What percentage drawdown means
Percentage drawdown is a way to express how far an account falls from a reference peak, stated as a percentage. A common framing is:
- Pick a baseline equity (or net asset value) series over time.
- Identify the running peak equity up to each point.
- Compute drawdown as the drop from that running peak.
- Express that drop relative to the peak using a percentage.
In symbols, with equity (E(t)) and running peak (P(t)=\max_{s\le t}E(s)), drawdown at time (t) can be written as (D(t)=P(t)-E(t)) and percentage drawdown as (\text{DD%}(t)=\frac{P(t)-E(t)}{P(t)}\times 100). A negative or zero equity can cause practical interpretation issues, so “equity” should be defined clearly (for example, whether it includes unrealized gains/losses and whether it is net of fees).
How it works in practice: the key dependencies
Percentage drawdown is not a single universal number by itself; it depends on how the inputs are defined and maintained. Advanced considerations focus on these dependencies.
1) The definition of “equity” and timing convention
Two people can compute different percentage drawdowns from the same trades if they:
- Mark equity at different moments (e.g., end-of-bar versus tick time).
- Include or exclude unrealized P/L.
- Use account currency conversions differently.
If unrealized gains/losses are included in the equity series, intraday fluctuations can create new peaks and therefore change later percentage drawdowns.
2) The peak-tracking rule
The running peak (P(t)) is crucial. Edge cases include:
- Reset rules: Does the peak reset after withdrawals/deposits, or is the series normalized to remove those cash flows?
- Multi-currency reporting: Does the peak occur in converted reporting currency, which may change even when local performance is unchanged?
- Data gaps: Missing days or incomplete history can suppress peaks or exaggerate troughs.
Because percentage drawdown uses the peak in the denominator, a slight difference in when a peak is detected can materially change DD%.
3) Denominator choice and near-zero baselines
The denominator is the peak (P(t)). If (P(t)) becomes very small (or zero), (\text{DD%}) can become unstable or undefined. Even if such a situation is rare, the computation should specify what happens when the peak is zero or near zero.
A practical modeling approach is to state an explicit assumption: “Peak is positive for all time points used in the calculation,” or define a fallback handling rule for non-positive peaks.
4) Choosing the evaluation horizon
Percentage drawdown can be computed over different scopes:
- All-history drawdown (peak-to-trough over the entire series).
- Windowed drawdown (for example, last N months).
- Event-based drawdown (for example, from a strategy start date).
Different horizons yield different “maximum drawdown” values because the peak history differs.
5) Costs and execution effects
Even if you conceptually rely on the equity series, the equity series is affected by variable components such as trading costs, spreads, and execution timing. If two providers or backtests assume different cost models or execution assumptions, percentage drawdown will differ.
This is an implementation constraint: drawdown is “about the realized accounting,” not only “about price movement.” Costs can increase both the depth of troughs and the frequency of new peaks.
A simple worked example with explicit assumptions
Below is an example model intended to show how dependencies change the output. It is not based on live data.
Assumptions:
- Equity is marked continuously over time.
- The running peak is updated whenever equity reaches a higher value.
- Percentage drawdown uses (\frac{P-E}{P}\times 100).
Example equity path (in account currency):
- Day 1: (E=100) → running peak (P=100) → DD% = 0%
- Day 2: (E=90) → (P=100) → DD% = (\frac{100-90}{100}\times 100=10%)
- Day 3: (E=110) → new peak (P=110) → DD% = 0%
- Day 4: (E=95) → peak remains (P=110) → DD% = (\frac{110-95}{110}\times 100\approx 13.64%)
Key observation: the “same trough depth” can produce different DD% depending on the peak used. Day 4 produces a larger DD% because the peak is higher (110) than the earlier peak (100).
Material limitation: if the equity marking rule differs (for example, if peak updates are delayed), you may get a different sequence of peaks and therefore a different DD%.
Limitations and failure modes to watch
Percentage drawdown is widely used, but advanced interpretation needs caution. Common failure modes include:
1) Ambiguity from cash flows (deposits and withdrawals)
If equity changes due to deposits or withdrawals, a raw DD% calculation can mix performance with funding. Two series that differ only by cash-flow treatment can show different drawdowns. A careful definition should state whether equity is “starting from realized performance only” or “including cash flows.”
2) Non-comparable equity series
Percentage drawdown is only comparable when equity series are computed consistently. Comparisons break when one series includes unrealized P/L while another includes only realized P/L, or when cost and conversion assumptions differ.
3) Sensitivity to peak timing
Because the denominator is the peak, DD% is sensitive to the timing and detection of new peaks. Data granularity (daily vs intraday) can change the maximum DD%.
4) Interpretation vs predictive meaning
A larger historical percentage drawdown does not guarantee future results, and a smaller historical value does not imply future stability. Historical relationships do not establish future outcomes.
5) Over-reliance on a single summary number
Many analyses compress all history into “maximum percentage drawdown.” That summary can hide duration and recovery behavior. For example, two equity paths can share the same maximum DD% but differ in how long they remain near the trough.
How to verify claims about percentage drawdown
If you are reviewing a reported percentage drawdown value, verification means reproducing the computation from the stated inputs and rules.
A checklist approach:
- Obtain the equity time series used.
- Confirm the equity definition (unrealized included or not, accounting currency, net of costs or not).
- Confirm peak-tracking rules (how running peaks are updated, and how cash flows are handled).
- Confirm the exact drawdown formula and whether results are rounded.
- Recompute DD% and compare the maximum drawdown value.
Verification can still fail if the source does not provide the equity series or if the methodology leaves key assumptions unspecified. In that case, treat reported values as dependent on unknown conventions.