What Is a Trading Drawdown? Definition & Measurement

A trading drawdown is a decline in account or strategy equity from a reference peak to a later, lower value, measured for as long as equity stays below that peak. It describes a path, not a single loss or bad day: the same dollar decline can sit inside a shallow, brief drawdown or a deep, prolonged one, depending on how far equity fell and how long it has stayed below peak.

That distinction matters because “I’m in a drawdown” is often used to describe a feeling — a stretch of losing trades — without specifying what is actually being measured. Depth, duration, and which equity value counts as the peak are three separate questions, and a trader can answer one without answering the others. This article defines the term precisely enough to support the decisions that depend on it: when to adjust risk, when a decline is unusual, and what a recovery plan is actually restoring.

What trading drawdown means

A trading drawdown is measured from a reference peak — the highest value the chosen equity series has reached so far — against the current value of that same series:

reference peak = highest value previously reached by the chosen equity series
current drawdown = (reference peak − current value) ÷ reference peak × 100

A drawdown begins when the series moves below its reference peak, and the reading changes with every new observation until equity returns to that peak.

Four related terms describe different points on this path. The current drawdown is the reading at one specific moment. A drawdown episode is the full stretch from when equity first drops below a reference peak until it returns to that peak. The maximum drawdown is the largest current-drawdown reading recorded during an episode, or over any stated observation period. A new high-water mark is established only when equity exceeds — not merely reaches — the previous reference peak.

For example: peak equity of $100,000 falls to a trough of $92,000, then recovers to a current value of $96,000. The maximum depth reached in this episode so far is 8% ($100,000 to $92,000); the current drawdown is 4% ($100,000 to $96,000). The episode is still open, because equity has not returned to $100,000. At exactly $100,000, the current drawdown returns to zero. Only above $100,000 does equity establish a new high-water mark and a new reference peak.

A drawdown is not the same as a losing trade or a losing streak. A losing trade is a trade-level outcome; a drawdown is a measurement of the equity path. A loss taken from a high-water mark can start a new drawdown, and a loss taken while a drawdown is already open can deepen it — but the number of losing trades involved does not by itself determine a drawdown’s depth or duration.

Peak-to-trough depth and recovery duration

A drawdown has two dimensions, and they answer different questions.

Depth is how far equity fell — the peak-to-trough decline, usually reported as a percentage of the reference peak or in R, the trader’s own risk unit. Depth alone says nothing about how long the decline lasted.

Duration is how long equity has stayed below the reference peak, typically measured in calendar days, sessions, or trades from when the drawdown began. A drawdown returns to zero — and its duration ends — once equity returns to the prior reference peak; that is a separate event from a new high-water mark, which requires equity to exceed the old peak, not merely reach it.

Dimension	Question it answers	What it does not answer
Depth	How far below the peak did equity go?	How long it took to get there or return
Duration	How long has equity stayed below the peak?	How severe the decline was at its worst point

The two can move independently. A drawdown that reaches 8% and returns to the reference peak within five sessions is a different event from one that reaches 8% and stays open for eleven weeks, even though both share the same depth figure. Because recovery arithmetic is asymmetric — a 10% decline requires an 11.1% gain on the reduced base to return to the reference peak, and a 20% decline requires 25% — depth alone already understates the work required to close a deeper drawdown, before duration is considered at all.

How the measurement basis is chosen

The reference peak and current value must come from the same consistently defined equity series; two separate choices in that definition change the reading independently.

The equity series itself. Common choices include realized or closed-trade balance, which counts completed trades only, and marked-to-market equity, which adds the current value of open positions to realized balance. Neither is a universal standard: the right choice depends on what the measurement is for.

Observation frequency. Separately from which series is used, it can be sampled after each closed trade, at the end of a session or day, or continuously through the session. A temporary intraday decline can appear in a continuously sampled series and disappear from an end-of-day series measuring the same trading, even when both use marked-to-market equity.

A reading from one series and sampling frequency isn’t directly comparable to one from another; fixing both before a decline occurs, rather than after, is what makes a drawdown figure auditable.

For the conventional statistical definition above, the reference is normally a historical high-water mark in the selected series. Evaluation and funded-account programs commonly use related but distinct terminology for a different purpose — an account-level loss limit — covered next.

Provider drawdown and loss-limit rules

Prop-firm and evaluation-account programs often use terms such as drawdown, maximum drawdown, maximum loss, trailing drawdown, and static drawdown for an account-level loss-limit mechanism — related to the statistical measurement above, but not the same calculation: a provider’s loss limit is a threshold designed to stop trading, not a description of an equity path after the fact.

Two designs are common:

Static loss floor. A fixed account threshold that does not move upward as account value increases — a threshold design, not the statistical reference peak fixed at a starting balance, though the two can look similar.
Trailing loss floor. A loss threshold that moves upward according to a provider-defined reference value as the account gains. Trailing systems can differ in what value updates the threshold, when it updates, whether it eventually stops trailing, and whether a breach is tested against balance, equity, or unrealized P&L — so “end-of-day trailing” names when the threshold updates, not necessarily when a breach is evaluated.

Evaluation and funded-account rules differ by provider and account type; the provider’s current official documentation governs any specific account. This article defines the general statistical concept and does not compare or endorse provider rules.

Normal variance versus execution deterioration

A drawdown by itself does not indicate a cause. The same peak-to-trough decline can occur because a genuine, unchanged edge produced an unlucky sequence, because execution drifted from the trader’s own rules, or both. Depth and duration describe the shape of the decline; they do not diagnose why it happened.

Separating those explanations requires comparing the drawdown against what the trader’s own historical win rate and payoff distribution would produce by chance, and separately checking whether process metrics — rule adherence, position-size accuracy, setup qualification — moved during the same window. How to tell normal drawdown variance from execution deterioration works through that full diagnostic; this article’s role is limited to defining the decline being diagnosed.

Drawdown in a risk ladder

Because a drawdown is a measurable state, it can serve as the trigger in a pre-defined risk-management rule: a stated depth crossing a boundary moves new trades into a reduced-risk state, and a separate restoration condition governs when normal risk returns. The definition matters here in a practical way — a risk ladder built on closed-trade equity will trigger at different moments than one built on continuously sampled equity, even applied to identical trading. Adjusting risk during a drawdown covers how to build that ladder, choose multipliers, and gate restoration. This article’s contribution is upstream of that mechanism: fixing the measurement so the ladder’s triggers mean the same thing every time they fire.

What to measure during recovery

A drawdown “recovering” can refer to two different things, and conflating them produces false confidence:

Equity recovery — the measured drawdown returning to zero as equity returns to the reference peak.
Process recovery — rule adherence, position sizing, and setup qualification returning to baseline levels, independent of what equity is doing.

Equity can recover while process metrics remain degraded, and process metrics can recover while equity is still below the reference peak — a trader can restore aligned execution and still be trading a losing stretch. Neither form of recovery certifies the other. A multi-day drawdown needs to track both across sessions rather than reset the question each morning; recovering from a multi-day drawdown covers that continuation.

When not to infer a cause

Three situations are commonly treated as evidence they are not:

A deep or unusual-looking drawdown, by itself, is not proof of execution failure or edge decay. Depth and duration describe the decline; they require a separate process and statistical comparison before either explanation is defensible.
A shallow or brief drawdown is not proof execution was sound. A short decline can still contain misclassified trades, oversized positions, or unqualified entries that a small sample simply didn’t reveal.
Reaching a new equity peak does not retroactively validate the trades that produced it. Equity recovery answers a financial question; it does not answer whether the trades along the way followed the trader’s own rules.

Treat “drawdown detected” as the trigger for a defined review, not as a conclusion about cause, discipline, or strategy quality on its own.

Where this connects

This article defines the state; it does not decide what to do about it. See adjusting risk during a drawdown for changing position risk while a drawdown is active, behavioral drawdown measurement for separating normal variance from execution drift, multi-day drawdown recovery for a drawdown spanning multiple sessions, restoring position size after a drawdown for stepping size back up as it resolves, how cognitive bias distorts decisions during a trading drawdown for how being below the reference peak changes judgment, not just measured risk, and how outcome bias distorts decisions during a trading drawdown for how the episode’s own ending, once known, can distort the later review of every decision made inside it.

Where Costante fits

Costante supports the behavioral review layer around a drawdown the trader has already identified from their own equity records: recording the risk-state triggers, session notes, and rule-adherence evidence tied to a decline, so execution can be reviewed alongside the equity path tracked separately.

Costante does not calculate drawdown, does not monitor live account equity, does not connect to a broker or funded-account provider, does not determine which equity series, observation frequency, or reference-peak method applies to any program, does not verify compliance with a firm’s loss-limit rule, and does not decide whether a decline reflects normal variance or execution deterioration. The trader defines the measurement rule, tracks the equity series it depends on, and applies it.

Frequently asked questions

Is a drawdown the same as a losing streak?

No. A losing streak is a count of consecutive losing trades; a drawdown is a measurement of the equity path from a reference peak. A short streak that includes one large loss can produce a deep drawdown, and a longer streak of small losses can produce a shallow one — trade count alone does not determine depth or duration.

What counts as a “large” drawdown?

There is no universal threshold — it depends on the trader’s own risk capacity, the strategy’s historical depth and duration range, and, for funded or evaluation accounts, the program’s stated limit. A depth routine for one strategy’s history can be unusual for another’s.

When is a drawdown recovered?

A drawdown returns to zero once equity returns to its reference peak; that is when the episode, and its duration, ends. A new high-water mark is a separate, later event that requires equity to exceed — not just reach — that peak. A decline can be fully recovered without yet having produced a new high.

What is the difference between drawdown and maximum drawdown?

A drawdown reading describes the distance from the reference peak at one specific moment. Maximum drawdown is the largest such reading recorded during an episode, or over a stated historical period — a historical fact about the equity series and sampling rule used, not a live reading or a guaranteed ceiling on the next decline. An episode still in progress can already be the deepest one on record before it has returned to its reference peak.

Why do prop firms and my own spreadsheet show different drawdown numbers?

They are likely using different equity series, observation frequency, or reference/threshold rules — for example, one measuring marked-to-market equity sampled continuously against a trailing loss floor, and the other measuring closed-trade equity sampled daily against a historical high-water mark. Confirm the exact method in the program’s current official documentation rather than assuming it matches your own tracking.

Costante provides educational workflow tools, not financial advice. Trading involves risk.

For the broader performance framework around operating through a drawdown, see trading performance.