The Calmar ratio (named for a newsletter tradition, not a physics constant) summarizes return per unit of worst pain. A common definition:
Calmar ≈ annualized return / |max drawdown|
Use the absolute value of max drawdown so the ratio stays positive when drawdown is reported as a negative percent.
Why it exists
Sharpe penalizes all volatility. Investors often care specifically about peak-to-trough loss because that is what triggers abandonship, margin stress, or mandate violations.
Calmar ties headline return directly to the deepest hole in the equity curve.
Example intuition
| Strategy | Annualized return | Max drawdown | Calmar (approx.) |
|---|---|---|---|
| A | 12% | 15% | 0.8 |
| B | 10% | 8% | 1.25 |
Strategy B earns less but experienced a shallower worst decline, so Calmar can favor B even when raw return favors A.
Calmar vs Sharpe vs Sortino
| Metric | Reward | Risk penalty |
|---|---|---|
| Sharpe | Excess return | Total volatility |
| Sortino | Excess return | Downside volatility only |
| Calmar | Annualized return | Single worst drawdown |
None replaces the others. Calmar ignores path shape between peaks and ignores how long recovery took.
When Calmar helps
- Comparing trend or macro strategies where blowups matter more than daily noise.
- Screening backtests for return relative to tail loss before deeper diligence.
- Talking to stakeholders who ask "what was the worst drop?" first.
Pitfalls
- Short samples - One lucky period without a crash inflates Calmar.
- Single event dominance - One crisis defines max drawdown forever until a worse one occurs.
- Leverage - Can boost return and drawdown together; Calmar may look stable until it is not.
- Different start dates change max drawdown (choosing a backtest start).
Always inspect the equity curve and recovery time, not only the ratio.
Using Calmar responsibly
Report Calmar with the same window as return and drawdown. Note whether returns are net of fees. Stress-test with walk-forward or Monte Carlo when possible.
Calmar is a compact summary of "was the return worth the worst slump?" - useful, but never sufficient on its own.
