30 wins from 50 trades: what's the real win rate?
Posted: Fri Oct 30, 2026 10:38 am
Fifty trades logged, 30 wins. That is 60%. It looks like a solid system.
Run the standard error. With p = 0.6, the standard error is the square root of (0.6 x 0.4 / 50). That is the square root of 0.0048, about 0.069. A rough 95% interval is the estimate plus or minus 1.96 standard errors, or 0.6 +/- 0.136. That gives roughly 46% to 74%.
Sit with that range. The true win rate could be 46%, which at 1:1 loses money, or 74%, which is spectacular. Fifty trades cannot tell those apart. Fancier methods like the Wilson interval shift the edges a little, but the width stays about the same.
For contrast, the same 60% over 500 trades gives a standard error of about 0.022 and an interval of roughly 57% to 64%. Ten times the trades, but only about a third of the width. The width shrinks with the square root of the sample, so halving it takes four times the trades.
Another way to see it: 30 or more wins out of 50 happens about 10% of the time for a pure coin flip. So a 50 trade record at 60% is not even strong evidence against a plain 50% game.
All of this assumes the trades are independent and taken in similar conditions, which rarely holds for a setup traded across different sessions.
How wide is the interval on your own last 50 trades, and did it change how you see the setup?
Run the standard error. With p = 0.6, the standard error is the square root of (0.6 x 0.4 / 50). That is the square root of 0.0048, about 0.069. A rough 95% interval is the estimate plus or minus 1.96 standard errors, or 0.6 +/- 0.136. That gives roughly 46% to 74%.
Sit with that range. The true win rate could be 46%, which at 1:1 loses money, or 74%, which is spectacular. Fifty trades cannot tell those apart. Fancier methods like the Wilson interval shift the edges a little, but the width stays about the same.
For contrast, the same 60% over 500 trades gives a standard error of about 0.022 and an interval of roughly 57% to 64%. Ten times the trades, but only about a third of the width. The width shrinks with the square root of the sample, so halving it takes four times the trades.
Another way to see it: 30 or more wins out of 50 happens about 10% of the time for a pure coin flip. So a 50 trade record at 60% is not even strong evidence against a plain 50% game.
All of this assumes the trades are independent and taken in similar conditions, which rarely holds for a setup traded across different sessions.
How wide is the interval on your own last 50 trades, and did it change how you see the setup?