Monte Carlo simulation generates thousands of random scenarios by sampling from assumed probability distributions. For trade planning, this means instead of predicting a single outcome, you generate a distribution of possible outcomes and evaluate your strategy against the full range.
The basic setup for a trading Monte Carlo involves specifying the expected return and volatility of your strategy, the number of trades or the time horizon, and any position sizing rules. The simulation then runs thousands of paths, each with randomly generated returns drawn from your specified distribution. The result is a fan of possible equity curves showing the range of outcomes you might experience.
What makes this powerful is that it reveals probabilities that intuition cannot calculate. Questions like what is the probability that I experience a 30% drawdown within the first year or what is the median time to double my account become answerable with reasonable precision. These are practical planning questions that static expected value calculations cannot address.
For position sizing, Monte Carlo is particularly valuable. You can simulate the same strategy with different position sizes and observe how the distribution of outcomes changes. Larger positions increase both the upside (higher median terminal wealth) and the downside (higher probability of ruin). The simulation lets you find the position size that balances growth against survivability based on your specific risk tolerance.
The Kelly Criterion gives you the theoretically optimal bet size, but it assumes you know the exact probability and payoff of each trade. In practice, those inputs are estimated with uncertainty. Monte Carlo allows you to simulate around your uncertain estimates and see how sensitive the optimal position size is to estimation errors. Often, betting a fraction of the Kelly amount (half-Kelly or quarter-Kelly) produces nearly as much growth with dramatically lower drawdown risk.
One important consideration is the distribution assumption. If you simulate with normally distributed returns but your actual strategy produces fat-tailed returns, the simulation will underestimate extreme outcomes. Using historical return distributions or fitting a Student-t distribution can produce more realistic tail behavior.
Sequence risk is another insight that Monte Carlo reveals. Two traders with identical average returns can have very different outcomes depending on the order in which those returns arrive. Early drawdowns are more damaging than late drawdowns because they reduce the base from which compounding operates. Monte Carlo shows you how much sequence risk affects your specific strategy.
The practical workflow is to run the simulation before sizing a new position or launching a new strategy. Define what constitutes an unacceptable outcome (a specific drawdown level or probability of total loss), and check whether your planned position size keeps that probability below your threshold. If it does not, reduce the size until it does. This is not conservative. This is how you stay in the game long enough for your edge to compound.