Drawdown review: the core idea
Drawdown review is the process of examining how large declines happen in a performance record, typically from a prior high point, and how those declines evolve over time. In plain terms, if a performance series rises to a peak and then drops, the drawdown measures the drop relative to that peak.
Beginners often start with a risk-first question: “Did the outcomes include periods of steep decline, and were those declines consistent with how I expected risk to behave?” Drawdown review supports that question by making declines visible and quantifiable.
How it works in practice (and what you must define first)
A drawdown review depends on definitions and inputs. Before any implication, you need to decide what “performance” means and how drawdown is calculated.
Choose the performance series
Most reviews use an equity curve concept (a running value over time). Your equity curve could represent account value, portfolio value, or a backtest balance series. The key is consistency: the same definition must be used when comparing periods or providers.
Define drawdown mathematically
A common approach uses the distance from a previous maximum:
- Track the running peak of the performance series.
- Compute the drawdown at each point as the percentage (or absolute amount) below that running peak.
- The “maximum drawdown” is the worst drawdown experienced during the chosen time window.
If you use percentage drawdown, state the formula you applied (percent below peak). If you use absolute drawdown, state the currency or units. Changing measurement changes the result.
State assumptions for examples
Any numerical example needs explicit assumptions, such as:
- time window length (e.g., 3 months vs. 2 years)
- whether results include transaction costs and financing effects
- how missing data or weekends are handled
- whether the series is smoothed or uses raw marks
Without these assumptions, two people can discuss “maximum drawdown” and still be talking about different calculations.
Evidence and a simple worked example (with explicit assumptions)
Assume a performance series measured once per day, including all relevant costs for that backtest or record.
- Day 1: equity = 100 (peak = 100)
- Day 2: equity = 92 (drawdown = (92−100)/100 = −8%)
- Day 3: equity = 96 (running peak still 100, drawdown = −4%)
- Day 4: equity = 90 (drawdown = (90−100)/100 = −10%)
In this example, the maximum drawdown over the window is 10%. A drawdown review would then also consider timing: when the worst decline occurred, and whether there were long recovery periods.
The material point for beginners is not the number itself, but the method: you can independently verify the drawdown only if the performance series, time stamps, and calculation approach are clearly defined.
Limitations and risks (common failure modes)
Drawdown review is useful, but it has limitations. At least one failure mode matters whenever you rely on drawdown metrics.
1) Data and definition mismatch
Two “drawdown reviews” can disagree because they use different equity curves, different timestamps, or different drawdown definitions (percentage vs. absolute, point-in-time vs. segment-based). If costs are excluded in one series but included in another, declines may look artificially smaller.
2) Cost and execution effects can dominate
Even if historical calculations show a certain decline size, real outcomes can differ when execution, spreads, slippage, or financing differ from the assumptions used to build the series. Historical relationships do not establish future results.
3) Survivorship and selection bias
If the review only includes periods that “look good” or excludes events that are inconvenient to compute, drawdown statistics can be understated. A robust review uses complete data for the chosen period.
4) Overfitting to past drawdown shape
Beginners sometimes focus on the “shape” of past declines and treat that shape as predictive. That is risky: future markets can produce different volatility regimes and different path-dependent outcomes.
Verification and next questions
To verify drawdown review facts independently, check four things:
- the exact performance series definition
- the drawdown formula and units
- the chosen time window
- whether costs and timing assumptions are included
Then ask what the review cannot answer. For example, drawdown review describes declines in the past record; it does not guarantee safety, stability, or future performance.