Profit factor formula: how to calculate it and what a good number is

What the number is, exactly

Profit factor is the sum of all winning trades divided by the absolute sum of all losing trades. A profit factor of 1.0 means the wins and losses cancel. Above 1.0 means the strategy made money over the sample; below 1.0 means it lost.

Two properties follow immediately. It is a ratio, so it tells you nothing about scale: a strategy that made £10 and lost £5 has the same profit factor as one that made £10 million and lost £5 million. And it is computed over a sample, so it inherits every problem that sample has.

It is also worth reading the inverse: 1 / profit factor is the loss per unit of profit. A profit factor of 2.0 means you lost 50 cents for every pound won. A profit factor of 1.2 means you lost 83 cents for every pound won. Stated that way, the margin between a good strategy and a marginal one looks considerably thinner than the headline ratio suggests.

The four checks that make it meaningful

1. Trade count. A profit factor of 2.5 on 18 trades is a description of 18 trades. The same figure on 1,800 trades is a description of a process. Always read the count before the ratio; the ratio is a statistic and the count is how much statistics you have.

2. Cost sensitivity. Profit factor is usually computed on gross trade results. Add realistic spread, commission and swap, and re-run. A strategy with many small trades can lose a large share of its profit factor to costs, and a backtest that omits them is reporting a number that does not exist in a live account.

3. Single-trade dependence. Remove the single best trade and recompute. Then remove the best three. If a profit factor of 2.0 falls below 1.0 when two trades are removed, the strategy is not a process, it is a story about two trades. This is the single most revealing check on the list and it takes one line of arithmetic.

4. The distribution behind it. Profit factor compresses an entire distribution into one ratio, so it hides the shape. A strategy with many small wins and three catastrophic losses can post the same profit factor as one with a smooth spread of outcomes — until the moment the tail shows up. Look at the largest loss relative to the average loss, not just the sum.

A concrete example of how it misleads

Take fourteen trades: thirteen winners of 1 unit each, and one loser of 4 units. Gross profit is 13, gross loss is 4, and profit factor is 3.25. That looks like a strong strategy, and it is the shape that high-win-rate strategies take.

Now add realistic costs of 0.1 units per trade — spread and commission, charged on every trade including the winners. Gross profit becomes 13 − 1.3 = 11.7, and the loss becomes 4 + 0.1 = 4.1. Profit factor falls to about 2.83. Still excellent, and the strategy is still fine.

The instructive part is what happens when the shape is slightly different. If the loser is 8 units instead of 4, the pre-cost profit factor is 1.63 and the post-cost figure is about 1.43. One trade changing size moves the headline from very good to merely acceptable, and the report gives no hint that a single observation is doing that much work.

This is why profit factor belongs next to the win rate and the payoff ratio rather than instead of them. The three together describe the shape; the profit factor alone describes only the outcome.

Why it is still worth reporting

Despite all of the above, profit factor does something a win rate cannot: it refuses to let a strategy hide its losses. A strategy that wins 90% of the time and gives it all back in a handful of trades will show a profit factor below 1.0, which is an instant, honest verdict that the win rate would have concealed.

It is also comparable across strategies on the same instrument and window in a way that raw returns are not, because it is normalised by the losses rather than by capital. That makes it a reasonable first-pass screen, provided it is never the only number you look at.

We record it for our own published strategies alongside the window, the bar count and the cost assumption, because a profit factor without those three qualifiers is not a measurement — it is a number that happened under conditions nobody wrote down.

How to read a profit factor in someone else's report

Ask four questions in this order. How many trades? Were costs included, and at what rate? What is the profit factor after removing the best few trades? And what is the largest single loss relative to the average loss?

If the report cannot answer the first two, the number is not comparable to anything and should not influence a decision. If it cannot answer the last two, you do not yet know whether you are looking at a process or an accident.

Then apply the same standard to your own reports. The useful discipline is not to demand better numbers from your strategy; it is to demand more context for the numbers you already have.

What a profit factor cannot tell you

It cannot tell you whether the edge will persist. It is a backward-looking ratio computed on data you already have, and it carries no information about the future beyond what your assumptions put there.

It cannot tell you the risk of ruin. A strategy with a healthy profit factor can still blow up at a position size that is too large for its loss distribution — that is a question about sizing, not about the ratio.

And it cannot be compared across instruments or timeframes without care. Profit factor depends on how many trades the market offered and how large the moves were, both of which differ enormously between, say, an hourly gold strategy and a daily equity index strategy.