Maximum drawdown explained

What maximum drawdown measures

Start with a cumulative return or account value series. At each date, track the running peak — the highest value seen so far. Drawdown at time t is how far below that peak you sit:

drawdown_t = (value_t − peak_t) / peak_t

Values are negative or zero while underwater. Maximum drawdown is the minimum (most negative) drawdown over the window — the deepest hole from any prior high. Example: account grows $100k → $140k → $91k → $120k. Peak at $140k; trough at $91k. MDD = (91 − 140) / 140 ≈ −35%. The subsequent rally to $120k does not erase MDD for that window; MDD records the worst episode, not current status.

Plot drawdown over time and you get the underwater curve — flat at zero at new highs, dipping negative during selloffs. Long flat stretches below zero signal extended recovery periods that annualized return statistics understate.

Recovery math: why drawdowns are asymmetric

Percentage losses and gains are not symmetric. To recover from drawdown d (expressed as a positive fraction, e.g. 0.35 for −35%), you need gain:

recovery gain = d / (1 − d)

A −10% drawdown needs +11.1% to break even. −20% needs +25%. −33% needs +50%. −50% needs +100% — double your remaining capital. −75% needs +300%. This compounding asymmetry is why professional risk managers cap loss limits before they chase returns: a single deep hole can dominate a decade of good years.

Recovery time

Time to recovery counts how long from the MDD trough until the account exceeds the prior peak. The S&P 500's 2008–2009 drawdown took years to recover in nominal terms; inflation-adjusted recovery took longer still. Crypto cycles can show −70% drawdowns with multi-year underwater periods. When evaluating a strategy, report both MDD depth and months underwater — a −25% drawdown that recovers in four months is a different product than −25% that lingers for three years.

MDD vs volatility metrics

Use MDD when the investor's constraint is capital preservation or psychological tolerance — endowments with spending rules, retirees, traders with daily loss limits. Pair MDD with Sortino when you also care about the shape of monthly shortfalls. Neither replaces the other.

Calmar ratio: return per unit of pain

The Calmar ratio (named for California Managed Accounts Reports) divides annualized return by the absolute value of maximum drawdown over a lookback — typically three years for hedge funds:

Calmar = annualized return / |MDD|

A strategy returning 15% per year with −30% MDD has Calmar ≈ 0.5. One returning 9% with −10% MDD also scores 0.9 — often preferable to allocators who weight survival over headline CAGR. Calmar is sensitive to window selection: ending the sample right after a crash vs right after a rally changes MDD dramatically. Always align lookback length when comparing funds and stress-test with overlapping windows.

Asset-class drawdown benchmarks (rough guide)

Historical peaks are not guarantees, but they calibrate expectations:

• U.S. large-cap equities — episodic −50%+ (Great Depression, 2008–09); typical bear markets −20% to −35%.
• Investment-grade bonds — usually shallow drawdowns; 2022 showed −15% to −20% when rates spiked fast.
• 60/40 balanced — 2008 and 2022 both stressed the classic mix; correlation spikes in crises widen joint drawdowns.
• Hedge funds / CTAs — wide dispersion; trend followers may show smaller equity-like drawdowns but long flat periods.
• Crypto (BTC, alts) — −70% to −90% cyclical drawdowns are common; recovery times measured in years.
• Leveraged ETFs — path-dependent decay; MDD can approach total loss in volatile sideways markets.

Context from bear market history and volatility regimes helps set personal loss limits before you need them emotionally.

Leverage, concentration, and liquidity

Drawdown depth scales nonlinearly with leverage. A 2× levered long position roughly doubles drawdown versus unlevered exposure to the same asset — before margin calls force liquidation at the worst price. Margin converts paper drawdown into realized loss when maintenance requirements bite.

Concentration amplifies MDD: a single-stock portfolio inherits that company's idiosyncratic crash risk. Illiquidity extends underwater periods — private funds, thin altcoins, and gated redemptions mean you cannot rebalance when the model says cut exposure. Stress-test drawdown assuming you cannot exit for 30–90 days.

Portfolio construction and drawdown budgets

Sophisticated allocators set a maximum portfolio drawdown budget — e.g. never exceed −20% on a rolling three-year basis — and size sleeves backward from that constraint. If equities might deliver −35% in a bad decade and bonds −15%, a 70/30 mix has a plausible joint stress drawdown wider than either leg alone when correlations spike. Tools:

• Volatility targeting — reduce exposure when realized vol rises.
• Stop-loss / trend filters — cut exposure after drawdown thresholds (with whipsaw tradeoff).
• Rebalancing — buy depressed assets if policy allows; see rebalancing discipline.
• Diversification — imperfect but lowers single-factor MDD when correlations stay below one.
• Cash and T-bills — dry powder limits depth at the cost of upside.

Retirees face sequence risk: the same average return with withdrawals during a −30% early retirement year can deplete principal permanently. Drawdown limits matter more than CAGR in decumulation.

Decision table: when drawdown is the right lens

Common mistakes

• Using too-short windows — three bull years hide cyclical −40% equity drawdowns.
• Ignoring recovery time — shallow but permanent underwater periods still fail spending needs.
• Backtest MDD without costs — fees, slippage, and gap risk deepen live drawdowns.
• Assuming max DD is max future DD — historical MDD is a sample, not a ceiling.
• Replacing diversification with leverage — levered calm markets explode MDD in regime shifts.
• Panic selling at trough — locks in drawdown; behavioral discipline is part of risk management.
• Comparing gross vs net returns — fund MDD should be on investor-facing net series.

Investor checklist

• Compute MDD on net returns over at least one full market cycle when possible.
• Chart the underwater curve — visualize duration, not just depth.
• Record recovery time from each major peak-to-trough episode.
• State personal or mandate max drawdown budget before allocating.
• Report Calmar alongside Sharpe and Sortino on the same window.
• Stress-test joint drawdown for multi-asset portfolios (correlation = 1 scenario).
• For leveraged or crypto sleeves, size so a historical worst-case MDD does not breach policy.
• Revisit limits after major life events (retirement start, leverage increase).

Key takeaways

• Maximum drawdown is the worst peak-to-trough percentage loss — the depth of the biggest hole.
• Recovery is asymmetric — −50% requires +100% to break even; shallow drawdowns compound better.
• Volatility metrics are not drawdown — pair MDD with Sharpe/Sortino for a fuller picture.
• Calmar links return to drawdown depth but depends heavily on sample window.
• Path matters for retirees and levered accounts — time underwater and sequence risk can dominate CAGR.

Related reading

• Sharpe ratio explained — volatility-based risk adjustment that complements but does not replace drawdown analysis
• Sortino ratio explained — downside deviation and MAR hurdles for asymmetric return profiles
• Risk management and position sizing explained — per-trade limits and portfolio heat tied to drawdown budgets
• Sequence of returns risk explained — why drawdown timing devastates retirement withdrawals