Common mistakes when defining drawdown
Drawdown measures how far performance drops from a previous high point. Common mistakes usually come from skipping that “from a peak” part, mixing percent and absolute forms, or using unclear assumptions. The result is a number that sounds precise but can mean different things across reports.
This matters because drawdown is used to describe downside pressure and recovery time. If the definition is inconsistent, comparisons between strategies, time periods, or platforms become unreliable.
Drawdown mechanics: what the definition should specify
A clear drawdown definition includes at least four mechanics:
- Reference point (peak): drawdown is computed from the most recent maximum value before the decline.
- Measurement basis: you must state whether you use absolute change (units of account) or percentage change.
- Time alignment: you need the same sampling frequency for all comparisons (for example, daily vs per trade). Different sampling can hide or exaggerate extremes.
- Data inputs: define whether the calculation uses net value (after relevant costs) or gross value (before costs).
A frequent misunderstanding is to treat drawdown as “the amount you lost.” In practice, drawdown is not simply cumulative loss; it is a decline from a prior peak, and it can recover without changing the peak logic.
Evidence and examples: how mistakes change the result
Consider the same account values over time: 100 → 120 → 90.
- Peak-to-trough absolute drawdown: 90 − 120 = −30 (a 30-unit drop from the peak).
- Peak-to-trough percent drawdown: (90 − 120) / 120 = −0.25 = −25%.
A common mistake is to report only one form and compare it to a report that uses the other. Another mistake is to use a “minimum since start” framing. If you instead measure from the initial value 100, the drop 100 → 90 is −10 (−10%), which is not the same as peak-to-trough drawdown.
A second example failure mode is unstated assumptions about timing. If one provider samples equity only at end-of-day while another tracks intraday minima, the reported worst drawdown can differ even if the underlying performance is identical.
Finally, a mistake that creates misleading confidence is mixing real outcomes with model outputs without noting what was included (for example, costs, execution slippage, or gaps). Historical relationships often fail to predict future behavior, so verification should focus on definition consistency, not just on past numbers.
Limitations and risks: where drawdown definitions can fail
At least one material limitation is that drawdown depends on the quality and timing of the input series. If you cannot verify the sampling rule, the peak rule, and the cost treatment, the number is not fully interpretable.
Other common risks include:
- Percent vs absolute confusion: the same path yields different magnitudes and different perceived severity.
- Inconsistent “net” vs “gross” definitions: costs can materially change declines.
- Recovery masking: some summaries emphasize the maximum drawdown and ignore how quickly recovery happened, which affects risk perception.
- Jurisdiction and reporting differences: provider reporting conventions can vary, so you should not assume two drawdown metrics are computed the same way.
Because outcomes vary with market conditions, costs, execution, and reporting conventions, a drawdown figure alone cannot confirm safety.
Verification checklist and next question to ask
To verify a drawdown definition, check these items in the same order:
- Ask what the “peak” is (most recent high, global high, or another rule).
- Confirm absolute or percent and ensure comparisons use the same basis.
- Check the sampling frequency (daily, per trade, intraday minima) and the timestamp logic.
- Confirm net/gross treatment (whether costs are included in the series).
- Look for stated assumptions: without them, two drawdown numbers might not be comparable.
If you want to go one step further, the next question is: “Can I reproduce the drawdown from the reported equity/value series using the stated rules?” A reproducible calculation is a strong sign the definition is internally consistent.