Monte Carlo simulation for trading strategies is a stress-testing technique that randomly reshuffles a trader’s historical trade results thousands of times to model the full distribution of possible future equity curves — not just the average outcome. Where a simple backtest replays trades in historical order, Monte Carlo reveals what happens when a losing streak clusters at the start of a run, exposing the true range of risk embedded in any strategy.
Key Takeaways
- A positive-expectancy strategy can still produce a 30%+ drawdown if losses cluster early — Monte Carlo quantifies the exact probability of this happening with your trade data.
- The 5th-percentile equity curve is the critical output: if it breaches your maximum drawdown limit, reduce per-trade risk until the simulation passes.
- Monte Carlo requires at least 50 trades to produce useful results; 100+ trades yield meaningful confidence intervals.
How Monte Carlo Simulation Works
Monte Carlo applies bootstrapped resampling to a trade log. Each simulation iteration draws trades randomly (with replacement) from the historical sample and recalculates the equity curve in that new sequence. After 5,000–10,000 iterations, the result is a fan of equity curves spanning best-case to worst-case outcomes.
The two core methods:
- Bootstrapped Monte Carlo: Resamples actual trade P&L dollar amounts or R-multiples. Preserves fat tails and asymmetric distributions. More accurate for discretionary and trend-following strategies.
- Parametric Monte Carlo: Generates synthetic trades from a model (mean return, standard deviation). Faster, but assumes a normal distribution that understates the frequency of large losses.
For most retail traders, bootstrapped Monte Carlo run on their actual journal data produces the most actionable results.
Key outputs to read from any simulation:
5th percentile curve → worst realistic equity path
50th percentile curve → median expected outcome
95th percentile curve → best realistic equity path
P(drawdown above 20%) → % of simulations that hit a 20% drawdown
P(drawdown above 25%) → % of simulations that hit a 25% drawdown
Quick Reference
| Aspect | Detail |
|---|---|
| Method | Random resampling (bootstrapping) of historical trade P&L |
| Standard Iterations | 10,000 for stable output |
| Key Output | 5th/95th percentile equity curves + probability of ruin |
| Minimum Sample | 50 trades (100+ recommended) |
| Good Signal | 5th-percentile max drawdown stays within your risk limit |
| Warning Sign | More than 10% of simulations breach your drawdown threshold |
Practical Example
A futures trader has 120 completed ES (S&P 500 futures) trades logged: 54% win rate, average winner +$412, average loser -$285. Expectancy: +$93 per trade. On paper, the strategy looks solid.
They run 5,000 Monte Carlo simulations by randomly resampling those 120 trades. Results:
- Median ending balance after 120 forward trades: +$11,160 — matches expectation.
- 5th-percentile maximum drawdown: -$8,200 (32% on a $25,000 account) at some point during the run.
- Probability of breaching 20% drawdown: 18% of simulations at current 1-contract sizing.
The trader’s risk rule is “never exceed 20% drawdown.” At 18% probability, current sizing fails that rule. The fix: reduce per-trade exposure by 30% (equivalent to 0.7 contracts). Re-running the simulation with reduced size shows only 6% of paths breach the 20% threshold — within an acceptable range.
This is the direct output no win-rate calculation or backtesting replay can provide: a concrete position size derived from the trader’s own data, calibrated to a personal drawdown limit.
Monte Carlo simulation reshuffles your historical trades thousands of times to show the full range of outcomes your strategy can produce. It reveals worst-case drawdowns and helps you size positions so losses stay within your risk tolerance.
Common Mistakes
- Running simulation on too few trades. Below 50 trades, the resampled distribution is too narrow to capture real variance. A 30-trade sample will systematically underestimate tail risk.
- Ignoring the 5th-percentile curve. Traders focus on the median outcome and miss the worst realistic scenario. The 5th percentile is where position sizing decisions should be anchored.
- Treating simulation output as a forecast. Monte Carlo shows the distribution of outcomes given past behavior, not a prediction. If market regime shifts, the historical sample is no longer representative.
- Skipping re-simulation after changing strategy rules. Any change to entry criteria, stop placement, or target size invalidates the prior simulation. Re-run whenever the underlying trade parameters change meaningfully.
How JournalPlus Tracks This
JournalPlus logs every trade’s P&L, R-multiple, and outcome, giving traders the clean historical sample that Monte Carlo simulation requires. The trade export includes the data structure — win rate, average winner, average loser — needed to feed directly into a bootstrapped simulation. Traders can export their full trade log from JournalPlus and run simulation analysis against the actual numbers from their journal rather than hypothetical parameters.
