Max Drawdown Explained With Real Examples

Max drawdown is the largest drop in an account's value from a previous high to a later low, written as a percentage. If an account grows to $20,000, falls to $12,000, and then recovers, its max drawdown for that stretch is 40%. It tells you the worst loss you would have had to sit through, which is often a better guide to whether you could stick with a strategy than its average return.

Below we show how to calculate it, how it relates to recovery time, and what real backtest drawdowns look like next to their benchmarks.

How to calculate max drawdown

The formula is simple:

Drawdown = (lowest value after a peak − the peak) ÷ the peak

You track the running high of the account. Every day, you measure how far the current value sits below that high. The biggest of those gaps over the whole period is the max drawdown.

A worked example with made-up numbers:

The max drawdown here is 35%, the fall from $13,000 to $8,450. The earlier 20% dip does not count as the maximum because a bigger one came later. Note also that month 6 is still in a drawdown: the account has not yet regained its $13,000 high.

Two details matter. First, drawdown is measured from a peak, not from your starting deposit. An account that doubled and then fell 40% is still up overall, but the person who joined at the top lost 40%. Second, a strategy that checks prices daily will usually show a larger drawdown than the same strategy measured monthly, because daily data catches lows that monthly data skips over.

Why losses need bigger gains to recover

The most useful fact about drawdowns is that recovering from one takes a larger percentage gain than the loss itself.

A 50% loss turns $10,000 into $5,000, and $5,000 has to double to get back to $10,000. This is why small differences in drawdown at the deep end matter so much. Going from a 40% to a 55% fall does not sound dramatic, but the gain needed to recover goes from 67% to more than 120%.

Drawdown duration: the number people forget

Max drawdown tells you how deep. Duration tells you how long. A 30% fall that recovers in three months feels very different from a 30% fall that takes four years to recover.

When you read any backtest, look for three dates: the peak, the bottom, and the day the old peak was finally regained. The time from bottom to recovery is where most people give up on a strategy.

Our Gold Breakout strategy is a good illustration of why duration matters. All numbers below come from a historical backtest, a computer test on past gold prices from June 2004 to November 2026 at 2% risk per trade. Nobody earned them, and future results will differ. Gold Breakout's caveat: doubling trading costs cuts its yearly result roughly in half.

• Over the full period, the backtest returned +15.1% a year with a worst fall of -41%. Gold simply bought and held returned +11.0% a year with a worst fall of -45% over the same period.
• The deepest fall ran from 2018-05-15 to 2019-10-28, a drop of 41%.
• The account did not regain its old high until 2023-03-23, which is 41 months after the bottom.

So a person following this strategy would have watched the account sink for about a year and a half, then waited more than three more years to get back to even. The headline yearly return hides that experience completely. You can read more about recovery periods in how long strategies take to recover.

Real drawdowns compared with their benchmarks

A drawdown number on its own does not mean much. You need to see the same measurement for a simple alternative over the same period. The table below shows each AlgolaneBots strategy's worst fall in its historical backtest next to the benchmark's worst fall over the same dates. All figures are simulated.

Each line needs its context:

• Leveraged Growth rotates between leveraged growth funds, volatility funds and bonds. Its backtest returned +129.1% a year against +13.9% for the S&P 500 and +18.2% for the Nasdaq 100, but its worst fall of 53%, during the Covid crash in early 2020, was much deeper than either index. That fall recovered in about 3 months. Its caveat: its rules may be fitted to the years after 2010, and leveraged funds can lose value quickly.
• Market Switch holds one fund at a time and switches when the market looks stretched. Backtest: +71.9% a year, worst fall -42%, against +15.0% and -34% for the S&P 500 over its period. The 42% fall in March 2020 recovered about 3 months after the bottom. Its caveat: nearly half of its tested profit came from UVXY, and it failed outside the US.
• Big Tech Trend returned +37.3% a year in the backtest with a -29% worst fall. That fall began in June 2026 and had not recovered at the end of the data. Its caveat: the stock list was chosen with hindsight, and three stocks made 81% of the profit.
• Steady Growth returned +26.5% a year with a -20% worst fall that recovered within about a month. Its caveat: its seven names were picked because they already did well, which flatters the test.

The drawdown you have not seen yet

A backtest's max drawdown is the worst fall that happened in the data. It is not the worst fall that can happen. A longer history, a different market or simply a new kind of crisis can produce something deeper.

Our own data shows this clearly. For Market Switch and Leveraged Growth we also ran simulated three-fund versions back through the dot-com crash. In that simulated test, Market Switch fell 100%, meaning the account would have been wiped out, compared with a 47% fall for the S&P 500, and it took 259 months to get back to its pre-crash value against 78 months for the index. The simulated Leveraged Growth version fell 91% over the same crash and took 115 months to recover. Neither of those losses appears in the 2011-2026 headline numbers.

That is the practical lesson: treat a backtest's max drawdown as a floor for how bad things could get, not a ceiling. For a broader view of worst cases, see how much a trading bot can lose.

How to use max drawdown when choosing a strategy

A few ways to put the number to work:

1. Translate it into dollars. "A 40% drawdown" is abstract. "My $10,000 could fall to $6,000 and stay there for three years" is concrete. Ask whether you would keep following the rules at that point.
2. Compare it with simply holding the market. If a strategy's drawdown is deeper than the index's, it needs to earn that extra pain. If it is shallower, check whether that came from a genuine rule or from hindsight.
3. Look at drawdown and return together. Some people use a ratio of yearly return to max drawdown as a rough quality score. It is a blunt tool, but better than return alone. Our related post on CAGR vs total return covers the return side.
4. Check how risk settings change it. In Gold Breakout's backtest, raising risk per trade from 1% to 5% took the worst fall from -23% to -74%. Higher risk raised the backtest return too, but the drawdown grew much faster than most people would be comfortable with.
5. Experience it at small scale first. Watching a paper account fall 15% teaches you more about your own tolerance than reading any table.

Key takeaways

• Max drawdown is the largest fall from a running high to a later low, measured as a percentage.
• Losses need larger gains to recover: a 50% fall needs a 100% gain to get back to even.
• Duration matters as much as depth; Gold Breakout's deepest backtest fall took 41 months to recover after the bottom.
• Always compare a strategy's drawdown with its benchmark over the same period.
• A backtest's worst fall is not the worst possible fall; longer or different histories can show deeper ones.

This is general education, not personal financial advice. All strategy figures above are from simulated historical backtests.

See drawdowns in real time

Numbers on a page do not tell you how a falling account feels. AlgolaneBots lets you run all five strategies on a $10,000 paper account for a free 90-day trial, with fills at real market prices and no card needed. Start your paper trial and read the risk page for the full picture.