Maximum drawdown = (peak − trough) ÷ peak: the largest fall from a high in an equity curve, as a percent of that high. Paste an equity curve or a returns series to get the max drawdown, how long the fall took and how long it took to recover, with a reference table of historical drawdowns in BTC, ETH, SOL, the S&P 500, the Nasdaq, gold and a 60/40 portfolio.
Paste equity-curve values, comma- or space-separated. Min 2 values.
Significant: comparable to traditional-market bear cycles (the S&P 500 lost about 57% peak-to-trough in 2007–09).
| Asset / Portfolio | Cycle | Max DD | Duration |
|---|---|---|---|
| BTC | Dec 2017 → Dec 2018 | −84% | ~12mo |
| BTC | Nov 2021 → Nov 2022 | −77% | ~12mo |
| ETH | Nov 2021 → Jun 2022 | −82% | ~7mo |
| SOL | Nov 2021 → Dec 2022 | −96% | ~13mo |
| S&P 500 | Oct 2007 → Mar 2009 | −57% | ~17mo |
| Nasdaq Composite | Mar 2000 → Oct 2002 | −78% | ~31mo |
| Gold (USD spot) | Aug 2011 → Dec 2015 | −45% | ~52mo |
| 60% S&P 500 / 40% US Treasuries | Jan 2022 → Oct 2022 | −21% | ~9mo |
Sources: CoinGecko, Yahoo Finance, S&P historical data, World Gold Council. Cycles are completed market events; numbers don’t change.
Max drawdown is the largest peak-to-trough decline an equity curve has experienced, expressed as a fraction of the prior peak:
max_dd = max((peak − trough) / peak) over all rolling peak/trough pairsThe calculator walks the equity curve once, tracking the running peak. At each step it computes the current drawdown vs that peak; the maximum observed is the max DD. We also surface duration (bars from peak to trough) and recovery (bars from trough back to prior peak, or “not yet recovered”).
Why the reference table matters. A 35% max DD reads very differently for a crypto strategy than an equity portfolio. The reference numbers below, completed market-cycle drawdowns, let you frame your own result: is your strategy “within normal crypto range” or “abnormal even for crypto”? The closest historical analog is highlighted automatically.
Numbers in the reference table are sourced from CoinGecko, Yahoo Finance, S&P historical data, and World Gold Council. They’re completed historical cycles and don’t change.
Sharpe and Sortino describe risk on average; max drawdown describes the worst stretch. Two strategies can share a Sharpe ratio while one loses a little every quarter and the other runs smoothly for years and then gives back 40% in one go. The second is far harder to keep running. The Calmar ratio, annual return divided by max drawdown, puts that worst stretch into the score; the Sortino ratio calculator shows it beside the Sharpe and Sortino ratios.
Each of these gives up some return in strong markets. Look at the max drawdown of your backtest and ask whether you would really have held through it. If the honest answer is no, lower the risk.
Take the equity curve $100 → $130 → $80 → $115 → $140.
These are the usual readings for a trading strategy. They are conventions; what matters most is whether you would keep running the strategy through its worst fall.
| Max drawdown | The usual reading |
|---|---|
| below 10% | shallow, rare for a crypto strategy |
| 10% to 20% | moderate |
| 20% to 35% | normal for a trend strategy on crypto |
| 35% to 50% | aggressive |
| 50% and above | severe; most traders abandon a strategy before it recovers |
Across the 18 strategies in Keel's public library, backtested on Hyperliquid perps from Aug 15, 2024 to Jul 5, 2026, the median max drawdown is 32.4%; half of them sit between 22.9% and 40.2%. Those drawdowns are on the account value, after fees, slippage and funding. See the strategies and their backtests.
For scale, holding BTC through its Dec 2017 to Dec 2018 fall was a drawdown of 84%.
Keel is a strategy builder for Hyperliquid. Describe an idea, backtest it with fees, slippage and funding, and run it live on Hyperliquid’s crypto perps. Start free, and connect a wallet when you’re ready to trade.
Max drawdown is the largest peak-to-trough decline in an equity curve, expressed as a percentage of the prior peak. It tells you the worst you would have sat through in the past, which is a floor, not a ceiling, for what you may sit through next.
Track the highest value the curve has reached so far, measure how far each later value sits below it as a percent of that high, and keep the largest. For the curve $100 → $130 → $80 → $115 → $140, the high is $130 and the low after it is $80, so the max drawdown is ($130 − $80) ÷ $130 = 38.5%.
Volatility (standard deviation) treats upside and downside the same. Drawdown measures only the painful side: actual loss from peak. Two strategies can have the same Sharpe ratio with very different drawdown profiles; the one with lower max DD is far more survivable in practice, because path-dependent risk-of-ruin is real.
Depends on the asset class and strategy. Liquid trend-following strategies often recover in months; deep-bear crypto positions can take years (BTC 2018: ~2.5 years to break even). The reference table on this page shows recovery durations for major historical drawdowns. A strategy that never recovers from its max DD is broken, regardless of past Sharpe.
It depends on the strategy, but the usual readings are: below 10%: shallow, rare for a crypto strategy; 10% to 20%: moderate; 20% to 35%: normal for a trend strategy on crypto; 35% to 50%: aggressive; 50% and above: severe; most traders abandon a strategy before it recovers. Across the 18 strategies in Keel's public library, backtested on Hyperliquid perps from Aug 15, 2024 to Jul 5, 2026, the median max drawdown is 32.4%; half of them sit between 22.9% and 40.2%. For scale, holding BTC through its Dec 2017 to Dec 2018 fall was a drawdown of 84%.
More than you lost, because the gain is on a smaller balance. The gain needed is 1 / (1 − drawdown) − 1: a 20% drawdown needs 25% to get back to the peak, a 50% drawdown needs 100%, and a 70% drawdown needs about 233%. The deeper the drawdown, the faster the climb back grows.
Max drawdown is the single worst peak-to-trough fall. Average drawdown is the mean depth across all the drawdowns in the curve. Max drawdown shows how bad it got; average drawdown shows the usual pain. A low average with a very deep max points to hidden tail risk; a high average with a bounded max is a choppy but contained strategy.
Every Keel backtest reports max DD alongside Sharpe, total return, win rate, and duration. Open any backtest in the Keel app; the metrics card includes max drawdown by default. The /strategies/funding-carry page is a worked example: its backtest, max drawdown included, comes from the strategy library package.