Module I·Article X·~2 min read

Sortino Ratio

Portfolio Thinking and Governance Framework

Turn this article into a podcast

Pick voices, format, length — AI generates the audio

Sortino Ratio

Sortino Ratio: only bad risk The Sortino Ratio is an improved version of Sharpe, which accounts only for negative volatility. Developed by Frank Sortino to solve the main drawback of the Sharpe — "punishment" for positive surprises.

Key Idea Investors are not afraid of upside volatility — they fear losses. If an asset grows by 50% in one month and by 3% in another, Sharpe "punishes" for the first month. Sortino does not.

Formula Sortino Ratio = (Rp - Rf) / σd σd (Downside Deviation) — volatility of only negative returns Only periods are counted when Ri

Calculation of Downside Deviation σd = √[Σ min(Ri - MAR, 0)² / n] All positive deviations are zeroed, only negative ones are counted.

Sharpe vs Sortino: practical example Two funds over a year (monthly returns):

MonthFund AFund B
January+2%+15%
February+1%-3%
March+3%+8%
April-1%-5%
May+2%+20%
June+1%+2%
MetricFund AFund B
Average return1.33%6.17%
Total volatility (σ)1.37%9.58%
Downside Deviation (σd)0.58%2.89%
Sharpe (Rf=0)0.970.64
Sortino (Rf=0)2.292.13

Conclusion: Sharpe shows that A is better than B. Sortino shows that they are almost equal — volatility of B is mostly "good" (growth).

When Sortino is especially important

  • Cryptocurrencies — extreme volatility, but often upward
  • Venture investments — most deals are unprofitable, but rare "unicorns" give 100x
  • Options strategies — selling puts has low overall volatility, but high downside risk
  • Momentum strategies — periods of strong growth alternate with sharp drops
  • Emerging Markets — asymmetric return distributions

Interpretation of Sortino values

SortinoAssessment
Loss-making strategy0 - 1
1 - 2Good ratio return/downside risk
2 - 3Excellent strategy
> 3Exceptional (check for hidden risks)

Limitations of Sortino

  • Choice of MAR — which threshold to consider "bad"? 0%? Rf? Inflation?
  • Few data — if negative periods are few, σd is unstable
  • Does not account for magnitude of losses — -1% and -30% "weigh" differently, but Sortino does not fully capture this

Combined analysis Professional CIO uses both metrics together:

SituationSharpeSortinoInterpretation
Both high↑↑Ideal strategy
Sharpe low, Sortino high↓↑Volatility is "good" — consider
Sharpe high, Sortino low↑↓Dangerous! Hidden downside risk
Both low↓↓Avoid

§ Act · what next