StructureX

Home › Tools

Win-rate range calculator: how much can a small sample tell you?

A win rate from 20 trades looks precise. It isn't. Enter the counts from your journal and see how uncertain the estimate still is.

Free tool · observed win rate · 95% Wilson interval · no account connection

Count every trade made under the same written rules. Do not leave out the losses.

Read the range, not just the win rate. Fifty wins in a hundred trades gives an observed 50% win rate. It does not tell you that the next hundred will also be 50%. A short test leaves room for a wide range of possible win rates.

The range uses the 95% Wilson score interval for a binomial proportion (z = 1.96). It is a statistical estimate, not a promise. It assumes the trades are reasonably independent and the method being tested has not changed. Market conditions and changing rules can break those assumptions.

How to use this with a journal

  1. Write one strategy's entry, exit and skip rules before testing.
  2. Count every completed test trade and every win. Keep the exact same rules.
  3. Enter the two counts. Read both ends of the interval.
  4. Check the average size of wins and losses with the expectancy calculator. Win rate by itself cannot show profitability.

When the number misleads

Ten wins in a row feels convincing, but they can all come from one strong trend. A backtest with peeking, changed rules or only favorite charts is not fixed by a confidence formula. Test trending and sideways periods, record spreads and costs, and compare the same setup across the sample. See how many trades to test and how to backtest without peeking.

Make this a habit, not a one-off. Lesson 2 of our free email course is all risk: the 1% rule, where your stop goes, and the 2-loss rule. Send me the free lessons →

Follow along

XTradingViewFacebook