Research
Sharpe vs Sortino vs Calmar: Choosing the Right Metric
Three risk-adjusted return metrics; three different stories. Knowing which to lead with for a given strategy and audience is the difference between honest reporting and selective reporting.
Sharpe, Sortino, and Calmar measure risk-adjusted return — but with three different definitions of risk. Each tells a different story. Choosing which to lead with for a given strategy and audience is part of the research stack.
What each penalizes
| Metric | Risk denominator | Penalizes |
|---|---|---|
| Sharpe | Total volatility | All variability — upside and downside |
| Sortino | Downside deviation | Only below-target returns |
| Calmar | Maximum drawdown | The single worst peak-to-trough loss |
When to lead with Sharpe
- Symmetric strategies where upside and downside variability are similar.
- High-frequency strategies with many independent return observations.
- Standard institutional reporting — Sharpe is the universal default.
- Comparative ranking across strategies with similar shapes.
When to lead with Sortino
- Asymmetric strategies with positive skew — trend-following, momentum. Big winners depress Sharpe but they're the point.
- Negatively-skewed strategies that you want to be honest about — options-selling, carry. Sortino doesn't let you hide the downside.
- When the target return matters — Sortino against an investor's specific target is more meaningful than against zero.
When to lead with Calmar
- Allocator-facing reporting. Allocators ask about worst-case experience, and Calmar maps directly.
- Strategies with persistent drawdown profiles — carry, vol-arb. The drawdown is the strategy's defining feature.
- Single-strategy rather than diversified portfolios. Drawdown of a single strategy is more meaningful than its volatility.
- CTA / managed-futures positioning. Calmar is the default in that universe.
Why all three matter
Each metric encodes different assumptions about what "risk" means:
- Sharpe assumes risk = variability, all variability equal.
- Sortino assumes risk = below-target outcomes only.
- Calmar assumes risk = peak-to-trough loss.
A strategy with great Sharpe and bad Calmar is volatility-friendly but drawdown-vulnerable. A strategy with great Calmar and mediocre Sharpe is drawdown-resilient but volatility-heavy. Reporting only one number hides part of the picture.
Sentivue's standard reporting
Three metrics, always all three:
- Sharpe for institutional comparability.
- Sortino for asymmetric strategies where Sharpe penalizes the wrong tail.
- Calmar for allocator-facing context on lived experience.
Plus the two raw metrics: annualized volatility and maximum drawdown. Allocators can recombine into whatever derived metric they prefer.
Practical takeaways
- There is no single best metric. Three numbers, each answering different questions.
- Leading with one metric and burying the others is selective reporting.
- An honest report shows all three plus the underlying volatility and drawdown.