Calculator
Enter periodic returns (one per line or comma-separated). You may also attach a CSV and choose the return column.
Example data table
Sample monthly returns (%) with MAR 0.5. Use “Load Example” to insert these values.
| # | Return (%) | Downside diff (Return − MAR) | Contribution (min(0,diff)^2) |
|---|---|---|---|
| 1 | 2.50 | 2.00 | 0.0000 |
| 2 | -1.20 | -1.70 | 2.8900 |
| 3 | 3.10 | 2.60 | 0.0000 |
| 4 | 0.80 | 0.30 | 0.0000 |
| 5 | -0.60 | -1.10 | 1.2100 |
| 6 | 1.70 | 1.20 | 0.0000 |
| 7 | -2.40 | -2.90 | 8.4100 |
| 8 | 2.20 | 1.70 | 0.0000 |
| 9 | 0.50 | 0.00 | 0.0000 |
| 10 | -0.90 | -1.40 | 1.9600 |
Formula used
This calculator uses a standard downside-risk approach:
- Downside difference:
d_i = min(0, r_i − MAR) - Downside deviation:
DD = sqrt( (1/D) * Σ(d_i²) ), whereDisN,N−1, or downside-only count (based on your settings). - Sortino ratio:
Sortino = (R̄ − B) / DD, whereBis MAR or risk-free rate. - Annualized ratio (common approximation):
Sortino_annual ≈ Sortino * sqrt(PPY)
Engineering note: use consistent time units when comparing scenarios, projects, or system performance streams.
How to use this calculator
- Paste returns into the box, or upload a CSV and choose its column.
- Select whether values are percent or decimal format.
- Set MAR (your required threshold) and optional risk-free rate.
- Choose numerator mode, downside basis, and denominator style.
- Enable annualization if you want yearly comparable outputs.
- Press Calculate. Export CSV or PDF if needed.
FAQs
1) What does the Sortino ratio measure?
It measures performance relative to harmful volatility. Only returns below your chosen threshold contribute to risk, so it focuses on downside behavior rather than total variability.
2) What should I use as MAR?
Use the minimum acceptable return for your context: a hurdle rate, required project yield, or policy target. Keep MAR in the same periodicity as your input returns.
3) When should I subtract risk-free instead of MAR?
Subtract risk-free when comparing to a near-guaranteed baseline. Subtract MAR when you care about meeting a specific required threshold for a program, design, or portfolio objective.
4) Why is downside deviation sometimes zero?
If no return falls below MAR, downside differences are all zero. The denominator becomes zero, so the ratio tends toward infinity when excess return is positive.
5) What is the difference between N and N−1?
Using N treats the data as the full population of interest. Using N−1 is a sample adjustment that slightly increases the estimated downside deviation when you have limited observations.
6) Should I use arithmetic or geometric mean?
Arithmetic mean is common for average periodic return comparisons. Geometric mean reflects compounding, but it requires all returns above −100%. Choose one and stay consistent across comparisons.
7) What does annualization change?
Annualization converts periodic metrics to a yearly scale for easier comparison. The calculator scales downside deviation by sqrt(periods per year) and provides an annualized ratio approximation.
8) How do winsorization and caps help?
They reduce the impact of extreme values from sensor glitches, logging issues, or rare shocks. Use them sparingly and record the settings so others can reproduce your reported ratio.