Profit factor is the ratio of gross profits to gross losses across a set of trades.

Profit factor is the ratio of gross profits to gross losses across a set of trades. It measures how much profit is generated for each unit of loss before trading costs are applied. A profit factor of 1.0 means gross profits equal gross losses. A value above 1.0 means gross profits exceed gross losses. A value below 1.0 means losses exceed profits.

This ratio is useful because it is simple and directly tied to trade outcomes. If a strategy produces 105 units of gross profit and 100 units of gross loss, its profit factor is 1.05. That means the strategy has only a small gross edge. The difference between profits and losses is narrow.

That narrow gap matters because gross results are not the same as net results. Real trading includes commissions, bid ask spreads, slippage, and market impact. These frictions reduce realized performance. When profit factor is only slightly above one, there is very little room for those costs before the strategy moves from a small gross profit to a net loss.

The mechanism is straightforward. Profit factor compares two gross quantities, profits and losses, but trading frictions act like an additional drag on the trade distribution. They either reduce winning trades, deepen losing trades, or both. As a result, the small excess of gross profits over gross losses can disappear once costs are included. A strategy with a profit factor barely above one therefore has little error tolerance. Small changes in execution quality, spread conditions, or assumptions used in a backtest can materially change the conclusion.

This is closely related to the general problem of backtest fragility. Sonar’s material on overfitting and multiple testing emphasizes that in-sample performance can look better than what survives out of sample, and that apparent edges should be discounted when they are small and statistically fragile. The same logic applies to profit factor. If the ratio is only marginally above one, any optimism in the backtest, including cost underestimation or model selection bias, can be enough to erase the edge.

A higher profit factor provides a larger cushion against this erosion. The concept is not that any single threshold guarantees robustness. It is that more distance between gross profits and gross losses leaves more room for inevitable implementation frictions and estimation error. Conversely, a profit factor that barely clears one offers essentially no margin of safety.

Covered in depth in the Strategy research fundamentals pillar hub.