Monte Carlo Simulation

Monte Carlo simulation tests your strategy's robustness by running thousands of randomized variations. If a strategy only works with trades in the exact order they occurred, it's fragile. Monte Carlo reveals how it performs under different conditions.

This is one of the most common reasons strategies fail in live trading — see Why Strategies Fail for the full picture.

Where to find it: In the app sidebar, open Monte Carlo. It has four tools, each covered below:

A fifth tool, Crisis Dependency, also lives under the Benchmark section and is covered on its own page.

On the Robustness Test and Health Validation pages you pick a strategy, set Starting Capital and Number of Simulations/Method in the Simulation Settings card, and click Run Simulation (or Run Validation). Progress is shown per method as the engine works through Shuffle, Bootstrap, Block Bootstrap, Parametric, and Stress+.

What Makes AlgoChef's Monte Carlo Different

Most Monte Carlo tools ask one question: "What is my net profit at the 95th percentile?"

AlgoChef asks a fundamentally different question: "Does my strategy's composite quality score stay stable across thousands of randomized scenarios?"

Here's why that distinction matters. Net profit at a single confidence level tells you one number about one outcome. It says nothing about whether your strategy's risk profile, statistical reliability, or overall quality holds up under randomization — or collapses.

AlgoChef's Monte Carlo recalculates the full CSI (Casey Strategy Index), along with the separate Profitability, Risk and Confidence scores, for every single simulation run, across all five methods. Each run produces a complete quality assessment, not just a P&L figure.

The result: AlgoChef measures the variability and stability of your strategy's composite score across thousands of scenarios. If the CSI and the other scores remain stable across simulations, the strategy is marked as robust. If they collapse or vary wildly, the strategy is fragile — regardless of what the net profit line shows.

This gives you confidence intervals across 20+ metrics, displayed as gauges — the same visual system as the Strategy Analyzer — so you can immediately see where your strategy holds up and where it doesn't.

Simulation Methods

AlgoChef offers five simulation methods, each testing a different aspect of robustness:

1. Shuffle (Fisher-Yates)

Randomly reorders your trades. Tests whether the strategy's profitability depends on the specific sequence of wins and losses.

Best for: Checking if results are order-dependent.

2. Bootstrap (With Replacement)

Randomly samples trades from your history, allowing the same trade to appear multiple times. Simulates "what if I had more/fewer of certain types of trades."

Best for: Estimating the range of possible outcomes with your trade distribution.

3. Block Bootstrap

Preserves consecutive trade sequences (streaks) while shuffling blocks. Maintains the temporal structure of winning and losing streaks.

Best for: Strategies where streaks matter (mean-reversion or momentum systems).

4. Parametric

Generates synthetic trades based on your strategy's statistical profile (mean, standard deviation, skewness). Creates entirely new trade sequences.

Best for: Testing if the statistical edge holds in synthetic scenarios.

5. Stress+ (Shuffle + Noise)

Shuffles trades AND adds random noise to each trade's P&L. The most aggressive test.

Best for: Worst-case scenario analysis with degraded trade quality.

Robustness Test (Full History)

What It Tests

The Robustness Test answers: "Is my strategy's edge real, or just a product of lucky trade sequencing?"

It runs thousands of simulations across your entire trade history, reshuffling and resampling trades to see if your strategy's quality scores hold up — or collapse.

How to Read Results

Robustness Grade

The headline metric. It measures tier stability — the percentage of simulations where your strategy's CSI stays in the same tier as the median result.

Tier Stability %

The exact percentage of simulations that stayed in the same CSI tier. A strategy with 94% tier stability means only 6% of simulations dropped (or rose) to a different tier.

Score Robustness

Shows how stable each of the other scores is across the simulations:

• Profitability Score — Does return quality hold up under randomization?
• Risk Score — Do drawdowns and volatility stay controlled?
• Confidence Score — Does statistical reliability remain stable?

If one score is significantly less stable than the others, it shows where the strategy is fragile.

Equity Fan Chart

A visual showing all simulation equity curves overlaid. The spread of the "fan" tells you how much path dependency affects outcomes. A tight fan = robust; a wide fan = order-dependent.

Confidence Intervals

These intervals are calculated across 20+ metrics, not just net profit.

Health Validation (IS/OOS)

What It Tests

Health Validation answers: "Does my strategy's edge persist out-of-sample?"

It splits your trade history into an earlier baseline (in-sample) and a recent period (out-of-sample), then runs Monte Carlo simulations on each, in line with the Health Score. If the IS and OOS distributions diverge significantly, your strategy may be overfit to historical data.

How to Read Results

• Health Score Comparison — Compares the database health score with the Monte Carlo-calculated health score to validate consistency
• Method Histograms — Distribution of scores across simulation methods, showing overlap (or divergence) between IS and OOS
• IS vs OOS Divergence — Quantifies how much the out-of-sample results deviate from in-sample expectations

Capital Sizing

What It Tests

Capital Sizing answers: "How much capital do I actually need to trade this strategy safely?"

It uses Monte Carlo simulations to model Risk of Ruin (RoR) and maximum drawdown probabilities at different capital levels, giving you data-driven capital recommendations.

Configuration Options

Click Run Capital Analysis to start.

How to Read Results

Dual Curve Chart

Shows two curves plotted against starting capital:

• Risk of Ruin curve — Share of simulations that hit your ruin rule at each capital level
• Max Drawdown curve — The deepest fall from peak at your confidence level (for example P98: 98% of simulations fall no further)

Where these curves cross your targets shows roughly how much capital you need. Goal Seek calculates a single estimate and marks it on the chart.

Goal Seek: Dual-Constraint Capital Planning

Finds the minimum starting capital where both of these hold:

• Risk of Ruin is at or below your target
• Maximum drawdown (at your confidence level) is at or below your target

The minimum is solved directly from the simulations, not read off the chart's capital steps, then rounded up to the next $100. The Risk of Ruin and drawdown shown next to it are measured at that rounded figure (rounding up only adds room, so both stay within your targets in the simulations), and the same figure appears on the chart as the Required marker.

The ruin rule

A simulated account is "ruined" when it breaks the rule you pick. Risk of Ruin is the share of simulations that hit the rule. Capital Sizing and the Survival Test use the same rule.

A few details:

• The buffer is "plus X% on top". With the default 10%, the line is margin × 1.10. Set it to 0% to use plain margin.
• The "highest point so far" includes your starting capital, so the drop is measured from the best the account has been, not from where it started.
• Contracts traded is the largest contract count recorded in the strategy's trades. If your import has no contract column, the tool assumes 1 contract and says so next to the rule.
• The margin line is a fixed dollar amount. More starting capital does not move it. It gives the account more room above it, which is why Risk of Ruin falls as capital rises.
• A futures strategy whose instrument has no margin requirement set uses the floor rule instead, and a note under the rule links to the Instruments page. Set the instrument's margin there to use the margin rules. The tool never guesses a margin for you.
• The floor, like the margin line, is a fixed dollar amount and is "below" (strict): an account sitting exactly on the floor is not ruined. A floor of $0 means the drop alone decides. If your starting capital is at or below the floor, the tab warns you, because every simulation is ruined at the start or by its first loss.

Worked example (simulated): margin is $10,000, you trade 1 contract, and the buffer is 25%. The ruin line is $10,000 × 1 × 1.25 = $12,500. Every simulation that ends a trade with the account below $12,500 counts as ruined. Starting with $30,000 leaves $17,500 of room above the line; starting with $15,000 leaves only $2,500. If 90 of 1,000 simulations cross the line, the simulated Risk of Ruin is 9%.

The rule is shown in plain words under the controls, for example "Ruined if the account goes below $12,500 (margin $10,000 × 1 contract + 25%)". For the floor rule it reads, for example, "Ruined if the account goes below $3,500 or falls 50% or more below its highest point." Changing the rule, buffer, floor or drop recalculates the curves instantly from the saved simulations, with no re-run.

Survival Test (Stress Testing)

What It Tests

The Survival Test answers: "Will my strategy survive degraded market conditions?"

It systematically degrades your strategy's performance metrics — reducing win rate, shrinking profits, and increasing losses — then runs Monte Carlo simulations under each scenario to see if the strategy remains viable.

Predefined Scenarios

Seven scenarios with progressively worse conditions:

The degradation scenarios progressively stress the three core performance levers. The consecutive loss scenarios test streak resilience independently.

How to Read Results

Survivability Score (0–100)

The headline metric. A weighted average of normalized survival rates across all scenarios, where more severe scenarios carry less weight (since they represent increasingly unlikely conditions).

Survival Rate per Scenario

Each scenario shows what percentage of simulations survived (didn't hit ruin). Ruin here uses the same ruin rule as Capital Sizing, so a futures strategy is judged against its margin line (plus buffer) and any other strategy against its dollar floor and its drop from peak. The thresholds for "good" survival vary by severity — a 50% survival rate under Extreme Stress is actually rated Excellent, while the same rate under Baseline would be Concerning.

Severity-Aware Thresholds

Set Starting Capital ($), the ruin rule and Number of Simulations, then click Run Stress Tests. The education tabs above the results (Overview, Baseline, Max Consecutive Losses, Mild/Moderate/Severe/Extreme Stress) explain each scenario in detail.

Related: Crisis Dependency Analysis

After stress-testing with Monte Carlo, you may also want to check whether your strategy depends on market crises. Crisis Dependency Analysis removes trades from major crisis periods (2008 GFC, COVID, etc.) and recalculates your scores to reveal if your edge disappears without market panics.

Robustness vs Survivability: Quick Reference

These two tests answer different but complementary questions:

Tips for Getting the Most from Monte Carlo

1. Run all five methods — Each reveals different fragilities. A strategy that passes Shuffle but fails Parametric may have distributional assumptions baked in.
2. Check Score Robustness — If the Profitability score is stable but the Risk score is not, look closely at how the strategy handles losing periods.
3. Use Capital Sizing before going live — Goal Seek gives you a simulation-based estimate at your chosen confidence level, under your assumptions, not a guess.
4. Don't ignore consecutive loss scenarios — Even strategies with high win rates can be devastated by unexpected streaks.
5. Run Health Validation before trusting robustness — A strategy that's robust in-sample but diverges out-of-sample is likely overfit.
6. Check crisis dependency — A strategy that only works during market panics will underperform in calm markets. Use Crisis Dependency Analysis to find out.