Maximum Drawdown: What It Is & How to Calculate It 2026

What Is Maximum Drawdown?

Maximum drawdown (MDD) is the largest peak-to-trough decline in the value of a portfolio, fund, or strategy before a new peak is reached. It measures the worst-case cumulative loss an investor would have experienced if they bought at the highest point and sold at the lowest point during a given period.

If a fund grows from £100 to £150, then falls to £105, then recovers to £160, the maximum drawdown is the drop from £150 to £105 - a 30% decline. It doesn't matter that the fund eventually recovered. The drawdown captures the worst pain an investor actually sat through.

Maximum drawdown is expressed as a percentage and is always negative or zero. In practice, people often state the absolute value - saying "the fund's max drawdown was 30%" rather than -30%. The sign convention varies, but the meaning is the same: the strategy lost 30% from its peak before recovering.

Fund managers, allocators, and risk management teams use maximum drawdown as a primary measure of downside risk. Unlike volatility, which treats upside and downside moves equally, maximum drawdown focuses exclusively on the worst loss experience. It answers the question every investor eventually asks: "How much could I have lost at the worst possible time?"

In 2026, maximum drawdown remains one of the most widely reported risk statistics in hedge fund due diligence, strategy evaluation, and regulatory reporting. It feeds directly into performance metrics like the Calmar ratio, and allocators frequently set hard drawdown limits as part of their investment mandates.

---

The Maximum Drawdown Formula

The maximum drawdown formula compares every trough to its preceding peak across the entire return series, then takes the largest decline.

MDD = (Trough Value - Peak Value) / Peak Value

Where:

• Peak Value = the highest portfolio value reached before a subsequent decline
• Trough Value = the lowest portfolio value reached after that peak, before a new peak is established

The result is a negative number (or zero). By convention, most practitioners report the absolute value.

Worked Example

Suppose you're tracking a portfolio with the following month-end values over eight months:

Step 1: Identify the running peak at each point. The running peak is the highest value reached up to and including that month. It starts at £100,000, rises to £112,000 in February, stays there through June (since the portfolio is below its peak), then updates to £115,000 in July.

Step 2: Calculate the drawdown at each point. The drawdown is the percentage decline from the running peak. In May, the drawdown is (£89,000 - £112,000) / £112,000 = -20.5%.

Step 3: Find the maximum drawdown. The largest drawdown in the series is -20.5%, occurring in May. This is the maximum drawdown.

MDD = (£89,000 - £112,000) / £112,000 = -20.5%

The portfolio's worst peak-to-trough loss was 20.5%. An investor who entered at the February peak and held through the May trough would have lost just over a fifth of their capital before the eventual recovery.

---

Drawdown Duration and Recovery

Maximum drawdown measures the depth of the worst loss, but depth alone doesn't tell the full story. How long a portfolio stays underwater - and how long the recovery takes - matters just as much to real investors.

Key Terms

• Drawdown duration - the total time from peak to the point where the portfolio first exceeds that peak, encompassing both the decline and the recovery
• Time to trough - the period from the peak to the lowest point
• Recovery time - the period from the trough back to a new high

A 20% drawdown that recovers in two months is a very different experience from a 20% drawdown that takes three years to recover. Both have the same maximum drawdown figure, but the second scenario ties up capital for far longer, imposes greater opportunity cost, and tests investor patience more severely.

Why Duration Matters

Duration is critical for several reasons. Investors paying management fees during a drawdown are losing money in two ways - the capital loss itself and the ongoing fees on a diminished balance. Leveraged strategies face margin calls and forced liquidation risk during extended drawdowns, regardless of whether the strategy would eventually recover. And psychologically, long drawdowns cause the most redemptions. Academic research consistently shows that drawdown duration, not just depth, predicts when investors capitulate.

Some practitioners supplement maximum drawdown with metrics that account for duration. The Ulcer Index, for example, measures the root mean square of drawdowns over a period, penalising both depth and persistence. The pain ratio divides return by the mean drawdown, capturing the typical underwater experience rather than just the worst case.

When evaluating a strategy, always look at the drawdown chart alongside the maximum drawdown number. A strategy with a 15% max drawdown might look acceptable until you see that it spent 70% of its life below previous highs.

---

Calculating Maximum Drawdown in Python

Here's a practical implementation for computing maximum drawdown from a return series using NumPy and pandas. This is the standard approach used in quantitative strategy research and production risk systems.

import numpy as np
import pandas as pd


def maximum_drawdown(returns: pd.Series) -> dict:
    """
    Calculate maximum drawdown and related statistics from a return series.

    Parameters
    ----------
    returns : pd.Series
        Period returns (e.g. daily or monthly).

    Returns
    -------
    dict
        Dictionary with max drawdown, peak date, trough date, and recovery date.
    """
    cumulative = (1 + returns).cumprod()
    running_max = cumulative.cummax()
    drawdown_series = (cumulative - running_max) / running_max

    # Maximum drawdown value and trough location
    max_dd = drawdown_series.min()
    trough_date = drawdown_series.idxmin()

    # Peak is the date of the running max just before the trough
    peak_date = cumulative.loc[:trough_date].idxmax()

    # Recovery date is the first date after the trough where cumulative
    # exceeds the previous peak value
    peak_value = cumulative.loc[peak_date]
    post_trough = cumulative.loc[trough_date:]
    recovery_dates = post_trough[post_trough >= peak_value].index

    recovery_date = recovery_dates[0] if len(recovery_dates) > 0 else None

    return {
        "max_drawdown": max_dd,
        "peak_date": peak_date,
        "trough_date": trough_date,
        "recovery_date": recovery_date,
        "drawdown_series": drawdown_series,
    }


# --- Example usage ---
np.random.seed(42)
dates = pd.bdate_range("2023-01-02", periods=756)  # ~3 years of daily data

daily_returns = pd.Series(
    np.random.normal(0.0003, 0.012, len(dates)),
    index=dates,
)

# Inject a drawdown period for a realistic example
daily_returns.iloc[200:240] -= 0.006

result = maximum_drawdown(daily_returns)

print(f"Maximum Drawdown:  {result['max_drawdown']:.2%}")
print(f"Peak Date:         {result['peak_date'].strftime('%Y-%m-%d')}")
print(f"Trough Date:       {result['trough_date'].strftime('%Y-%m-%d')}")

if result["recovery_date"]:
    duration = (result["recovery_date"] - result["peak_date"]).days
    print(f"Recovery Date:     {result['recovery_date'].strftime('%Y-%m-%d')}")
    print(f"Drawdown Duration: {duration} days")
else:
    print("Recovery Date:     Not yet recovered")

A few things to note about this implementation:

• The equity curve is built from compounded returns, not from raw prices. This ensures the drawdown calculation is correct regardless of whether you're working with daily, weekly, or monthly data.
• The running maximum tracks the high-water mark at each point in time using pandas' cummax() method. The drawdown at any given date is simply how far the equity curve has fallen relative to that high-water mark.
• Recovery detection scans forward from the trough to find the first date where the cumulative value exceeds the peak. If no such date exists, the strategy hasn't recovered yet - a common situation in live monitoring.
• Set the random seed for reproducibility. In production, you'd replace the simulated returns with actual strategy returns from your data pipeline.

For a quick, minimal version that returns only the maximum drawdown percentage:

def max_dd_simple(returns: pd.Series) -> float:
    """Return the maximum drawdown as a positive decimal."""
    cumulative = (1 + returns).cumprod()
    running_max = cumulative.cummax()
    drawdowns = (cumulative - running_max) / running_max
    return abs(drawdowns.min())

---

What Is a Good Maximum Drawdown?

What counts as an acceptable maximum drawdown depends entirely on the strategy type, the leverage employed, and the asset class traded. There is no universal threshold - a 30% drawdown that would be unremarkable for a long-only equity fund would be catastrophic for a market-neutral strategy.

The table below shows typical maximum drawdown ranges for different strategy types based on historical performance data. These are rough guides, not hard rules.

Interpreting These Ranges

A long-only equity investor who tracks a broad index should expect drawdowns of 30% or more at some point in any 10-year period. The S&P 500 fell roughly 55% during 2007-2009 and about 34% during the March 2020 COVID sell-off. For passive strategies, the question isn't whether a large drawdown will happen but when.

Active strategies are judged more harshly. A long/short equity fund that suffers a 40% drawdown has failed at the hedging that justifies its fees. An allocator would reasonably ask why they're paying 2-and-20 for an experience worse than an index tracker.

For quantitative strategies, the expected maximum drawdown is closely tied to the strategy's frequency and capacity. A high-frequency market maker turning over its book every few minutes has structurally small drawdowns because positions are never held long enough to accumulate large losses. A monthly rebalancing factor portfolio has much more time between adjustments for losses to compound.

The Drawdown Budget

Many institutional allocators set explicit drawdown budgets. A pension fund might allocate to a manager with the constraint that maximum drawdown must not exceed 15%. If the manager breaches that threshold, the allocation is reviewed or redeemed. Understanding these constraints is essential when designing strategies for institutional capital. A strategy that maximises Sharpe ratio but occasionally produces 25% drawdowns may be un-investable for clients whose mandate caps drawdowns at 15%.

---

Maximum Drawdown in Portfolio Evaluation

Maximum drawdown is central to how allocators, fund-of-funds managers, and institutional investors evaluate and compare investment strategies.

Due Diligence and Manager Selection

During due diligence, allocators examine a manager's maximum drawdown alongside the circumstances that caused it. A 20% drawdown during a broad market crisis (where peers also suffered) is very different from a 20% drawdown during calm markets caused by a concentrated position going wrong. The drawdown number alone isn't enough - the story behind it matters.

Allocators typically want to see:

• The maximum drawdown over the full track record
• Rolling 12-month and 36-month maximum drawdowns
• How the drawdown compares to peers running similar strategies
• How long the recovery took
• Whether the manager changed their process in response

The Calmar Ratio Connection

The Calmar ratio divides annualised return by maximum drawdown, giving a direct measure of how much return the strategy generates per unit of its worst loss. A Calmar ratio above 1.0 means the strategy earns back its worst drawdown within a year on average. Above 3.0 is considered excellent.

Calmar Ratio = Annualised Return / |Maximum Drawdown|

For a strategy returning 12% annually with a maximum drawdown of 8%, the Calmar ratio is 1.5. For a strategy returning 25% with a 30% drawdown, the Calmar ratio is 0.83 - despite the higher absolute return, the risk-adjusted picture is worse.

Drawdown Constraints in Portfolio Construction

When building multi-manager portfolios, allocators often set a portfolio-level drawdown budget and allocate it across managers. If the total acceptable drawdown for the portfolio is 10%, and a given manager historically has a 20% max drawdown, that manager might receive a smaller allocation than one with a 10% max drawdown - even if the first manager has higher expected returns. The goal is to keep the portfolio's aggregate worst-case loss within bounds.

This is why risk management frameworks in institutional settings treat maximum drawdown as a constraint rather than just a metric. It's not simply a number to report after the fact - it's a limit that shapes position sizing, leverage, and manager selection in real time.

---

Limitations of Maximum Drawdown

Maximum drawdown is a useful metric, but relying on it exclusively gives an incomplete picture of risk. It has several well-known weaknesses.

It's Backward-Looking

Maximum drawdown tells you the worst loss that did happen, not the worst loss that could happen. A strategy with a 10% maximum drawdown over three years of benign markets may be taking risks that could produce a 40% drawdown in a different environment. The historical figure depends entirely on what market conditions occurred during the measurement period. A strategy that hasn't been tested by a genuine crisis will show a flatteringly low maximum drawdown.

It's a Single Data Point

The entire metric rests on one event - the worst peak-to-trough decline. A strategy that experienced one 20% drawdown and otherwise had smooth returns has the same maximum drawdown as a strategy that experienced nine 18% drawdowns and one 20% drawdown. The two have very different risk profiles, but maximum drawdown can't distinguish between them. It captures the extreme, not the pattern.

It Doesn't Capture Frequency

Maximum drawdown ignores how often drawdowns occur. A strategy that produces a 15% drawdown once in five years is far more comfortable to hold than one that produces 12-14% drawdowns every six months. Both might report similar maximum drawdowns, but the investor experience is radically different. Metrics like the Ulcer Index or average drawdown provide better insight into the typical drawdown experience.

It's Regime Dependent

Drawdown behaviour changes across market regimes. A momentum strategy that performed well through trending markets might show a small maximum drawdown, but that figure tells you nothing about what would happen during a regime shift to mean-reverting markets. Extrapolating past drawdowns into a different market environment is one of the most common errors in strategy evaluation.

It Grows With Time

Maximum drawdown can only increase or stay the same as the observation window lengthens. A strategy with a 10-year track record will almost certainly show a larger maximum drawdown than the same strategy evaluated over three years - simply because there's been more time for a bad draw to occur. This makes direct comparisons between strategies with different track record lengths problematic.

---

Maximum Drawdown vs Other Risk Metrics

Maximum drawdown is one of several metrics used to measure investment risk. Each captures a different dimension of loss, and no single metric tells the whole story.

MDD vs VaR

Value at Risk estimates the loss threshold at a specific confidence level over a specific time horizon - for example, "95% of the time, daily losses won't exceed £500,000." Maximum drawdown looks at the cumulative worst case across the full period. VaR is a probability statement about single-period losses; maximum drawdown is a realised measure of multi-period cumulative loss. A strategy can have a modest VaR but a large maximum drawdown if small daily losses compound over weeks or months into a sustained decline.

MDD vs Volatility

Volatility measures the average dispersion of returns. It treats a 2% up day the same as a 2% down day. Maximum drawdown cares only about the downside - specifically, the worst sustained downside. A strategy with high volatility (large swings in both directions) might have a small maximum drawdown if the down swings are always brief and quickly reversed. Conversely, a low-volatility strategy that slowly bleeds value over months can have a surprisingly large maximum drawdown despite appearing calm on a day-to-day basis.

MDD vs Expected Shortfall

Expected Shortfall (CVaR) measures the average loss in the worst X% of cases. It captures tail severity better than VaR but, like VaR, focuses on single-period losses. Maximum drawdown captures cumulative multi-period losses. The two metrics answer different questions: Expected Shortfall tells you about the average bad day, while maximum drawdown tells you about the worst bad stretch.

Using Them Together

In practice, sophisticated allocators and risk teams use all of these metrics in combination. Maximum drawdown shows the worst historical experience. VaR provides daily risk bounds. Volatility feeds into position sizing and portfolio optimisation. Expected Shortfall captures tail severity. The Sortino ratio measures return per unit of downside volatility. The Calmar ratio measures return per unit of maximum drawdown. Each adds a dimension that the others miss.

---

Frequently Asked Questions

What is the difference between drawdown and maximum drawdown?

A drawdown is any decline from a peak to a subsequent trough in portfolio value. At any point in time, the current drawdown is how far the portfolio sits below its highest-ever value. Maximum drawdown is the single largest of all drawdowns observed over the measurement period - the worst peak-to-trough loss. A portfolio might experience dozens of drawdowns over its lifetime, each of varying depth and duration. Maximum drawdown picks out the deepest one. For example, if a portfolio had drawdowns of -5%, -12%, -8%, and -18%, the maximum drawdown would be -18%.

How do you calculate maximum drawdown from returns?

To calculate maximum drawdown from a return series, first build the cumulative equity curve by compounding the returns: multiply (1 + r) for each period return r. Then compute the running maximum of the equity curve at each point - this is the high-water mark. The drawdown at each point is the percentage decline from the running maximum: (current value - running max) / running max. The maximum drawdown is the most negative value in this drawdown series. In Python, you can do this in three lines using pandas: compute the cumulative product with (1 + returns).cumprod(), the running max with .cummax(), and the drawdown series by dividing the difference by the running max.

What is a good maximum drawdown for a hedge fund?

A good maximum drawdown for a hedge fund depends heavily on strategy type. Market-neutral and statistical arbitrage strategies typically aim for maximum drawdowns below 10-15%. Long/short equity funds generally consider anything above 25-30% to be unacceptable. Trend-following and CTA strategies typically experience maximum drawdowns of 15-25% during range-bound markets. As a rough benchmark, most institutional allocators become uncomfortable when a hedge fund's maximum drawdown exceeds its annualised return - that is, when the Calmar ratio drops below 1.0. A fund returning 12% annually with a maximum drawdown of 8% (Calmar ratio of 1.5) would be viewed favourably by most allocators.

Does maximum drawdown account for recovery time?

No. Maximum drawdown measures only the depth of the worst peak-to-trough decline, not how long the portfolio spent underwater or how long the recovery took. A 25% drawdown that recovers in two months and a 25% drawdown that takes three years to recover produce the same maximum drawdown figure. This is one of the metric's key limitations. For a fuller picture, pair maximum drawdown with drawdown duration (time from peak to recovery) and consider metrics like the Ulcer Index, which penalises both the depth and persistence of drawdowns. Many allocators now report both maximum drawdown and maximum drawdown duration as standard practice.

Can maximum drawdown be used to compare different strategies?

Maximum drawdown is useful for comparing strategies within the same category but misleading when comparing across different strategy types. A maximum drawdown of 20% means something very different for a long-only equity fund (which is expected to experience large drawdowns in bear markets) than for a market-neutral strategy (which should rarely see drawdowns that large). When comparing strategies, normalise by expected drawdown for the strategy type or use risk-adjusted ratios like the Calmar ratio. Also be cautious about comparing strategies with different track record lengths - a strategy with ten years of history will almost certainly show a larger maximum drawdown than one with two years, simply because more time means more opportunities for a bad draw to occur.