metric, shows how they relate, and then works through a 20-trade example journal with every number calculated, including what those 20 trades can and cannot tell you.

Why R is the right unit

R is the amount you risked on a trade: the distance from entry to stop loss, expressed in money. If you risk 100 and make 250, the trade was +2.5R. If you lose the full 100, it was minus 1R. If you close early for a loss of 50, it was minus 0.5R.

Measuring in R rather than money or pips does two things. It makes trades comparable when position sizes differ, and it separates the quality of your decisions from the size of your account. A strategy with an expectancy of +0.2R per trade earns 20 per trade at 100 risk and 200 per trade at 1,000 risk. The position sizing guide covers how to set R consistently, and the trading journal guide covers how to record it.

Win rate

Win rate is the number of winning trades divided by the total number of trades. It is the metric people quote first and the one that misleads most often, because it ignores size. A 40% win rate is excellent if winners are three times the size of losers and terrible if they are half the size.

Decide in advance how to count scratches (trades closed at roughly zero). Counting them as losses, wins or neither changes the win rate without changing anything else, so be consistent.

Average win, average loss and payoff ratio

Average win is gross wins divided by the number of winning trades; average loss is gross losses divided by the number of losing trades. The payoff ratio is average win divided by average loss. It is related to the planned risk-reward ratio but not the same thing: risk-reward is what you plan before entry, payoff ratio is what actually happened after early exits, partial closes and slippage.

Win rate and payoff ratio pull against each other. Wider targets mean bigger wins but fewer of them. The win rate you need to break even before costs is 1 divided by (1 + payoff ratio). With a payoff ratio of 2, you need to win just over 33% of trades; with a payoff ratio of 0.5, you need to win just under 67%.

Expectancy in R

Expectancy is the average result per trade. In R:

• Expectancy = (win rate × average win in R) minus (loss rate × average loss in R)

Or, more simply, total R divided by the number of trades. A positive expectancy means that, over the sample, the strategy made money per trade on average after whatever costs are included in the results. It is the number every other metric is in some way describing.

Expectancy per trade is not the whole story, because a strategy with +0.5R per trade that trades twice a month may earn less over a year than one with +0.15R that trades twenty times a month. Multiply expectancy by the number of trades per period to compare them, and be sure the more active strategy's costs are fully included.

Profit factor

Profit factor is gross wins divided by gross losses (as a positive number). If your winners made 1,550 in total and your losers lost 1,150, profit factor is 1,550 divided by 1,150, or about 1.35.

• Below 1: the strategy lost money over the sample.
• Exactly 1: it broke even.
• Above 1: it made money over the sample.

Profit factor and expectancy are two views of the same data. Profit factor equals (win rate × average win) divided by (loss rate × average loss). It is useful because it is independent of trade count and position size, and because it shows the margin of safety: at 1.1, costs only slightly higher than assumed could turn the result negative, whereas at 1.5 there is more room. It says nothing about how the gains and losses were distributed in time, which is what drawdown measures.

What counts as "good" depends on the style. Strategies that trade often with small targets tend to show lower profit factors than those that trade rarely with large targets. A very high profit factor on a small sample is more often a sign of luck or curve-fitting than of an exceptional strategy.

Maximum drawdown

Maximum drawdown is the largest fall in equity from a peak to a subsequent low, measured in money, percentage or R. It answers the question "how bad did it get?". Two strategies with the same expectancy can have very different drawdowns depending on how wins and losses cluster.

Drawdown matters for two reasons. Psychologically, a deep drawdown is when traders abandon or change a strategy. Mathematically, losses require larger gains to recover: a 20% drawdown needs a 25% gain to get back to the peak. The drawdown and recovery guide covers this in detail, and the drawdown calculator shows the recovery needed for any drawdown.

The maximum drawdown in your records is a lower bound on what the strategy can do, not an upper bound. The worst drawdown in a sample is almost always exceeded eventually if you keep trading. The risk of ruin guide explains how to estimate the chance of a much deeper fall.

Recovery factor

Recovery factor is net gain divided by maximum drawdown. A recovery factor of 3 means the strategy made three times its worst peak-to-trough fall over the period. It rewards strategies that earn a lot relative to the pain along the way. Like every metric, it depends heavily on the length of the sample: a longer period gives more time for net gains to build but also more time for a larger drawdown to appear.

Sharpe ratio basics

The Sharpe ratio measures return per unit of volatility. It is calculated from periodic returns (usually daily or monthly):

• Sharpe ratio = (average period return minus the cash rate for that period) divided by the standard deviation of period returns

The cash rate is the return you could have earned on cash or very short-term government bills over the same period. Monthly figures are usually annualised by multiplying the monthly ratio by the square root of 12.

Two cautions for traders. First, the Sharpe ratio treats upside and downside volatility alike, so a strategy with occasional large gains is penalised as if they were losses. The Sortino ratio, which uses only downside volatility, is a common alternative. Second, it needs many periods to be meaningful. A Sharpe ratio based on four or six months of returns can look spectacular purely by chance, as the worked example below shows.

Worked example: a 20-trade journal

Illustrative assumptions: a 10,000 account, a fixed risk of 100 per trade (1R = 100, not recalculated as the balance changes), and results already net of costs. The 20 trades, in order, with the running balance:

1. +2R, +200, balance 10,200
2. minus 1R, minus 100, balance 10,100
3. minus 1R, minus 100, balance 10,000
4. +1.5R, +150, balance 10,150
5. minus 1R, minus 100, balance 10,050
6. +2.5R, +250, balance 10,300
7. minus 1R, minus 100, balance 10,200
8. minus 1R, minus 100, balance 10,100
9. +2R, +200, balance 10,300
10. minus 0.5R, minus 50, balance 10,250
11. +1R, +100, balance 10,350
12. minus 1R, minus 100, balance 10,250
13. minus 1R, minus 100, balance 10,150
14. minus 1R, minus 100, balance 10,050
15. +3R, +300, balance 10,350
16. minus 1R, minus 100, balance 10,250
17. +2R, +200, balance 10,450
18. minus 1R, minus 100, balance 10,350
19. +1.5R, +150, balance 10,500
20. minus 1R, minus 100, balance 10,400

The metrics

• Win rate: 8 winners out of 20 trades = 40%.
• Gross wins: 2 + 1.5 + 2.5 + 2 + 1 + 3 + 2 + 1.5 = 15.5R, or 1,550.
• Gross losses: eleven trades of minus 1R and one of minus 0.5R = 11.5R, or 1,150.
• Net result: 15.5R minus 11.5R = +4R, or +400, a 4% gain on the starting balance.
• Average win: 15.5R divided by 8 = 1.94R (193.75).
• Average loss: 11.5R divided by 12 = 0.96R (95.83).
• Payoff ratio: 1.9375 divided by 0.9583 = 2.02.
• Break-even win rate for that payoff ratio: 1 divided by 3.02 = 33.1%. The actual 40% is above it, which is why the result is positive.
• Expectancy: 4R divided by 20 trades = +0.2R per trade, or +20. Check with the formula: 0.4 × 1.9375 minus 0.6 × 0.9583 = 0.775 minus 0.575 = 0.2R.
• Profit factor: 1,550 divided by 1,150 = 1.35. Check: 0.775 divided by 0.575 = 1.35.
• Longest losing streak: 3 trades (trades 12 to 14).
• Maximum drawdown: the balance peaked at 10,350 after trade 11 and fell to 10,050 after trade 14, a drawdown of 300, or 3R, or 2.9% of the peak.
• Recovery factor: net gain 400 divided by maximum drawdown 300 = 1.33.

A Sharpe ratio from this journal

Suppose the 20 trades took four months, five trades per month. The monthly results were +0.5R, +2R, +1R and +0.5R, which as a percentage of the balance at the start of each month are about 0.50%, 1.99%, 0.98% and 0.48%. Assume an illustrative cash rate of 3% a year, or 0.25% a month.

• Monthly excess returns: 0.25%, 1.74%, 0.73% and 0.23%, an average of 0.74%.
• Standard deviation of the monthly returns: about 0.71%.
• Monthly Sharpe ratio: 0.74 divided by 0.71 = about 1.04.
• Annualised: 1.04 × the square root of 12 (about 3.46) = about 3.6.

An annualised Sharpe ratio of 3.6 would be remarkable for a professional fund over many years. Here it comes from four data points and a 4% gain, and one bad month would change it completely. That is the point of the exercise: the formula always produces a number, but the number is only as good as the sample behind it.

What 20 trades can actually tell you

The per-trade results in this journal have a standard deviation of about 1.51R. The standard error of the expectancy, a measure of how far the sample average could be from the strategy's true average, is the standard deviation divided by the square root of the number of trades: 1.51 divided by the square root of 20 (about 4.47) = about 0.34R.

A rough 95% range for the true expectancy is the sample figure plus or minus two standard errors: 0.2R plus or minus 0.67R, or roughly minus 0.47R to +0.87R. In plain terms, these 20 trades are consistent with a strategy that loses nearly half an R per trade and with one that makes nearly nine-tenths of an R. They cannot tell you whether the edge is positive.

How many trades would it take for an expectancy of +0.2R to be two standard errors above zero, assuming the same spread of results? The standard error would need to be 0.1R, so the square root of the number of trades must be 1.51 divided by 0.1, about 15.1, which means around 230 trades. That is a rough rule of thumb rather than a precise test (it assumes trades are independent and the strategy does not change), but it shows why one month of results proves very little.

Sample size: why 30 trades says little

The same arithmetic applies to every metric. With 30 trades:

• Win rate is uncertain by a wide margin. If the true win rate is 40%, the standard error of the measured win rate over 30 trades is the square root of (0.4 × 0.6 divided by 30), about 8.9 percentage points, so a measured result anywhere between about 22% and 58% would not be surprising.
• Profit factor can swing far above or below its long-run value on the strength of one or two large winners.
• Maximum drawdown will usually be smaller than what the strategy produces over a longer period, simply because there has been less time for a bad run.

A few practical consequences follow. Do not change a strategy after one losing week, and do not raise size after one winning week. Judge changes on hundreds of trades where possible, and combine live results with backtesting and forward testing on a demo account to build a larger sample. Be especially wary of any metric reported for a small sample in marketing material, including signal services and account managers; the forex scam red flags guide covers that.

How the metrics fit together

A useful way to read a set of results:

1. Is expectancy positive after costs? If not, nothing else matters.
2. How confident can I be? Check the number of trades and the standard error.
3. What is the margin of safety? Profit factor shows how much worse costs or execution could get before the edge disappears.
4. How painful was the path? Maximum drawdown, longest losing streak and recovery factor.
5. Is it efficient? Sharpe ratio over a long run of periods, and expectancy multiplied by trades per month.

Then size positions so that a drawdown worse than the worst you have seen remains survivable. The risk of ruin calculator lets you test how win rate, payoff and risk per trade interact.

Common mistakes

• Optimising for win rate. Taking gains early and letting losers run raises the win rate and destroys expectancy.
• Measuring in pips. A 20-pip win on a tiny position and a 20-pip loss on a large one are not equivalent. Use R or money.
• Leaving costs out. Spread, commission, swap and slippage belong in the results. A profit factor of 1.15 before costs can easily be below 1 after them.
• Changing the rules mid-sample. If you change entries, exits or sizing, start a new sample; mixing versions makes every metric meaningless.
• Cherry-picking the period. Metrics computed from the best three months are not the strategy's metrics.
• Treating a short-run Sharpe ratio or recovery factor as proof. They need long samples.

Frequently asked questions

What is a good profit factor?

There is no universal threshold. Above 1 means the strategy made money over the sample; many traders look for comfortably above 1 after all costs, so that small changes in execution do not wipe out the edge. A very high profit factor on few trades is more often luck or curve-fitting than skill. Judge it together with sample size and drawdown.

How do you calculate expectancy in trading?

Add up the results of all trades in R and divide by the number of trades. Equivalently, multiply the win rate by the average win and subtract the loss rate multiplied by the average loss. In the example journal, 4R over 20 trades gives +0.2R per trade.

Is a high win rate good?

Only in combination with the payoff ratio. A 70% win rate with average losses three times the size of average wins loses money. A 35% win rate with winners three times the size of losers makes money. Look at expectancy, not win rate alone.

How many trades do I need to evaluate a strategy?

More than most people expect. With a typical spread of results, the uncertainty around expectancy after 30 trades is usually larger than the expectancy itself. A few hundred trades, from live trading, forward testing and careful backtesting combined, give a much more reliable picture.

What is the difference between profit factor and the Sharpe ratio?

Profit factor compares gross wins with gross losses across trades and ignores timing. The Sharpe ratio compares average period returns above the cash rate with the volatility of those returns, so it describes how smooth the equity curve was. A strategy can have a decent profit factor and a poor Sharpe ratio if its results are very uneven over time.

What is recovery factor?

It is net gain divided by maximum drawdown over the same period. In the example journal, a net gain of 400 and a maximum drawdown of 300 give a recovery factor of 1.33. Higher is better, but it depends heavily on the length of the sample.