Maximum drawdown (MDD)

Maximum drawdown is the largest observed drop from a peak in an investment's value to the lowest point reached before a new peak is set. It captures the worst loss an investor would have experienced by buying at the top and holding through the decline, over a given history.

It is calculated as (Trough value - Peak value) / Peak value, expressed as a negative percentage, taken over the largest such decline in the period examined. Unlike volatility, which measures the dispersion of returns generally, maximum drawdown looks specifically at the single worst cumulative loss, making it more intuitive to investors who think in terms of how much they could have lost.

Maximum drawdown is widely used to evaluate strategies in backtesting and to set risk limits, because large drawdowns are what actually drive investors to abandon a strategy, even one with a strong long-run average return, unlike a smoother risk figure such as the sharpe ratio or sortino ratio. The key pitfalls: it is a single historical event, so it says nothing about how likely a similar or worse drawdown is going forward, and it depends heavily on the observation period — a strategy's worst drawdown may simply not have happened yet in the sample used to measure it.