Monte Carlo Backtesting: Why One Backtest Is Never Enough

The Problem With Traditional Backtesting

A traditional backtest runs your strategy against historical data and produces a single equity curve. That curve represents one specific sequence of wins and losses out of potentially millions of possible orderings. As David Aronson wrote in *Evidence-Based Technical Analysis* (Wiley, 2006), "a single backtest is a sample of one, and it tells you what happened, not what could happen."

Shuffle the order of those same trades and you get a completely different equity curve, with different drawdowns and a different final balance.

What Monte Carlo Simulation Does Differently

Monte Carlo simulation shows a range of outcomes for the plan you enter, instead of one equity curve. The name comes from work at Los Alamos in the 1940s. The useful part for a trader is the result: a weaker case, a typical case, and a stronger case. Not a recipe, and not financial advice.

This shows you:

• P5 scenario: What happens in the worst 5% of cases
• P50 scenario: The median, most likely outcome
• P95 scenario: What happens when things go well

How to Interpret Results

If your P5 shows a 60% drawdown but your P50 shows a 30% gain, the strategy is viable but volatile. You need to decide if you can survive the worst case to capture the median.

If even your P50 shows negative returns, the strategy likely doesn't have a real edge.

Why Order Matters

The exact same set of trades can produce:

• Path A: Early wins, smooth growth, 45% return
• Path B: Early losses, 35% drawdown, then recovery to 15%
• Path C: Clustered losses, account ruin at trade 67

All three paths contain the same trades. Only the sequence changed. According to Nassim Taleb in *Fooled by Randomness* (Random House, 2001), "the difference between a profitable trader and a ruined one is often nothing more than the order in which the same trades occurred."

Try it with your own plan in the Risk of Ruin Calculator. Or use RangeSight for a Monte Carlo view of the range of outcomes. Not financial advice.

Conceptually this provides a framework. In practice, outcomes vary across simulations. Testing multiple scenarios reveals how consistent a strategy actually is.