Monte Carlo Simulation¶

Equity curve probability cloud from trade-return shuffling.

Overview¶

Shuffles historical trade returns, adds noise, and runs many simulations to estimate probability distributions of outcomes. Tests for overfitting by comparing original backtest result against shuffled permutations.

Data Input¶

Upload CSV with columns: date, trade_return
Or use built-in sample data — 200 trades with slight positive edge (+0.15% per trade) and loss clustering

Simulation Engine¶

for each simulation i:
    shuffled = permute(trade_returns) + noise(±0.3%)
    equity[t+1] = equity[t] * (1 + shuffled[t])

Default: 1,000 simulations, starting capital $100,000.

Statistics¶

Probability Metrics¶

Probability of loss — % of simulations where final value < starting capital
Probability of 20%+ drawdown — % of simulations with max drawdown ≥ 20%
Probability of 30%+ drawdown — % of simulations with max drawdown ≥ 30%

Percentile Distribution¶

5th, 25th, 50th, 75th, 95th percentiles of max drawdown distribution

Overfitting Check¶

Compares original backtest final value against shuffled outcomes:

original_pctile = mean(final_values < original_final_value)

>90th percentile — HIGH risk: original result looks too good, possible overfitting
>80th percentile — MEDIUM risk
≤80th percentile — LOW risk

Visualizations¶

Fan Chart¶

Cinematic probability cloud showing:

Individual curves — low opacity lines (0.015) in cyan
5th–95th band — wide probability range
25th–75th band — interquartile range
Median line — 50th percentile
Original backtest — solid white line

Histograms¶

Final portfolio values — distribution of ending values with original backtest line
Max drawdown distribution — distribution of worst drawdowns per simulation

Drawdown Distribution Detail¶

5th, 25th, 50th, 75th, 95th percentile values displayed as cards.

Interpretation¶

Auto-generated text summarizing probability of loss, drawdowns, median outcome, and original backtest percentile.

Usage¶

Upload CSV or use sample data
Set starting capital and number of simulations
Click "Run Simulation"