Understanding Monte Carlo Simulation

Monte Carlo simulation is a statistical technique that Gilito uses to stress-test strategies beyond their historical backtest results. Instead of asking "how did this strategy perform in the past," Monte Carlo asks "how might this strategy perform across thousands of possible future scenarios?"

What Monte Carlo Simulation Tells You

A single backtest shows one path — what happened historically. But the future will not repeat the past in the same order. Monte Carlo simulation takes the strategy's historical trade results and reshuffles them thousands of times to generate a range of possible outcomes. This reveals:

• How robust the strategy is — does it profit in most scenarios, or only in the specific historical sequence?
• The range of potential drawdowns — what is the worst case you might realistically face?
• The probability distribution of returns — what outcomes are most likely vs. edge cases?

Reading the Results

When you run a Monte Carlo simulation on a strategy result, Gilito displays several key metrics:

Probability of Profit

The percentage of simulated scenarios where the strategy ended with a positive return. For example, a 85% probability of profit means that in 85 out of 100 simulations, the strategy was profitable. Higher is better, and strategies with above 70% are generally considered robust.

Drawdown Distribution

A histogram showing the range of maximum drawdowns across all simulations. This tells you how deep the worst losing streak could realistically get. Key percentiles are highlighted:

• Median drawdown (50th percentile) — the most typical maximum drawdown you might experience
• 95th percentile drawdown — a worst-case scenario that would only occur 5% of the time
• 99th percentile drawdown — an extreme scenario for risk-sensitive evaluation

Confidence Intervals

The simulation produces confidence bands around the strategy's equity curve, showing the range of outcomes at different probability levels:

• 50% confidence band — the range where half of all simulations fall (the most likely outcomes)
• 80% confidence band — a wider range covering most realistic scenarios
• 95% confidence band — the near-complete range, excluding only extreme outliers

Narrow confidence bands indicate a consistent strategy. Wide bands suggest higher variance and less predictable outcomes.

Using Monte Carlo to Validate Strategies

Monte Carlo simulation is most useful as a validation step after reviewing a strategy's backtest results. Here is how to incorporate it into your workflow:

Step 1: Run a Backtest

Start with a standard backtest to see the strategy's historical performance. Look at the Gilito Score, total return, and maximum drawdown.

Step 2: Run Monte Carlo

From the strategy result page, click Monte Carlo Simulation to generate the analysis. The simulation runs thousands of scenarios and displays results within a few seconds.

Step 3: Evaluate Robustness

Compare the Monte Carlo results against your risk tolerance:

• Is the probability of profit above your minimum threshold (e.g., 70%)?
• Can you tolerate the 95th percentile drawdown without panic selling?
• Are the confidence bands narrow enough to trust the return projections?

Step 4: Compare Strategies

Run Monte Carlo on multiple strategies for the same asset and compare their probability of profit and drawdown profiles. A strategy with a slightly lower historical return but higher probability of profit and lower drawdown risk may be the better choice for real trading.