Monte Carlo simulation applied to trading means replaying your trade history in thousands of random orders to estimate the probability of ruin of a strategy. Unlike a classic backtest that produces a single equity curve based on the real chronological order, Monte Carlo simulation explores hundreds of thousands of possible sequences and reveals the probability that your account drops below a threshold defined as unrecoverable. This is the difference between knowing your strategy "worked in the past" and knowing whether it can survive unfavorable future conditions.
What is risk of ruin in trading?
Risk of ruin represents the probability that your equity permanently falls below a threshold beyond which recovery becomes impossible. It is not the probability of losing on a trade: it is the probability that you can no longer continue trading.
Definition and basic calculation
Risk of ruin depends on three fundamental parameters: win rate (percentage of winning trades), average win/loss ratio (how much you win relative to what you lose, expressed as an R-multiple), and the percentage of capital risked per trade. These three elements form the mathematical foundation of any risk of ruin estimate.
A simplified analytical formula exists for series of independent and identically distributed trades. If your win rate is 50%, your average win/loss ratio is 1.5R, and you risk 2% per trade, your theoretical risk of ruin toward a minus-50% threshold is approximately 3 to 5%. That sounds reassuring on paper, but it assumes all your trades are independent, which is never exactly true in live trading.
The analytical formula systematically underestimates risk of ruin because it ignores the actual consecutive loss sequences in your backtest. This is why Monte Carlo simulation produces more conservative and more realistic estimates.
Why actuarial tables are not enough
Classic risk of ruin tables (Ralph Vince, Van Tharp) assume normally distributed returns. However, trading returns exhibit fat tails: extreme events occur far more frequently than what the normal distribution predicts. Monte Carlo simulation on your own trades bypasses this problem by using the actual distribution of your results.
Why it is the forgotten backtest metric
The vast majority of traders who backtest their strategies look at profit factor, win rate, and historical maximum drawdown. Almost nobody calculates risk of ruin. This omission carries real consequences: the historical maximum drawdown is only one scenario out of an infinite number of possible sequences.
According to the European Securities and Markets Authority (ESMA), between 74% and 89% of retail trading accounts lose money on leveraged products. This statistic is not the result of bad strategy selection alone: it is largely the result of risk sizing that does not account for the inherent statistical variability of any trade series.
Knowing your risk of ruin before deploying a strategy live is exactly what separates a professional approach from an amateur one. See our article on common backtesting mistakes to understand the other biases to avoid.
How Monte Carlo simulation works
Monte Carlo simulation on a trade series works by replaying individual trade results in thousands of different random orders. Each replay produces a different equity curve. After 10,000 simulations, you have a complete statistical distribution of possible outcomes.
Principle: replaying 10,000 trade sequences
Imagine your backtest contains 200 trades with their individual results expressed as R-multiples (1R = one unit of risk per trade). The simulation will:
Collect the R-multiple series
Randomly permute
Calculate the equity curve
Repeat 10,000 times
The standard recommendation is to run at least 1,000 simulations to obtain a stable distribution. Beyond 10,000, marginal precision gains are negligible for most trading strategies. Source: Wikipedia, Monte Carlo methods in finance.
Required inputs: the R-multiple series from your backtest
The primary input for the simulation is the series of R-multiples from each trade. An R-multiple expresses the result of a trade as a multiple of the initial risk: if you risk $100 on a trade and gain $150, that is +1.5R. If you lose your $100, that is -1R.
Expressing results in R-multiples rather than dollars or percentages has one major advantage: the simulation is independent of account size and leverage used. You can recalibrate the risk per trade directly within the simulation to compare 1%, 2%, or 3% risk scenarios without re-running the backtest.
You will find a detailed explanation of how to build the R-multiple series and its impact on expectancy in our article on backtest metrics: expectancy and profit factor.
Interpreting Monte Carlo results
Monte Carlo results are read primarily through percentiles. The 5th percentile represents the "95% confidence pessimistic scenario": in 95% of simulations, results were better than this scenario. The 95th percentile represents the optimistic scenario.
Equity curve at the 5th percentile
The 5th percentile equity curve is the most useful reference for estimating realistic maximum drawdown. If your classic backtest shows a maximum drawdown of 12%, it is not unusual for Monte Carlo simulation to reveal a 5th percentile drawdown of 22 to 28% on the same trade series.
This gap is not a sign that your strategy is bad. It is the natural variability inherent in any random series. The question to ask is: can my account and my psychology absorb a 25% drawdown without me cutting the strategy prematurely?
Realistic drawdown vs historical drawdown
The historical maximum drawdown of your backtest represents the worst scenario that actually occurred, in the real chronological order. Monte Carlo simulation reveals the worst possible scenario if losses had been concentrated at the beginning of the series. The 5th percentile drawdown is systematically 1.5x to 3x greater than the historical drawdown: use this figure as your "design drawdown" for capital sizing.
Ruin threshold: at what percentage to stop
The ruin threshold is the loss level beyond which you consider the account unable to recover. For individual traders, this threshold is typically set at 50% of initial capital (partial ruin) or 100% (total ruin). For prop firm traders, the threshold is usually the maximum drawdown allowed by the firm (8 to 10% depending on the rules).
The risk of ruin calculated by the simulation represents the percentage of simulations in which equity touched or exceeded this threshold at any point. A risk of ruin below 5% is generally considered acceptable for professional traders. Above 10%, the strategy presents a structural risk of ruin that must be corrected.
To understand how prop firm drawdown rules fit into this framework, see our article on backtesting prop firm rules.
Comparing scenarios before and after risk per trade adjustment
One of the most powerful uses of Monte Carlo simulation is comparing risk scenarios. By varying the percentage risked per trade (from 0.5% to 3%), you can directly visualize the impact on risk of ruin:
| Risk per trade | P5 max drawdown | Risk of ruin (50% threshold) | Median annualized return |
|---|---|---|---|
| 0.5% | 8% | < 1% | 12% |
| 1% | 15% | 2% | 24% |
| 2% | 28% | 8% | 45% |
| 3% | 40% | 18% | 62% |
This table illustrates the fundamental trade-off between return and risk of ruin: doubling risk per trade does not double ruin risk, it multiplies it by a much larger factor. This is why most professional traders stay between 0.5% and 2% risk per trade.
Monte Carlo with Backtrex
Backtrex integrates Monte Carlo simulation directly into the backtesting pipeline, eliminating the friction normally associated with this analysis: no need to export trades to an external tool, configure parameters in a spreadsheet, or write code. The simulation runs immediately on the results of your visual backtest.
Running a simulation on your backtest results
Once your strategy is built using drag-and-drop blocks and the backtest is executed, Backtrex displays Monte Carlo simulation results directly in the performance analysis panel. The process is as follows:
- Build your strategy with visual blocks from the features page.
- Run the backtest over the desired period (up to 10 years of data).
- In the "Monte Carlo Analysis" tab, select the number of simulations (1,000 to 10,000), the ruin threshold (50% by default), and the risk per trade.
- Backtrex displays the equity curve distribution, the calculated risk of ruin, and the 5th percentile drawdown in under 30 seconds.
The advantage of native integration is data consistency: Backtrex uses exactly the same trades and the same anti-repainting rules as your backtest. No distortion from export or data format conversion.
Backtrex anti-repainting rule
All signals generated by Backtrex use close[1] (the previous confirmed bar) rather than close[0] (the current bar). This rule ensures that the Monte Carlo simulation runs on realistic backtest results, without the repainting bias that artificially inflates performance. See the anti-repainting feature page for technical details.
Stress scenarios: what happens during a drawdown period?
An advanced feature of Monte Carlo simulation is stress testing on specific sub-periods. Rather than simulating across the entire historical series, it is possible to restrict the simulation to the bottom 20% of performing trades (the "left tail" of the distribution) to simulate a prolonged drawdown period.
This stress scenario reveals how long it takes, in the worst case, for the strategy to recover to prior levels. If the simulation indicates that a period of 30 consecutively underperforming trades is possible with 10% probability, you can plan for this scenario in your trading plan and decide in advance at which threshold you will suspend the strategy for re-evaluation.
To complete this robustness analysis, see our article on Monte Carlo simulation in trading which covers shuffling and bootstrap methods in detail, as well as our guide on detecting and preventing overfitting in backtests. For a complete introduction to backtesting, see how to backtest a trading strategy.
Important Risk Warning
FAQ
As a general rule, you need at least 100 trades for Monte Carlo simulation results to be statistically significant. Below 30 trades, results are too sensitive to a few extreme trades to be usable. The ideal is a series of 200 or more trades: beyond that, simulation precision increases slowly and conclusions become very stable. The longer the series, the more the risk of ruin distribution converges toward its theoretical value.
Risk of ruin is the probability that an account balance reaches a loss level defined as unrecoverable, typically 50% or 100% of the initial capital. This risk depends on win rate, average win/loss ratio, and the percentage risked per trade. A risk of ruin below 5% is generally considered acceptable for a professional trader seeking to operate long term.
No. Monte Carlo simulation does not predict future losses: it models the distribution of possible scenarios by replaying historical trades in random orders. It is probabilistic, not predictive. It assumes that the future statistical characteristics of your strategy (win rate, win/loss ratio) will be similar to those of the backtest, which is never guaranteed. Its role is to quantify the risk inherent in your current trade series, not to predict the market.
No, the two methods are complementary. Monte Carlo simulation analyzes the statistical variability of a given trade series: it answers the question "what if my trades had occurred in a different order?". Out-of-sample testing validates the strategy's ability to generalize to data not seen during optimization. A robust strategy passes both tests: low risk of ruin in Monte Carlo and consistent performance out-of-sample.
For prop firm traders (FTMO, MFF, Topstep), the relevant ruin threshold is the maximum drawdown allowed by the firm, typically between 8% and 12% of the evaluated capital. A risk of ruin below 5% relative to this threshold is recommended. If Monte Carlo simulation reveals a 15% probability of hitting the prop firm's maximum drawdown, risk per trade must be reduced before attempting the challenge phase.
The shuffle method (sampling without replacement) randomly reorders your historical trades while keeping exactly the same results. The bootstrap method (sampling with replacement) randomly draws trades from your pool allowing repetitions, which can generate loss sequences longer than those observed historically. Shuffle gives a conservative estimate; bootstrap is more pessimistic and more revealing of tail scenarios. Backtrex offers both methods in its Monte Carlo module.
The most directly actionable variable is the percentage risked per trade. Halving the risk per trade typically reduces risk of ruin by a factor of 3 to 10, depending on the result distribution. Increasing the number of trades (by expanding the universe or frequency) also improves the statistical properties of the series. Finally, defining a strategy stop rule (suspend trading if drawdown exceeds X%) mechanically limits live risk of ruin. See the pricing page to explore how Backtrex helps you run these scenarios instantly.