Max drawdown
The biggest fall from a high point to a later low, before a new high.
4 min readLesson 7 of 57Interactive
In short
A −30% drawdown means that at the worst moment you'd have watched 30% disappear from the peak. It's a good guide to how painful a strategy is to stick with, which matters as much as its return.
Why it matters
Averages hide the worst moments, and the worst moments are what make people quit. A strategy that returns 10% a year but once fell by half is a very different thing to hold than one that returned 8% and never fell more than 15%. Many investors who own the first sell near the bottom and never see the recovery.
Max drawdown answers a blunt question: if you had invested at the worst possible time, how much would you have watched disappear before things got better? It's the number to look at when asking whether you could actually stick with something.
How it works
Walk through the values day by day, keeping track of the highest value so far. On each day, the drawdown is how far the value sits below that running peak, as a percentage of the peak. The max drawdown is the biggest of those. MarketLab works it out this way from daily closing values, so a dip during a day that recovered by the close doesn't count.
Losses and gains aren't symmetrical. After a fall, you're starting from a smaller amount, so you need a bigger percentage gain just to get back to where you were. A 10% loss needs an 11.1% gain; a 20% loss needs 25%; a 50% loss needs a 100% gain; an 80% loss needs a 400% gain.
That's why deep drawdowns are so costly. Recovering takes time, and the time spent getting back to even is time not spent growing.
Drawdown = 1 − value ÷ highest value so far; gain needed to recover = 1 ÷ (1 − drawdown) − 1
How far below its peak the value is, and the percentage rise it needs to get back there.
A worked example
$10,000 rises to a peak of $15,000, falls to $9,000, climbs back to $14,000, and finally reaches a new high of $16,000.
- Drawdown at the low: 1 − $9,000 ÷ $15,000 = 1 − 0.60 = 0.40, a 40% fall from the peak.
- The dip from $15,000 is measured from the peak, not from the $10,000 start, so the max drawdown is −40% even though the value never fell below $9,000.
- Gain needed to recover: $15,000 ÷ $9,000 − 1 ≈ 0.667, so a rise of about 66.7%.
- For a 50% fall: $100 drops to $50, and getting back to $100 means gaining another $50, which is 100% of $50. In the formula: 1 ÷ (1 − 0.5) − 1 = 1, or 100%.
Result: The max drawdown was −40%, and it took a 66.7% gain just to get back to the old peak. The total return of +60% hides how hard that would have been to sit through.
Try it
Find the worst fall
Over two years, the worst fall from a high was 73.7%, lasting 497 trading days from peak to bottom.
Max drawdown
−73.7%
Peak to bottom
497 days
Back to the peak
Not yet
…you need a gain of +100.0% to get back to where you started.
Generated prices for illustration, not market data.
How to read it
- 0% to −10%
- Mild. Normal for a cautious mix, or for stocks in a calm year.
- −10% to −20%
- A correction. Broad stock markets have had falls like this every year or two on average.
- −20% to −40%
- A bear market for an index, or an ordinary bad stretch for a single stock.
- Worse than −50%
- Severe. It takes a doubling to recover. Single stocks and narrow strategies can get here; broad indexes rarely do.
In the real world
- History has some famous ones. The S&P 500 fell roughly 55% from its 2007 peak to its 2009 low. In early 2020 it dropped about a third in a few weeks, then recovered within months. The tech-heavy Nasdaq fell nearly 80% from 2000 to 2002 and took around 15 years to regain its old high.
- Hedge funds and professional managers are often judged on drawdowns as much as returns. Some clients pull their money if losses pass a set limit, and many funds have rules to cut risk when a drawdown deepens.
- Financial planners use it to set expectations. Asking "could you watch your savings fall by a third and not sell?" is often more useful than talking about volatility.
Common mistakes
- Measuring from the starting value instead of the peak. A drawdown is always measured from the highest point reached so far.
- Thinking a 30% loss needs a 30% gain to recover. It needs about 43%.
- Assuming a backtest's worst drawdown is the worst that can happen. The future can always bring a deeper fall than the past.
- Ignoring how long a drawdown lasted. A 25% fall that took five years to recover can hurt more than a 35% one that recovered in months.
See it in MarketLab
Test yourself
Optional: 4 quick questions on the words, using the idea, and how it plays out in real markets.
Loading your answers…