What is maximum drawdown?

What it means

It is a single number describing one event, which is both its usefulness and its limitation. It tells you the worst thing that happened in the sample, and says nothing about whether something worse is likely in the future.

Why it matters

The most important thing to understand about maximum drawdown is that the observed figure is a floor, not a ceiling. A backtest over one year that shows 20% is telling you 20% was enough to contain that particular year — a different year with a different sequence of trades will produce a different number, frequently a larger one.

What this changes in practice

This is why the honest way to present it is alongside the period, the sample size, and an out-of-sample check. A maximum drawdown quoted without those is a marketing number. Ours are 21.2% and 20.5% across two separate periods, and both are on the order of ten trades per month, which is a small enough sample that neither should be treated as precise.