Every trader accepts, in theory, that losing streaks happen. Almost no trader has done the arithmetic on how long a normal streak actually is. The gap between those two numbers is where accounts — and confidence — quietly die.
The question we simulated: over 500 trades, what is the longest run of consecutive losses a trader should expect, purely from chance, with the edge working exactly as designed? 20,000 simulated sequences per win rate.
Independent trades at each fixed win rate. 95th-percentile streaks run ~40–50% longer than the medians shown.
Read your row and sit with it
| Win rate | Median longest streak (500 trades) | 95th percentile |
|---|---|---|
| 30% | 15 losses in a row | 22 |
| 40% | 11 | 16 |
| 50% | 8 | 12 |
| 60% | 6 | 9 |
| 70% | 5 | 7 |
A trend-following style method winning 40% of the time should expect eleven consecutive losses somewhere in its next 500 trades — and one such method in twenty will see sixteen in a row. Nothing is broken. The edge is intact. The coin simply does what coins do.
Why this number changes behaviour
First, it sets risk size from the other direction. Eleven straight losses at 1% risk is a manageable ~10% drawdown; at 3% risk it's a 28% crater in the account and a larger one in the psyche — mid-way through which you will be utterly convinced the method has died. Multiply your expected streak by your risk per trade before choosing that risk; the result is a drawdown you must be able to call "normal operation."
Second, it defines when a method is actually broken versus merely breathing. If your win rate is 45%, a streak of 9 is weather, not climate — abandoning the method there is how traders end up owning ten half-tested systems and zero track records. The rational tripwire lives out past the 95th percentile, not at the point where losing starts to hurt. (Your journal, not your memory, is where the real win rate that feeds this table comes from.)
What the lab says
- Long losing streaks are a certainty, not a symptom. Budget for the 95th-percentile streak at your win rate — it is coming in some 500-trade window.
- Size so the expected streak is boring. Expected streak × risk per trade ≈ a drawdown you must survive with judgment intact.
- Don't diagnose a method inside its normal noise. Quitting during a statistically ordinary streak destroys more edges than markets ever do.
- Low win rate ≠ low quality. A 35%-win method with big winners lives with 12-loss streaks as a design feature. Know your design.