Take one strategy and change exactly one number — the percent of the account risked per trade — and you get completely different trading careers. Same entries, same exits, same edge. Different lives.
We simulated it. The system: 45% win rate, winners worth 2R, losers worth −1R — a perfectly ordinary, positive-expectancy method (+0.35R per trade). We ran 4,000 simulated accounts of 400 trades each at four risk settings and looked at what actually happens to the equity curve, the drawdown, and the trader holding it.
System: 45% win rate, 2R winners, −1R losers, no trading costs, independent trades, $10,000 start, compounding. Log scale — equal vertical steps are equal multiples.
The part everyone sees: growth
Compounding rewards risk brutally well — while the edge holds. Median outcomes after 400 trades on a $10,000 start: 0.5% risk grows to about $19 903; 1% to $38 706; 2% to $136 690; and 5%, absurdly, to several million. Looking at that column alone, every trader would pick maximum risk. That column is the bait.
The part that decides whether you survive: drawdown
| Risk / trade | Median end (400 trades) | 10th percentile | Median max drawdown | Worst-decile drawdown |
|---|---|---|---|---|
| 0.5% | $19 903 | $16 385 | 5% | 8% |
| 1% | $38 706 | $26 256 | 11% | 16% |
| 2% | $136 690 | $63 131 | 21% | 29% |
| 5% | ~$3.5M | ~$527k | 46% | 61% |
Read the right side of that table slowly. At 5% risk, the median journey includes losing nearly half the account at some point — and one run in ten loses 60%+. That is the journey with a working, profitable edge and a machine executing it without emotion. No human executes through a 60% drawdown without changing something — cutting size at the bottom, skipping the setups that would have recovered it, or quitting. The simulation compounds; the human breaks.
The hidden assumption: that your edge is real
Everything above assumed the edge holds forever. Now the honest scenario: you believed you had 45%/2R, but live execution gives you 30%/2R — a mildly negative edge, which is exactly what most developing traders actually run. Same 400 trades: at 0.5% risk the account drifts to about $8,100 — annoying, survivable, and cheap tuition. At 2% it falls to roughly $3,900, and 78% of runs are more than half destroyed at some point — 69% finish there. At 5%: median $537 left, with 98% touching a halved account and 95% ending below it. Risk size is the price of discovering the truth about your edge. Small risk means you can afford the lesson; big risk means the lesson costs everything.
What the lab says
- Risk per trade is a volatility dial, not a profit dial. It scales the drawdowns as surely as the gains — and drawdowns, not averages, are what force human errors.
- Pick size from the drawdown you can genuinely sit through. If a 20% drawdown would change your behaviour, 2% risk is already your ceiling with a good edge.
- While your edge is unproven, trade at tuition size. 0.5–1% keeps the cost of being wrong survivable. Earn the right to size up with a track record, not confidence.
- The equity curve you backtest is not the one you'll live. Median curves are smooth; your single life is one noisy draw from the whole distribution — including the ugly percentiles.