You ran a backtest. The equity curve goes up. Profit factor is 3.2. Maximum drawdown is 8%. You’re ready to go live.
Here’s the problem: that backtest represents one possible ordering of trades. Change the sequence — put a cluster of losses at the start instead of the middle — and that 8% drawdown could become 25%. Your profit factor stays the same. Your win rate stays the same. But the path your equity takes is completely different, and that path determines whether you survive or get margin called.
This is exactly what Monte Carlo analysis is designed to reveal.
What Monte Carlo Simulation Actually Does
Monte Carlo simulation for trading takes your list of historical trades and reshuffles them thousands of times. Each shuffle creates an alternate equity curve — a parallel universe where the same trades happened in a different order. Across 1,000 or 5,000 of these reshuffles, you get a statistical distribution of every metric that matters: drawdown, final equity, consecutive losses, and peak-to-trough timing.
The mathematical foundation is straightforward. If your backtest has n trades, there are n! (n factorial) possible orderings. For 100 trades, that’s approximately 9.3 × 10157 possible sequences. Your backtest showed you exactly one. Monte Carlo sampling gives you a representative picture of the full distribution without needing to test every permutation.
There are two primary methods:
Permutation (shuffle without replacement): Every trade appears exactly once, just in a different order. The total profit is identical in every simulation — multiplication is commutative, so the product of all trade returns is the same regardless of sequence. What changes is the drawdown profile. This method isolates how sensitive your risk metrics are to trade ordering.
Bootstrap (resample with replacement): Trades are drawn randomly, allowing some to appear multiple times while others are skipped entirely. This produces variation in both the final equity and the drawdown. Bootstrap is more aggressive — it tests whether your overall result depends on a few outlier trades.
The Drawdown Illusion
Maximum drawdown is the most misleading metric in a backtest report. Not because the number is wrong, but because traders treat it as a ceiling. “My EA’s max drawdown is 12%, so I’ll set my risk to tolerate 15%.” This reasoning ignores a basic statistical reality: the backtest drawdown is a single sample from a distribution.
Run Monte Carlo on any EA with 100+ trades and you’ll typically find that the 95th percentile drawdown is 1.5x to 3x the original backtest drawdown. For some strategies — particularly those with clustered wins and losses — the ratio can exceed 4x.
What this means practically: if your backtest shows 10% drawdown and the Monte Carlo 95% confidence interval shows 22%, you should size your position as if 22% drawdown is realistic. Because statistically, there’s a 1-in-20 chance it will happen from trade ordering alone.
Confidence Intervals That Actually Matter
A proper Monte Carlo report gives you drawdown at multiple confidence levels:
50% CI (Median): Half of all simulations had drawdown below this value. Think of it as your “baseline expected” drawdown. If the median is significantly higher than your original backtest drawdown, the original sequence was likely lucky.
90% CI: Only 10% of simulations exceeded this drawdown. A reasonable stress test level — something that could happen in roughly 1 out of 10 equivalent trading periods.
95% CI: The industry-standard risk management threshold. Many professional traders use this for position sizing. If you can’t survive this drawdown, consider reducing your lot size.
99% CI: Near worst-case. Only 1% of simulations were worse. Use this when planning your absolute maximum capital reserve.
Our free Monte Carlo tool calculates all four confidence intervals automatically from any MetaTrader backtest report.
The Risk Multiplier
Here’s where Monte Carlo becomes directly actionable. Divide your original backtest drawdown by the 95% CI drawdown:
Risk Multiplier = Original DD ÷ 95% CI DD
If your backtest shows 8% drawdown and the 95% CI shows 16%, the multiplier is 0.50x. That means: to maintain the same drawdown level you saw in the backtest, you’d need to use roughly half your backtest lot size in live trading.
A multiplier above 1.0 means your original drawdown was actually worse than the 95% CI — the original trade order was unlucky on drawdown, and you may have room to scale up.
A multiplier below 0.5 should raise serious questions about whether the strategy’s risk is manageable at any meaningful size.
What Monte Carlo Cannot Tell You
Monte Carlo reshuffling tests one thing: sensitivity to trade ordering. It does not test:
Curve fitting: If your EA is overfitted to historical data, Monte Carlo won’t detect it. All reshuffled sequences use the same overfitted trades. You need out-of-sample testing and walk-forward analysis for this.
Regime changes: Monte Carlo assumes future trades will have similar characteristics to historical trades. If market conditions change fundamentally — volatility regime shift, structural break in correlations — the historical trade distribution may not apply.
Execution quality: Slippage, spread widening, and partial fills in live trading can change the profit distribution. Monte Carlo tests the theoretical trade list, not real-world execution.
Correlation between trades: Permutation testing assumes trades are independent. Some strategies have correlated trades — for example, grid strategies where multiple positions share the same directional bias. Shuffling these trades destroys the correlation structure and may understate risk.
Understanding these limitations makes Monte Carlo more useful, not less. It’s one essential validation step — not the only one.
Try It on Your EA
You can run a Monte Carlo analysis right now using our free Monte Carlo analyzer. Export your Strategy Tester report from MetaTrader 4 or MetaTrader 5 as an HTML file, upload it, and you’ll get a complete robustness analysis including drawdown confidence intervals, martingale detection, and a composite robustness score. Everything runs in your browser — your data never leaves your machine.
For comprehensive validation, Monte Carlo is one of 17 tests in Edge Matrix — our professional EA validation platform that also covers temporal stability, sample size validation, concentration risk, and edge decay detection.
The bottom line: a single backtest tells you what happened once. Monte Carlo tells you what could happen. The difference is the gap between hope and informed risk management.