A 55 percent win rate with +3 percent wins and -2 percent losses: the median deepest drawdown over 1,000 trades is 19.1 percent, one path in ten falls past 24.9 percent
Our own simulation, 7 November 2026 (UTC), for the article "How deep can drawdowns get for a strategy with a 55% win rate?". A model, not a track record: every trade is independent, wins +3 percent of equity and losses -2 percent, compounded. Way 1: 10,000 paths of 1,000 trades per win rate, fixed seed 20261107. Way 2: the probability of a losing run of k or more in 1,000 trades from the run count's Poisson approximation, 1 - exp(-1000 x 0.45^k x 0.55); the two agree within 3 points.
What we found
- At a 55 percent win rate the median path's deepest drawdown is 19.1 percent; one path in ten falls past 24.9 percent, one in a hundred past 32.2 percent, and 2.0 percent of paths fall 30 percent or more from a peak.
- A run of at least 8 straight losses occurs on 61.0 percent of 1,000-trade paths (analytic 60.3 percent); a run of 10 or more on 16.7 percent (analytic 17.1 percent).
- Eighteen straight 2 percent losses cut equity by 30.5 percent (0.98 to the power 18 = 0.695); a run that long occurs on about 3.1 paths in 10,000.
- At a 45 percent win rate the median deepest drawdown is 34.9 percent and 7.3 percent of paths lose half their equity at some point; at 65 percent the median is 13.2 percent.
Check it yourself
| Win rate | Expected gain per trade | Median deepest drawdown | One path in ten deeper than | One in a hundred deeper than | Paths with a 30 percent drawdown or more | Paths halved at some point |
|---|---|---|---|---|---|---|
| 45% | +0.25% | 34.9% | 47.9% | 61.6% | 74.2% | 7.34% |
| 50% | +0.50% | 24.6% | 32.8% | 42.1% | 18.4% | 0.07% |
| 55% | +0.75% | 19.1% | 24.9% | 32.2% | 2.0% | 0.00% |
| 60% | +1.00% | 15.3% | 20.2% | 25.9% | 0.2% | 0.00% |
| 65% | +1.25% | 13.2% | 16.6% | 20.9% | 0.0% | 0.00% |
Streaks at a 45 percent loss rate over 1,000 trades: at least 8 losses in a row, simulated 61.0 percent against analytic 60.3 percent; at least 10, 16.7 against 17.1.
What this cannot tell you: real trades are not independent and their sizes vary; a strategy's true win rate is unknown and drifts; the seed changes every figure by a few tenths of a point.
The same figures rounded as the article states them
- At a 55% win rate the median deepest drawdown is about 19%, one path in ten falls more than 25% from a peak, one in a hundred more than 32%, and about 2% of 1,000-trade paths fall 30% or more.
- Win rate 45%: median 35%, one in ten deeper than 48%, 74% of paths with a 30% drawdown or more, about 7% halved at some point. Win rate 50%: 25%, 33%, 18%. Win rate 60%: 15%, 20%, 0.2%. Win rate 65%: 13%, 17%, about 1 in 10,000 with a 30% drawdown.
- Streaks: a run of at least 8 losses on about 61% of paths, of 10 or more on about 17%; 18 straight losses cut equity by 30%, on about 3 to 4 paths in 10,000.