What Is Maximum Drawdown? And Why It Understates Your Risk

Maximum drawdown is the largest peak-to-trough decline in an equity curve, the deepest fall from a high-water mark to the low that follows before a new high is made. It is the standard way traders describe worst-case pain, and it is genuinely informative. It also understates risk in a way that is easy to miss, and the flaw is contained in two words people skip over: so far.

One draw from a distribution

Your maximum drawdown is a property of one specific sequence: the exact order in which your trades happened to occur. But that order was one of countless possible orderings of the same trades, and a worse one almost always existed. The same set of winners and losers, with the losers clustered more tightly together, produces a deeper drawdown from identical raw performance. You experienced one path drawn from a distribution of paths, and maximum drawdown reports that single path as though it defined the limit of what could occur. It does not define the limit. It records what happened to occur, which is usually not the worst arrangement your trades were capable of.

It grows with the length of the record

Maximum drawdown tends to increase the longer a strategy runs, because a longer history simply offers more opportunities for an unlucky run of losses to line up. This makes the raw number ambiguous: a small maximum drawdown can indicate a genuinely resilient strategy, or it can indicate a short track record that has not yet lived through its bad stretch. The figure on its own cannot distinguish the two, so a flatteringly small drawdown deserves the immediate question of how much history produced it.

It says nothing about time

Two strategies can share an identical maximum drawdown and offer completely different experiences. A twenty percent drawdown recovered within a month and a twenty percent drawdown that takes two years to climb out of are the same number and nothing alike in practice, both financially and psychologically. Depth is only half of a drawdown; duration is the other half, and the single figure captures only the first.

The honest replacement

The fix is not to quote a larger number, it is to stop quoting a single number at all. Resampling your own trades many times, in many orders, produces the distribution of drawdowns your strategy can generate, rather than the one it happened to generate. That turns the question from what was my worst drawdown into how bad can this reasonably get, which is the question your risk tolerance actually needs answered. Viewed that way, the worst drawdown you have personally seen is best treated as a floor on your risk, not a measurement of it. Building that distribution from a trader's own record is exactly what the resampling in our case study does.