π Sortino Ratio Calculator
Return per unit of downside risk, not total volatility.
Sortino ratio
Estimates for planning only, not financial advice.
The Sortino ratio answers the question the Sharpe ratio fumbles: how much return do you get per unit of bad volatility? A return series of 8, 12, -3, 5, and -1 percent against a 3 percent target has a Sortino ratio of 0.37. Upside jumps do not punish the score; only the drops below your target do.
How to Use This Calculator
Paste a series of periodic returns as percentages, separated by commas, then set your target return (often called the minimum acceptable return) and the risk-free rate. Use consistent periods: all monthly or all annual, never mixed. More observations give a steadier ratio; five data points is a toy, thirty is a starting point.
Enter returns as percentages: 8, 12, -3, 5, -1. Monthly fund returns or annual portfolio returns both work if you stay consistent.
Set the target return. This is your hurdle: the return you needed, often 0 percent, inflation, or your required rate.
Read the ratio plus the downside deviation. Compare strategies over the same periods and targets, never across mismatched ones.
The Formula
Downside deviation only counts the shortfalls: each period's return minus the target, floored at zero, squared, averaged, and square-rooted. Periods above the target contribute nothing. The numerator is excess return over the risk-free rate, same as Sharpe. The difference is entirely in the denominator, and it matters: a strategy with big upside spikes and small drawdowns scores far better on Sortino than on Sharpe, which is arguably the correct verdict.
Worked Example
Returns: 8, 12, -3, 5, -1 percent. Target 3 percent, risk-free 3 percent. Mean return: (8 + 12 - 3 + 5 - 1) / 5 = 4.2 percent. Shortfalls vs target: only -3 (shortfall 6) and -1 (shortfall 4); the rest floor at zero. Downside deviation: sqrt((36 + 16 + 0 + 0 + 0) / 5) = sqrt(10.4) = 3.22 percent. Sortino: (4.2 - 3) / 3.22 = 0.37. The same series scores a Sharpe of about 0.21 because Sharpe penalizes the +12 spike as "risk." That penalty is exactly what Sortino refuses to apply.
Sortino vs Sharpe: When Each Wins
Sharpe treats all volatility as bad, which punishes positively skewed strategies like option selling or trend following that have occasional large gains. Sortino only punishes downside, matching how investors actually experience risk: nobody complains about upside volatility. Use Sortino when comparing strategies with asymmetric payoffs, and Sharpe when you want the classic textbook number or when return distributions are roughly symmetric.
| Sortino ratio | Rough reading |
|---|---|
| Below 0 | Not beating the risk-free rate, reject |
| 0 to 1 | Suboptimal, common for single assets |
| 1 to 2 | Good, worth a closer look |
| Above 2 | Excellent, or too few observations |
Those bands are rules of thumb, not laws. A Sortino of 3 on twelve monthly observations is mostly noise; on ten years of data it is a genuine edge. Always ask how many periods produced the number.
Choosing the Target Return
The target, or minimum acceptable return, is where judgment enters. Common choices: 0 percent (any loss is downside), the inflation rate (preserving purchasing power is the hurdle), or your required return for the goal (7 percent for a retirement plan). A higher target mechanically lowers the ratio, so never compare two Sortinos computed with different targets. My default for fund screening is the risk-free rate as both target and rf, which keeps the comparison clean. For goal-based planning, pair this with our retirement savings calculator to set the required return honestly, and check raw growth with the compound interest calculator.
What Sortino Misses
Sortino fixes Sharpe's upside penalty but keeps its other blind spots. It says nothing about the shape of the downside: a strategy with frequent small losses and one with rare catastrophic losses can share a Sortino. It ignores sequence risk, the order of returns, which decides whether a retiree survives. And like every backward-looking ratio, it assumes the past distribution repeats. Use it to compare, not to predict. If the downside deviation comes from three observations, the ratio is a rumor, not a statistic.
Using Sortino in Fund Selection
Put Sortino to work as a screen, not a verdict. A practical pass: require a 3-year Sortino above 1.5, then rank survivors within their category, since absolute Sortino varies by asset class. Pair it with maximum drawdown: a fund with a great Sortino and a 40 percent drawdown got there with smooth small losses and one terrible stretch, which the ratio's averaging can hide. Finally, check the upside capture: two funds can share a Sortino while one participates far more in rallies. Sortino answers "how painful was the ride down"; upside capture answers "did I miss the ride up." You want both answers.
Estimates for planning only, not financial advice.
Frequently Asked Questions
What is the Sortino ratio?
The Sortino ratio measures return per unit of downside risk: (mean return minus risk-free rate) divided by downside deviation. Unlike the Sharpe ratio, it only penalizes volatility below your target return, so upside spikes do not hurt the score. Higher is better, and 1 to 2 is generally considered good.
How do you calculate downside deviation?
For each period, take min(0, return minus target), square it, average across all periods, and take the square root. Periods above the target contribute zero. For returns of 8, 12, -3, 5, -1 against a 3 percent target, downside deviation is 3.22 percent.
What is the difference between Sortino and Sharpe ratio?
Sharpe divides excess return by total standard deviation, punishing upside volatility too. Sortino divides by downside deviation only, punishing just the drops below target. For positively skewed strategies like trend following, Sortino gives the fairer verdict; Sharpe makes them look worse than investors experience them.
What is a good Sortino ratio?
Roughly: below 0 is unacceptable, 0 to 1 is suboptimal, 1 to 2 is good, above 2 is excellent. But context rules: a 2.5 on ten years of monthly data means something real, while a 2.5 on one lucky year means almost nothing. Always check the observation count behind the number.
What should I use as the target return?
Common choices are 0 percent, the inflation rate, or your required return for a goal. Using the risk-free rate as both target and rf keeps comparisons clean. Never compare two Sortino ratios computed with different targets, since a higher target mechanically lowers the score.
Can the Sortino ratio be negative?
Yes, when the mean return is below the risk-free rate. A negative Sortino means the strategy did not even beat cash, so there is no reason to hold it for return purposes. It can still have a role as a hedge, but the ratio itself is telling you to look elsewhere.
Why is my Sortino ratio undefined?
When no observation falls below the target, downside deviation is zero and the ratio divides by zero. That is genuinely good news, it means the target was never missed, but the statistic cannot be computed. Add more history or lower the target to get a defined number.
How many data points do I need?
More is better. Thirty-plus monthly observations (2.5 years) is a reasonable minimum for screening; five annual returns is a toy. Short samples make the ratio jump around with one bad month, which is why professional fact sheets show 3, 5, and 10-year versions side by side.
Does Sortino work for crypto?
It computes fine, but crypto's extreme skew and fat tails make any single ratio shaky. Sortino is arguably better than Sharpe for crypto since it ignores the upside explosions, yet one crash month can dominate the downside deviation. Use longer histories and treat the number as descriptive, not predictive.
What are the limitations of the Sortino ratio?
It ignores the shape of downside (many small losses vs one crash look alike), says nothing about return sequence, and is backward-looking. It also depends heavily on the chosen target. Use it to compare strategies over identical periods and targets, alongside drawdown and absolute return figures.