Calmar Ratio: Comparing Strategies by Return Divided by Maximum Drawdown

Returns alone cannot tell you whether a strategy earning 30% annually is better than one earning 60%. The picture changes if the account halved on its way to that 60%. The Calmar ratio puts that pain in the denominator to compare both with the same measure.

What Is the Calmar Ratio?

The Calmar ratio divides compound annual growth rate (CAGR) by maximum drawdown (MDD). Terry Young introduced it in 1991 through his newsletter CALMAR, California Managed Account Reports. Originally used to rank futures fund managers, it now commonly compares strategies and automated trading bots.

The formula is straightforward:

Calmar ratio = CAGR (%) ÷ Maximum drawdown (%)

It asks: How much annual return did I earn for each 1% of the worst drawdown experienced? A Calmar of 2.0 means an annual 2% return per 1% maximum drawdown; 0.5 means only 0.5% return per 1% drawdown.

The denominator, maximum drawdown, is the largest fall from a peak in the equity curve. Correct MDD calculation is essential. An incorrect denominator makes the whole ratio incorrect.

Calculation Example

Measure MDD from the peak, not starting capital. A rise from $10,000 to $13,800 is still +38%, but falling to $13,800 after reaching $18,400 creates a 25% drawdown. That 25% is the loss experienced by someone holding the account.

Different Calmar Ratios Behind Similar Returns

Calmar becomes useful when distinguishing strategies by the risk behind their returns.

Whether you can actually withstand an account nearly halving is a separate question. Abandoning the strategy during its 48% drawdown leaves the eventual 72% return as a number on paper. Drawdown psychology is a common reason for abandoning strategies.

How It Differs from Sharpe and Sortino

All three measure return relative to risk, but define risk differently.

The Sharpe ratio uses return standard deviation, treating both upward and downward fluctuations as risk. Large up days also count as volatility, so strategies with frequent surges can have lower Sharpe ratios.

The Sortino ratio uses only downside volatility, addressing that issue. However, it still measures average downside fluctuations rather than the single largest fall.

Calmar uses the worst actual drawdown instead of variance. This is intuitive because investors fear how far the account has fallen more than a standard deviation. Its critical weakness is that the denominator represents one event, discussed below.

These are complementary tools. In practice, examine them alongside profit factor. A high Calmar with a profit factor around 1.1 may indicate a sample that luckily avoided a large drawdown.

Leverage Usually Leaves Calmar Nearly Unchanged

This is a practical reason Calmar is useful for strategy comparisons. Doubling exposure roughly doubles both returns and drawdowns, so the numerator and denominator grow together and preserve the ratio.

Leverage cannot turn a low-Calmar strategy into a good one. It scales returns and drawdowns together, while fees and funding can make the result worse. Choosing leverage is a position-sizing decision, not a way to improve Calmar.

Three Calmar Traps

① The denominator contains one observation: MDD is a single event across the entire period. A strategy that luckily avoids the worst conditions appears strong. Run it one more month and a large drawdown may halve the ratio. One event can determine the entire metric.

② Short periods inflate it: A bot running for 3 months has had fewer opportunities for a major drawdown. A smaller MDD creates a larger Calmar. This explains the original 36-month convention; avoid comparing samples shorter than 12 months. Even long backtests can artificially suppress drawdown through overfitting, so a ratio without walk-forward validation is only a reference.

③ It ignores recovery time: A strategy recovering a 25% drawdown in two months can have the same Calmar as one taking fourteen months, despite very different investor experiences. Always report maximum drawdown duration with Calmar.

Practical Use

Use rolling values, not one number: Instead of one full-period value, plot rolling Calmar using a 12-month window advanced monthly. Remaining consistently above 1 indicates stability. A spike in one interval that lifts the average suggests a strategy suited only to that environment.

Compare matching periods and costs: A's 2023–2025 result is not comparable with B's 2024–2026 result. Also align fee and slippage treatment. Mixing net and gross returns compares numbers measured differently.

Use conservative reference levels: In practice, below 1 generally indicates inadequate compensation for drawdown; 1–3 is usable; above 3 warrants checking the sample and duration first. A value above 5 from a short backtest usually reflects a short sample rather than exceptional performance.

It describes the past, not the future: Calmar summarizes a historical equity curve and does not cap future drawdowns. To examine possible losses, Monte Carlo simulation can reorder trades and inspect the distribution. Actual MDD usually exceeds backtest MDD.

Three Key Points

① Calmar = CAGR ÷ MDD, measuring annual return per 1% of worst drawdown.
② Leverage scales numerator and denominator together, leaving the ratio nearly unchanged. It cannot rescue a low-Calmar strategy.
③ The denominator is one observation and short periods inflate the ratio. Use at least 12 months, rolling calculations, and recovery duration together.

Notice

The capital, returns, and fees are hypothetical calculation examples, not observations or forecasts of a particular strategy. Performance metrics summarize historical data and guarantee neither future returns nor drawdown limits. Leveraged trading can lose all principal. Investment decisions and responsibility are yours.