What Is the Sortino Ratio?
The Sortino ratio measures return per unit of downside volatility: excess return divided by the standard deviation of negative returns only. It differs from Sharpe by not penalising a strategy for large gains, which matters for anything with an asymmetric payoff.
What is the formula?
Sortino = (portfolio return − target return) ÷ downside deviation, where the downside deviation is the standard deviation computed over only the periods that fell below the target. The target is often zero or the risk-free rate.[1]
Worked example. Twelve monthly returns average 1.8% against a 0% target. Four of them are negative: −2.0%, −3.5%, −1.0% and −4.0%. The downside deviation over those four, measured against the target, is sqrt((4 + 12.25 + 1 + 16) ÷ 12) = sqrt(2.77) = 1.66%. Monthly Sortino = 1.8 ÷ 1.66 = 1.08, which annualises by sqrt(12) to about 3.7.
Note the denominator: the sum of squared shortfalls is divided by the total number of periods, not by the number of negative ones. Dividing by four instead of twelve would give 2.88% and a Sortino of 0.62 — the two conventions differ by a lot, so always ask which was used.
Why not just use Sharpe?
Because Sharpe treats a +9% month and a −9% month as equally risky. For a strategy with a long right tail — trend following, long options, anything that loses small and occasionally wins big — that is actively misleading, because the volatility being penalised is the source of the returns.[2]
The reverse case matters more. A premium-selling strategy shows tiny, regular gains and rare large losses. Its Sharpe looks excellent right up until the loss arrives, and its Sortino is the number that degrades first, because every one of those rare losses lands in the denominator.
What can it not tell you?
Path. Two strategies with identical Sortino can have very different worst drawdowns, because the ratio measures dispersion, not the depth of the hole.
Sample size. Twelve monthly observations is a small sample. A Sortino computed on a single good year is a description of that year.
Tail risk. A strategy that has not yet had its bad month has an outstanding Sortino by construction.
How does this connect to a track record?
The same statistical caution applies here as to Sharpe, and harder. Lo showed a Sharpe ratio is an estimate with its own standard error, and that annualising by the square root of 12 assumes independence that real return series often lack.[3] Sortino uses a subset of the observations — only the downside ones — so it is computed from fewer data points and its error bars are wider still.
Sortino is a ratio you compute over a return series, so it is only as good as the series it is computed on. A series assembled after the fact by the person being measured is where the trouble starts, not in the formula. A metric is a summary of a record, so it inherits every weakness of that record. Computed from trades selected after the fact, it is a number about the selection. kappi commits each trade before it resolves and publishes it on a Merkle-anchored log, so PnL, RME, correlation to SPX, mean R:R and trade count over 30, 100 and 200-day windows are computed over everything that was committed, losses included. $15/month to keep a record; reading one is free.
Sources
- Sortino & Price, 'Performance Measurement in a Downside Risk Framework', Journal of Investing 3(3), 1994, 59–64 read 2026-08-16
- Sharpe, 'The Sharpe Ratio', Journal of Portfolio Management 21(1), 1994, 49–58 read 2026-08-16
- Lo, 'The Statistics of Sharpe Ratios', Financial Analysts Journal 58(4), 2002, 36–52 read 2026-08-16
Frequently asked questions
What is a good Sortino ratio?
Above 1 is usually considered acceptable and above 2 strong, but the value depends heavily on the target return, the period length and the sample. Always check what target was used and over how many observations.
How is Sortino different from Sharpe?
Sharpe divides by the standard deviation of all returns; Sortino divides by the deviation of negative returns only. Sortino therefore does not penalise a strategy for its large gains.
How do you calculate downside deviation?
Take the shortfall below the target for each period, square it, sum the squares, divide by the total number of periods, and take the square root. Dividing by only the losing periods gives a much larger number.
Can the Sortino ratio be misleading?
Yes. A strategy that sells premium shows small regular gains and rare large losses, so its Sortino looks excellent until the first bad period arrives. It also says nothing about maximum drawdown.