Monte Carlo (trade order)
Replays a backtest's trades in 1,000 random orders to show how much of the result was the luck of the sequence.
Formula
Shuffle: the same trades, in random order Bootstrap: trades drawn at random with replacement Skip: each trade dropped with a 10% chance, then shuffled
Inputs
- Each path. Every trade's P&L as a fraction of starting capital, compounded through the path.
How to read it
Read the 5th-percentile return and drawdown as a plausible bad case for the same edge. Loss odds is the share of paths ending below the start; ruin odds the share ending at or below half of it.
How to use it
A strategy whose 5th-percentile path still makes money, with a drawdown you could sit through, is sturdier than one that needs its exact trade order.
Limitations
Treats trades as independent and cannot invent a market regime the backtest never saw. Needs at least 10 trades; a fixed random seed makes the same test read the same every time.
In Kereon
Strategies · All statistics. 1,000 paths for each of the three methods.
Also called: reshuffle, bootstrap, loss odds, ruin odds, robustness.