By Keel Research Team · Updated May 13, 2026
Max drawdown is the largest peak-to-trough percentage decline a trading strategy or asset has experienced. It's the single most honest risk metric — what you would have actually lost from peak — and a far better predictor of psychological survivability than volatility.
Max drawdown is the worst-case loss measurement. If a portfolio went $100 → $130 → $80, the max drawdown is (130 − 80) / 130 = 38%. The peak is $130; the trough is $80; the decline from peak is what max DD reports.
It's the single most important risk metric in trading for one practical reason: it's the only one that captures the path-dependent experience of actually holding a position. Volatility says nothing about whether your losses came in one painful crash or were spread evenly. Max DD says exactly that — how deep did you go, and how long did it take to climb back.
Walking through the math:
max_dd = max((peak − trough) / peak) over all rolling peak/trough pairsThe implementation is a single pass through the equity curve. Track the running peak. At each bar, compute the current drawdown as (peak − value) / peak. The maximum observed value is the max DD.
Three derived metrics matter alongside the headline number:
Sharpe ratio and Sortino ratio capture risk-adjusted return on average. Max drawdown captures the worst-case experience. The two can diverge sharply.
Consider two strategies that both have Sharpe 1.5 over a 5-year sample. Strategy A had its losses spread evenly — small 5% drawdowns once per quarter, never deeper. Strategy B was profitable for four years then took a single 40% drawdown in year 5 that took 18 months to recover. Same Sharpe. Very different lived experience.
The Calmar ratio (return / max DD) explicitly captures this — it punishes strategies for deep drawdowns regardless of how good the average-case is. Many professional traders use Calmar as the primary metric for strategy selection precisely because it weights tail risk appropriately.
For framing whether a strategy's max DD is "normal" or "extreme," here are well-known historical drawdowns:
A trading strategy with a 20-30% max DD on crypto sits in the "normal but uncomfortable" zone — comparable to a traditional bear market in equities. Above 50% means the strategy took crypto-bear-cycle risk. Above 70% is in line with holding spot crypto through 2018 or 2022 — survivable for true believers, ruinous for leverage.
Three things tighten max DD without sacrificing too much expected return:
The trade-off: each constraint shaves some upside in strong regimes. The right balance depends on personal tolerance for path pain. Compute max DD on your historical backtest, ask yourself whether you would actually have held through that drawdown, and adjust risk if the honest answer is no.
The max drawdown calculator takes either an equity series or a returns series. It returns the max DD, peak-to-trough duration, recovery time, and a historical-context reference table showing how your number compares to BTC, ETH, S&P 500, Nasdaq, and 60/40 across major market cycles. Use it to verify backtest results, to sanity-check live performance, or to frame whether a deep DD on a strategy is in line with the asset class.
Keel is a Strategy OS for AI-assisted systematic trading on Hyperliquid. Backtest, optimize, and run live strategies across single-stock perps, indices, and crypto majors — realistic fees, slippage, and funding modeled.
Free to start — connect a Hyperliquid wallet when you’re ready to go live.
Max drawdown is the largest peak-to-trough percentage decline an equity curve has experienced, measured at the worst point relative to the prior peak.
Volatility (standard deviation) treats upside and downside the same — a strategy with big winners and small losers can have high volatility but be very survivable. Max drawdown only measures the painful side — how much you actually lost from peak. Two strategies can have identical Sharpe ratios with very different drawdown experiences; the one with smaller max DD is far more practical to actually hold through.
Depends on asset class. For institutional-grade strategies on liquid markets, single-digit DDs (5-10%) are exceptional and rare. 15-25% is normal for trend-following on crypto; 30-40% is aggressive (carry strategies through regime flips often hit here); above 50% usually means the strategy took asymmetric risk that didn't pay. For comparison, holding spot BTC through 2022 was a -77% drawdown.
The math: a 50% drawdown requires a 100% return to recover. A 20% drawdown needs 25%. Recovery time scales nonlinearly with depth — the deeper the DD, the longer the climb back. For trend strategies on crypto, recoveries from 20% DDs typically take 3-9 months; deeper DDs (40%+) can take 1-2 years or never fully recover.
Max DD is the single worst peak-to-trough decline ever observed. Average drawdown is the mean depth across all drawdown periods. Max DD captures tail risk (how bad can it get); average DD captures typical pain (what you should expect on most cycles). Both matter — a strategy with low average DD but extreme max DD is hiding tail risk; one with high average DD but bounded max DD is just choppy.
Walk the equity curve once. Track the running peak. At each point, compute current drawdown as (peak − current) / peak. The maximum value observed is the max DD. Paste your equity values or returns into the max-drawdown calculator linked below — it returns max DD, duration, recovery time, and a historical-context table showing how your number compares to major-market drawdowns.
Compute max DD on your own equity or returns series, with historical-context reference table.
Risk-adjusted return ratios — the on-average risk metrics that complement max DD.
Risk-based sizing — the first line of defense against deep drawdowns.