Definition and core idea
Drawdown is a measure of how much a value falls from a prior high. In finance discussions, it is most often computed from a peak-to-trough change in an account’s equity (or another chosen “value” series).
A worked example starts by stating the exact inputs:
- Equity series: the sequence of account values over time (numbers you would record from your own account statement or a backtest output).
- Peak: the maximum equity value observed before the later low.
- Trough: the lowest equity value observed after that peak (until the next new peak, if you are tracking maximum drawdown).
Common output formats:
- Absolute drawdown = Peak − Trough
- Percentage drawdown = (Peak − Trough) / Peak
Worked numerical example (with explicit assumptions)
Assume the following, and note that these are choices you must apply consistently:
- We measure equity once per day at a specific timestamp.
- Equity includes unrealized P&L and realized P&L, and it already reflects relevant fees and financing that affect account value.
- We define the drawdown episode as the period from the peak day to the trough day.
Now consider an example equity timeline (chosen for illustration, not based on live data):
- Day 1 equity (peak) = 10,000
- Day 2 equity = 9,700
- Day 3 equity = 9,200 (trough)
Step-by-step:
- Peak = 10,000
- Trough = 9,200
- Absolute drawdown = 10,000 − 9,200 = 800
- Percentage drawdown = 800 / 10,000 = 0.08 = 8%
If you want maximum drawdown for a longer history, you would compute drawdowns for each peak-to-trough episode and then select the worst (largest) percentage drawdown observed across the dataset.
How this “works” and what can fail
The mechanics are simple, but interpretation depends on how you measure inputs. Key limitations and failure modes include:
- Measurement frequency changes the result: If you sample equity daily, you might miss intraday highs/lows and underestimate the true peak-to-trough swing.
- What “equity” includes is variable: Some series may exclude certain cashflows or costs, which changes peaks and troughs.
- Episode boundaries matter: Choosing the “next new peak” rule affects which troughs belong to which peaks.
- Costs and execution can distort comparability: If equity in one case includes different fees, spreads, or slippage assumptions, drawdown comparisons can be misleading.
- Historical relationships do not guarantee future behavior: Past drawdown patterns cannot establish that future drawdowns will be smaller, larger, or similar.
Verification and next question to ask
To independently verify a drawdown definition, check that any calculation you see is explicit about:
- the value series used (equity, balance, or another measure),
- the time sampling (daily, hourly, tick, etc.),
- the peak selection rule (peak before trough), and
- whether it reports absolute drawdown, percentage drawdown, or maximum drawdown.
A good next question is: “Does the drawdown definition in my source match my measurement choices, or does it change the interpretation?”