Every backtest tells you exactly one story: what would have happened if your strategy had traded a single specific historical path of prices. That single path is not the future — it is one realisation of a stochastic process that could have unfolded in thousands of different ways. Monte Carlo simulation is how we explore those other paths and understand what our strategy's performance distribution actually looks like.
At StockHunt, Monte Carlo simulation is embedded into every stage of our strategy development and live risk management process. It is not an academic exercise — it is a practical tool that has directly shaped our position sizing, drawdown limits, and capital allocation decisions.
What Is Monte Carlo Simulation?
Monte Carlo simulation is a computational technique that generates a large number of random scenarios based on statistical properties derived from observed data. In a trading context, the most common application is equity curve simulation: given a strategy's distribution of trade returns (mean, variance, skewness), generate thousands of random sequences of those returns and observe the resulting distribution of outcomes — equity curves, drawdowns, and final returns.
The power of this approach lies in what it reveals about tail risk. Your actual backtest shows you the median outcome — one specific sequence of wins and losses. Monte Carlo shows you the 5th percentile outcome (the bad luck path), the 95th percentile (the lucky path), and everything in between. As a risk manager, the 5th percentile is what you plan around.
Three Core Applications in Trading
1. Maximum Drawdown Distribution
A backtest might show a maximum historical drawdown of 15%. But that is the drawdown of one specific sequence of trades. Monte Carlo across 10,000 simulations might reveal that 10% of sequences produce a drawdown exceeding 30%, and 1% exceed 45%. This tells you something the backtest alone never could: even with this strategy's edge intact, there is a 10% chance of experiencing a drawdown twice as severe as anything in your historical record.
We use this distribution to set our live drawdown hard stop — the level at which we halt trading and reassess. Rather than basing it on historical maximum drawdown, we base it on the 95th percentile Monte Carlo drawdown. This ensures we are capitalised to survive even the unlucky sequences that did not happen to appear in our historical data.
2. Probability of Ruin
Probability of ruin is one of the most important risk metrics that most retail traders never calculate. It answers a simple question: given my strategy's edge, my position sizing, and my starting capital, what is the probability that I will lose a defined percentage of my account before hitting a defined profit target?
Monte Carlo makes this calculation straightforward. Run 10,000 simulations. Count the fraction that hit your ruin threshold before your profit target. That fraction is your probability of ruin at your current position sizing. If it is above 5%, your position sizing is too aggressive — full stop. This single calculation, properly applied, would prevent the majority of catastrophic retail trading losses.
3. Position Sizing Optimisation
Monte Carlo simulation is the correct tool for position sizing decisions because it accounts for the sequence-dependency that fixed-formula approaches (like Kelly Criterion) miss. A streak of losses early in a trading period has a different impact than the same streak late — Monte Carlo captures this path-dependency naturally.
Our position sizing process works as follows: for a given target maximum drawdown and confidence level, we run Monte Carlo simulations at varying position sizes and find the largest size that keeps the simulated drawdown below our threshold at the 95th percentile of outcomes. This is our maximum allowable position size — and it is always smaller than Kelly would suggest.
How to Run a Monte Carlo Simulation for Your Strategy
- 1Collect your complete trade log from backtesting — every individual trade's return expressed as a percentage of capital risked.
- 2Calculate the statistical properties of this return distribution: mean, standard deviation, and if desired, skewness and kurtosis.
- 3Generate N random sequences of trades by sampling with replacement from your actual trade returns (bootstrap method) or from a fitted distribution.
- 4For each sequence, calculate the cumulative equity curve and record the maximum drawdown and final return.
- 5Analyse the resulting distribution: what are the 5th, 25th, 50th, 75th, and 95th percentile drawdowns and returns? These percentiles define your realistic performance range.
- 6Repeat with different position sizes until you identify the sizing that meets your risk criteria at the 95th percentile confidence level.
The Bootstrap vs. Parametric Approach
There are two main methods for generating Monte Carlo scenarios. The parametric approach fits a probability distribution (typically normal or log-normal) to your trade returns and samples from that fitted distribution. The bootstrap approach directly resamples with replacement from your actual observed trade returns.
We strongly prefer the bootstrap approach for crypto trading strategies for one critical reason: crypto trade returns are not normally distributed. They exhibit heavy tails — large losses and large gains occur far more frequently than a normal distribution would predict. The bootstrap method naturally preserves these fat tails because it samples directly from observed data rather than fitting a distribution that may not capture them adequately.
“Monte Carlo does not predict the future — it maps the full landscape of possible futures that your strategy might inhabit. Knowing the terrain before you enter it is the entire point.”
Monte Carlo in Live Risk Monitoring
Beyond strategy development, we use Monte Carlo as a live risk monitoring tool. Every week, we update our simulation using the most recent 90 days of live trading returns, recomputing the current probability of ruin and expected drawdown distribution. If the updated simulation shows deteriorating risk metrics — increasing drawdown percentiles or rising ruin probability — it is an early warning signal that the strategy's edge may be eroding, triggering a deeper review before we hit our hard drawdown stop.
Run Monte Carlo before you go live, not after you blow up. Understanding your strategy's worst-case distribution is what separates disciplined risk management from hopeful gambling.
