Direct answer
Maximum drawdown in forex is a way to summarize how large an account’s decline can be, based on historical account value. Mechanically, it looks at the biggest “peak to later low” move within a chosen time window. It does not predict future performance, and the exact number depends on what account-value series you use and how you treat costs and missing data.
Mechanism and definition
Maximum drawdown (often abbreviated MDD) is defined using an ordered sequence of account values over time—commonly an equity curve. “Equity” typically means the account’s value including unrealized gains or losses; “balance” typically means value after closing positions. In practice, the distinction matters: equity can move intraday as prices change, while balance changes mainly when trades are closed.
A standard way to view the mechanism:
- Identify a “peak” moment: a time when the account value reaches a local high relative to earlier times.
- Look forward from that peak to find the “trough” moment: the lowest account value reached after the peak.
- Compute the drawdown magnitude from the peak to that trough.
- Repeat across all peaks in the window and select the largest drawdown magnitude. That largest value is the maximum drawdown.
A common mathematical expression uses drawdown as a percentage:
- Drawdown% at time t = (PeakValue − Value(t)) / PeakValue, where PeakValue is the maximum Value observed up to time t.
- Maximum drawdown = the maximum of Drawdown% over the selected window.
This model is an accounting-style transformation of a time series: the “work” of maximum drawdown is performed by the peak-tracking rule and the max-of-drawdowns rule, not by any special forex-specific market pattern.
Inputs, outputs, and the calculation sequence
Inputs you must define
To calculate maximum drawdown for forex activity, you need:
- A time window (start date/time to end date/time). Results change if you extend or shorten the window.
- A time-ordered account-value series (for example, equity sampled at a fixed interval). Each point becomes Value(t).
- The peak definition: typically the running maximum up to each time point.
- A convention for costs and valuation: whether fees, swaps/rollover, commissions, and spreads are already reflected in the account value you use.
If your dataset comes from a platform report, you should use exactly the series that platform labels as the basis for equity and ensure that all costs are included consistently.
Outputs you can interpret
From that calculation you get:
- Magnitude: the largest observed percentage (or absolute amount) decline from a prior peak to a subsequent trough.
- Context (optional but common): the time of the peak, the time of the trough, and the drawdown duration.
Even though drawdown magnitude is a single worst-case summary, duration can matter for operational risk (for example, how long the account stays below earlier peaks). Maximum drawdown itself does not fully encode recovery speed; it only marks the worst decline.
A worked example using explicit assumptions
Assume an equity series sampled at five time points (you must treat this as illustrative, not a claim about any live market):
- Time 1: equity = 100
- Time 2: equity = 120 (peak becomes 120)
- Time 3: equity = 110
- Time 4: equity = 80 (trough after peak)
- Time 5: equity = 90
Compute drawdown relative to the running peak:
- After Time 2, peak is 120.
- At Time 4, drawdown% = (120 − 80) / 120 = 0.3333… = 33.33%.
- The maximum drawdown over the window is 33.33% because it is the largest drawdown% observed after tracking the peak.
If, instead, you used balance rather than equity, the values and timing could differ and so could the maximum drawdown.
Limitations and risks (material failure modes)
1) The result is only as good as the equity series
Maximum drawdown is derived from the data you feed into it. If your equity series:
- uses balance instead of equity,
- excludes certain costs,
- or is sampled infrequently, then the maximum drawdown can be understated or overstated.
2) Sampling frequency can miss the true trough
If you record account value only at coarse intervals (e.g., daily), the equity may have dipped between observations and then recovered before the next data point. Your computed maximum drawdown would then miss the deeper trough.
3) Window selection changes the “worst” event
A short window might not include the worst peak-to-trough episode, while a longer window includes more peaks and more chances to encounter a larger drop. This is especially important for forex, where strategies and account conditions can change over time.
4) Historical relationships do not establish future risk
Maximum drawdown summarizes the historical worst decline inside a defined window. It is not a guarantee of future declines, because future market volatility, execution quality, and cost conditions can differ.
5) Different conventions lead to different numbers
Two analysts can compute different maximum drawdown values using the same underlying account activity but different conventions, such as:
- using equity vs balance,
- using percentage vs absolute units,
- using different sampling intervals,
- or applying a different definition of where the peak resets.
Verification and next questions
You can independently verify maximum drawdown by doing two checks:
- Recompute from the raw account value series using a clear window and the same sampling interval.
- Confirm the peak-tracking rule: at each time point, PeakValue should be the maximum value observed so far, not a future maximum.
A useful next question is what series and convention your platform or data source provides (equity, balance, whether costs are included, and the sampling frequency). If those choices are not explicit, you should treat the reported maximum drawdown as dependent on those underlying definitions.
If you want, you can also compare maximum drawdown with related measures like drawdown duration or recovery metrics to better understand not only how deep the worst decline went, but also how long it took to recover—while remembering each measure captures only a specific view of risk.