Maximum drawdown: the number that decides whether you can stay in

The definition, precisely

Maximum drawdown is the largest percentage fall from any equity peak to any subsequent equity trough. Notice the ordering: it falls after the peak, before the next peak. That matters because a strategy can be profitable overall and still have a 30% drawdown, and the two facts are not in conflict.

It is usually quoted with a date range and sometimes with a recovery time. The date range is essential — a 20% drawdown over three weeks and a 20% drawdown over fourteen months are very different experiences, even though the number is identical.

Why it matters more than return

Two effects push drawdown ahead of return in importance. The first is arithmetic, and it is unforgiving.

The second effect is behavioural, and it is not a weakness

A drawdown you cannot sit through is, for you, the same as a strategy that does not work. This is not a character flaw to be fixed with discipline articles. It is a constraint, like account size or available screen time, and a sensible response is to size positions so the historical worst case stays inside what you can actually tolerate.

That is why we publish drawdown next to return on every strategy page rather than in a footnote. A return without its drawdown is an advertisement; with it, it becomes information.

What normal looks like

It is easy to read a single strategy's numbers in isolation and conclude that anything above a few percent is unacceptable. To calibrate, here is the distribution across our public library.

Depth is not the whole story

Two more properties of a drawdown change how it feels. Duration is the first: a strategy that spends eighteen months underwater is painful even if the depth is modest, because the opportunity cost compounds alongside the loss. Recovery time is the second, and it is what most people underestimate when they look at a chart in hindsight.

A practical way to compare candidates is to look at the top three drawdowns rather than only the worst. If the second and third are close to the first, the strategy has a consistent shape. If the worst is far larger than the others, the strategy has one bad episode, and you should want to know what caused it.

Measuring it honestly

• Use the same cost assumptions you would trade with, or the drawdown is understated.
• Measure on the account size you will actually run, since sizing changes the curve.
• Include the period before the strategy was selected, if the data exists.
• Recompute after any parameter change — a drawdown that only appears after tuning is not the same strategy.
• Check that the drawdown was survivable at the position size you intend, not at one lot.

What to do with the number

Decide your tolerable drawdown before you look at any strategy. Then discard candidates whose historical worst case exceeds it, rather than trying to argue yourself into them. The strategies you keep will be less exciting on the return column and far more likely to still be running in a year.