Judge returns by the pain, not the swings
Volatility to the upside is not risk β it is what you want. Sortino corrects Sharpeβs biggest flaw by counting only downside deviation, giving strategies with sharp rallies a fairer score. Compare it side by side with the classic measure on the Sharpe ratio calculator. Before comparing two Sortino numbers from different sources, read why the target return (MAR) setting alone can swing a Sortino ratio by 30%+ on identical trade data.
Sortino: punishing only the bad volatility
The Sharpe ratio penalizes all volatility β including the upside kind, which nobody actually minds. Sortino fixes this by dividing excess return by downside deviation only: violent up-moves don't hurt the score, violent drawdowns do. For crypto strategies with asymmetric returns (small steady losses, occasional big wins β or the reverse), Sortino describes reality better than Sharpe.
Reading the number: below 1.0, the downside swings outweigh the returns β weak. 1.0β2.0 is respectable for a directional crypto strategy. Above 2.0 on a year of daily data is genuinely good; above 3.0, check your data before believing it β survivorship, short samples, and one lucky trade produce most miracle Sortinos.
Trend-following systems are the classic Sortino outperformers: their many small losses and few huge wins destroy their Sharpe but flatter their Sortino, which correctly identifies that their volatility lives mostly on the pleasant side.