Losing streaks in prop challenges: how many to plan for
Educational only, not financial advice. Simulations and backtests are models, not promises. Check every rule with your own firm.
Most challenges aren’t lost to one terrible trade. They’re lost to an ordinary losing streak that the position size couldn’t absorb. The good news: how long streaks get is pure probability, so you can plan for it.
Streaks are longer than they feel
People expect wins and losses to alternate more than they really do. Even a strategy that wins half the time will produce runs of losses that feel like proof it’s broken. The table below shows the exact chance of seeing at least one run of losses of a given length somewhere in a series of trades, assuming each trade is independent.
| Win rate | Trades | 5+ losses in a row | 8+ in a row | 10+ in a row |
|---|---|---|---|---|
| 40% | 100 | 98% | 49% | 20% |
| 40% | 200 | 100% | 75% | 38% |
| 45% | 100 | 92% | 31% | 10% |
| 45% | 200 | 99% | 53% | 20% |
| 50% | 100 | 81% | 17% | 4% |
| 50% | 200 | 97% | 32% | 9% |
| 60% | 100 | 46% | 4% | 1% |
| 60% | 200 | 71% | 7% | 1% |
Computed exactly with the standard run-length recursion (see Schilling, 1990), rounded to whole percentages. Our own calculation, not a survey of traders.
Two things stand out. First, a 5-trade losing streak is close to certain for most strategies over a challenge-length sample. Second, the lower your win rate, the longer the streaks: a trend-following style that wins 40% of the time should expect an 8-loss run about as often as not within 100 trades. That’s not a sign the strategy stopped working. It’s what a 40% win rate looks like.
Turn the streak into a risk limit
Once you pick a streak you want to survive, the maximum risk follows directly:
max risk per trade = (max drawdown − safety buffer) ÷ streak length
With a 10% max drawdown, a 1% buffer you never plan to use, and a 10-loss streak, that’s 0.9% per trade. Remember that touching the limit usually counts as a breach: at exactly 1% risk, a 10% drawdown survives 9 losses, not 10.
Then do the same check against the daily loss limit. A streak doesn’t politely spread itself across days. If your daily limit is 5% and you risk 1%, four losses in one day leave you one loss from failing. Many traders set a personal rule to stop for the day after two or three losses, well inside the firm’s limit.
The risk-per-trade planner does both calculations and shows them side by side for several risk levels.
Trailing drawdowns make streaks more dangerous
With a static drawdown, profits build a cushion: if you’re up 4% and the floor is 10% below your start, you have 14% of room. With a trailing drawdown, the floor follows your high, so a streak that starts right after a good run has only the original room to work with. If you trade under a trailing rule, plan your streak against the trailing distance, not against your profit. The drawdown visualizer shows the difference on your own trades.
Why smaller isn’t automatically better
Cutting risk lets you survive longer streaks, but it also means more trades before you reach the target. If your strategy has a real edge after costs, more trades is fine: the edge has more chances to show up. If it doesn’t, more trades just means paying the spread more often. That’s the trade-off covered in position sizing for prop challenges and in the real odds of passing with no edge.
Caveats
- The table assumes each trade is independent. Real results can cluster, for example when market conditions stop suiting your strategy for weeks. That makes long streaks more likely than the table says, not less.
- A "loss" here means a full 1R loss. Partial losses, slippage and gaps change the arithmetic, usually against you.
- Your measured win rate is an estimate. If it comes from a short or optimised backtest, assume it’s lower and plan for longer streaks.
Sources
- Schilling, M. F. (1990). The longest run of heads. The College Mathematics Journal, 21(3), 196–207. doi:10.1080/07468342.1990.11973306 [Peer-reviewed]
- Feller, W. (1968). An Introduction to Probability Theory and Its Applications, Vol. 1, 3rd ed. Wiley (chapter XIII, recurrent events and runs). [Book]
"Peer-reviewed" means published in a refereed journal. Working papers and books aren’t peer-reviewed. Figures from "our own tests" are our backtests and simulations, not peer-reviewed research.