A Monte Carlo simulation takes your strategy's trade results and reshuffles their order many times — hundreds or thousands of runs — to show the range of outcomes those same trades could produce. One backtest shows you a single ordering, which may be lucky or unlucky. The reshuffles show what else could plausibly have happened.
What is a Monte Carlo simulation, in plain words?
It's a deck-shuffling exercise performed on your trading history. Write each trade's result on a card — +$100, −$60, and so on — then shuffle the deck and deal the cards in a new order. The total profit stays identical (the same cards are in the deck), but the journey changes: losing cards might cluster at the start, or the end, or spread out evenly. Shuffle and deal a thousand times, recording the worst losing stretch of each run, and you get a picture of the strategy's possible futures instead of its one recorded past. No advanced math is required to grasp the idea — the computer is simply asking, over and over: "what if the same trades had arrived in a different order?"
Why isn't one backtest enough?
Because a single backtest is one specific ordering of wins and losses, and that ordering is partly luck. Two strategies with identical trades can look wildly different depending on when the losses happened to land. If your backtest's losing trades were conveniently scattered, the equity curve looks smooth and the worst drawdown looks mild — and you might size positions based on that mildness. But nothing guarantees the future will scatter its losses just as politely. The sequence of trades is one of the least repeatable features of any backtest: entries and exits may have logic behind them, but which order the wins and losses arrive in is substantially chance. Monte Carlo strips that accident away by trying many orderings.
How can the same trades produce different drawdowns?
Because drawdown measures the worst stretch, and clustering losses deepens the stretch — here's a small worked example. Take ten trades: five wins of +$100 and five losses of −$60. Total profit is 5 × $100 − 5 × $60 = +$200, in every possible order. Now compare two orderings:
| Ordering A: alternate win, loss | Ordering B: all 5 losses first | |
|---|---|---|
| Sequence | +100, −60, +100, −60, ... | −60 ×5, then +100 ×5 |
| Final profit | +$200 | +$200 |
| Worst drawdown | $60 | $300 |
In A, every loss is immediately repaired, so the account never sags more than $60 below its peak. In B, the account digs a $300 hole before the first win arrives. Same trades, same profit — a drawdown five times deeper. A trader sized for a $60 drawdown might abandon (or be unable to fund) the identical strategy living through ordering B.
