Analysis & metrics

Win rate

By Jarosław Wasiński • Published: Apr 9, 2026 • Updated: Oct 11, 2026

The proportion of trades ending in profit in a defined sample. Specify the treatment of costs, break-even trades and partial executions. Win rate alone does not establish profitability: 90% winners of one unit and 10% losers of ten units average −0.1 unit. Conversely, 35% winners of three units and 65% losers of one unit average +0.4 before additional costs. These are assumed payoffs, not outcomes promised by a target distance.

Example. Five complete trades after costs finish at +100, +80, −60, 0 and +100 USD. Winners are 3/5 = 60%. Explicitly excluding the zero result gives 3/4 = 75%; these figures use different conventions. Do not choose the more attractive denominator without telling the reader.

Define the trade first

Specify whether an observation is a complete position, an individual fill or a group of linked orders. Three partial exits from one entry are not automatically three independent winners. Decide how to treat zeros, positions still open and financing. A break-even result before commission can become a loss after charges. Report winning and total observation counts, rather than only a rounded percentage.

Win rate is not profitability

With assumed payoffs of +1 and −10 units, a 90% win rate produces a mean of 0.9 × 1 − 0.1 × 10 = −0.1 unit. Many small winners can conceal an occasional large loss. Examine gain and loss size, expectancy, costs and drawdown alongside the percentage. A planned target does not establish that its assumed payoff or probability will actually occur.

What does it tell us about future wins?

Under a hypothetical independent-trial model with constant success probability, 120 wins in 200 trades estimate a 60% win rate. An approximate two-sided 95% Wilson interval is 53.1–66.5%. This is a statistical-method interval, not a guaranteed range for next month’s win rate. Dependence, selecting the best system and regime changes require a different or extended model. See NIST: confidence intervals for proportions.

The denominator changes other summaries too

The five-trade example totals USD 220 and averages USD 44. Excluding the zero result raises the mean to USD 55 despite unchanged cash P&L. That selection can answer a particular question, but both reports must disclose counts and rules. Do not combine win rate excluding zeros with full-sample expectancy without explaining which observations each measure describes and how they should be interpreted.

Practical reporting

Record the period, grouping rules, positive, negative and zero result counts, and uncertainty. Compare similar conditions and evaluate rules outside development. Do not omit difficult sessions or trades violating the plan without a separate complete report. A change in win rate does not independently identify why performance changed. In the course: R:R, win rate and expectancy. Source checked on 11 October 2026.

Diagram and worked example

Educational infographic: Win rate — Educational diagram; model data and assumptions are stated in the graphic.
Open the full-size graphic

FAQ

Is a 60% win rate good?

There is no universal threshold. Payoffs, costs, uncertainty and the strategy rules matter.

May zero results be excluded?

That convention can be used, provided it is disclosed and applied consistently.

← Back to the full glossary