Sortino Ratio Calculator – Downside Risk-Adjusted Return

What Is the Sortino Ratio?

Two strategies post the same 12% annual return. Strategy A delivered it through a steady drift upward, with the worst month down 1.5%. Strategy B got there with the same average but two terrifying months of -8% and -10% offset by a few jubilant +15% spikes. By Sharpe ratio, the standard yardstick of risk-adjusted return, both look broadly comparable because Sharpe penalizes total volatility — including the upside spikes. To most investors that is plainly wrong: nobody loses sleep over a +15% month. The Sortino ratio, introduced by Frank Sortino and Robert van der Meer in 1991, fixes that asymmetry by replacing standard deviation with downside deviation — only volatility below your goal is treated as risk.[1]

The formula is simple in spirit: Sortino = (Return − MAR) / Downside Deviation. The numerator measures how much you beat your "Minimum Acceptable Return" — the threshold you would be unwilling to fall below. That can be the risk-free rate, zero (capital preservation), inflation, or a personal goal like a 6% liability rate for a pension. The denominator is the downside deviation: take only the periods when returns fell below your MAR, square the shortfalls, average across the full sample, and take the square root. The result is a unit-of-shortfall-risk measure rather than a unit-of-total-volatility measure. Two strategies with identical Sharpe ratios can have very different Sortino ratios when their distributions are skewed.[2, 10]

The mental model is simple: standard deviation treats every wiggle as risk, including good wiggles. Sortino only counts the wiggles that hurt. If your strategy has lottery-like upside (option-buying, momentum, trend-following), Sortino flatters you while Sharpe punishes the upside dispersion. If your strategy has lottery-like downside (option-selling, illiquid yield, leverage), Sortino punishes you more harshly than Sharpe because the rare blow-ups dominate the downside-deviation calculation. The premise that downside hurts more than upside helps is empirically robust: Brown, Imai, Vieider, and Camerer's 2024 meta-analysis of 166 studies pegs the loss-aversion coefficient λ in the 1.25–1.45 range — investors weigh a $1 loss roughly 30% more heavily than an equivalent $1 gain.[11, 28]

Interpreting Sortino Values: From Poor to Exceptional

There is no regulatory standard for what counts as a "good" Sortino ratio, but conventional industry practice — used by hedge fund analysts at Morningstar, allocators at large pension funds, and CFA Institute education — clusters around the following bands. Below 0.5 is poor: the strategy barely compensates investors for the downside risk taken. Between 0.5 and 1.0 is sub-optimal: acceptable but not differentiated. Between 1.0 and 2.0 is good: meaningfully positive risk-adjusted return. Between 2.0 and 3.0 is excellent: institutional-grade performance. Above 3.0 is exceptional and rare; sustained Sortinos that high typically indicate either a brilliant strategy or, more commonly, return smoothing, illiquidity, or short-history luck.[12, 9]

Avoid the temptation to compare absolute Sortino numbers across different MARs. A strategy with Sortino 2.0 against MAR 0% may have Sortino 0.5 against MAR 5%. The numerator collapses as the bar rises. Disclose your MAR alongside the ratio, every time. The Global Investment Performance Standards (GIPS) explicitly require this disclosure for any downside risk metric reported to clients.[9]

Sortino vs Sharpe: When the Two Disagree

For perfectly symmetric return distributions — what statisticians call zero skewness — the relationship between Sortino and Sharpe is mechanical: Sortino ≈ Sharpe × √2. Half of the variance comes from the upside, half from the downside; Sortino discards the upside half, so its denominator is smaller by a factor of √2 and the ratio is larger by the same factor. If your data is roughly normal — broadly diversified equity returns, for instance — knowing the Sharpe lets you guess the Sortino within a tight band. The deeper question is which side of the distribution to weight more heavily; behavioral finance settles it, with Brown, Imai, Vieider, and Camerer (2024) showing that the empirical loss-aversion premium is robustly above one — losses simply matter more.[1, 28]

The interesting cases are the asymmetric ones. Long-volatility strategies — like option-buying, trend-following CTAs, or barbell portfolios — have positive skew: many small losses in normal markets, occasional large gains in crises. Sortino loves these because the upside spikes that drive Sharpe down do not enter its denominator. Conversely, short-volatility strategies — option-selling, carry trades, illiquid credit, leveraged short-vol ETFs — have negative skew: many small wins, rare catastrophic losses. Sortino punishes them more than Sharpe does, because those rare losses dominate the downside-deviation calculation. Rollinger and Hoffman (2013) document this gap empirically across hedge fund styles, finding that Sortino flips the rank order against Sharpe for roughly a third of fund pairs.[10]

In practice, sophisticated allocators rarely look at one in isolation. They report both, alongside Treynor, Calmar, Information Ratio, and drawdown statistics. The pair is more informative than either alone: agreement between Sharpe and Sortino is reassurance that the result is not driven by skew artifacts. A wide gap is a flag — either the strategy genuinely has asymmetric returns (which can be a feature, not a bug) or the data is too short, too smoothed, or too illiquid for either ratio to be trusted.[19, 18]

Choosing Your MAR: Risk-Free, Inflation, Goal, or Zero?

The Minimum Acceptable Return is the most consequential single choice in computing Sortino. Four conventions dominate practice, each anchored in a different framing of what risk means.

First, the **risk-free rate** — academic convention. As of April 2026, the 3-month T-bill yields roughly 3.68% (FRED DTB3), with the Federal Funds rate in the 3.50%–3.75% target range and the Fed H.15 release publishing daily constant-maturity yields. Using the risk-free rate makes Sortino directly comparable to Sharpe and to academic CAPM derivatives. It answers: "did this beat what I could have gotten for free?"[16, 14, 15]

Second, **zero** — capital preservation framing. Common in family-office and high-net-worth contexts where the implicit liability is "do not lose money." MAR = 0 makes the Sortino ratio answer: "how often, and how badly, did I post negative returns relative to my upside?" It is the most conservative and the most psychologically aligned choice for retail investors who have already saved enough and are now in protect-mode.[8]

Third, **inflation** — real-return framing. The MAR is the recent CPI rate, currently around 2.5%. Below MAR means the portfolio lost real purchasing power. This framing is popular for endowments, sovereign wealth funds, and retirement portfolios with long horizons, where the goal is preservation of real capital rather than nominal capital. Pension funds occasionally combine inflation with an actuarial spread (e.g., CPI + 4%) as their MAR.

Fourth, a **custom goal rate** — liability-driven framing. A pension plan with a 7% actuarial assumption uses 7%. A retiree with a 4.5% safe withdrawal target uses 4.5%. A growth-mode investor benchmarking against the S&P 500 long-run real return uses 6.5%. This is the most personally meaningful framing because the Sortino ratio now answers exactly the question that matters: "am I beating my actual goal, and when I miss, by how much?" The PMPT literature, from Sortino (2009) onward, treats this Desired Target Return as the conceptually correct MAR. For equity-heavy mandates, a forward-looking floor can be derived from the implied equity risk premium: Damodaran's 2025 ERP study pegs the U.S. implied premium at 4.33% as of January 1, 2025, modestly above the 1960–2024 mean of 4.25%, suggesting that a MAR around 8% — the risk-free rate plus a forward ERP — is a defensible 2026 benchmark for equity-heavy portfolios.[4, 22]

Downside Deviation: The N vs n_below Debate

There is a small but consequential disagreement in how downside deviation is computed. The original Sortino & Price (1994) form sums the squared shortfalls of below-MAR returns and divides by N — the full sample size. A common error in retail tools and even some published papers is to divide by n_below — the count of below-MAR observations only. The two diverge wildly: a strategy with 1 below-MAR return out of 36 has a downside deviation 6× too large under the n_below method, making the Sortino ratio 6× too small.[2, 10]

The N convention is correct. Conceptually, downside deviation is a population statistic of the squared-negative-deviation random variable; the proper estimator divides by N regardless of how many observations exceed the threshold. Mathematically, dividing by n_below would mean the metric is undefined whenever n_below = 0 — which is silly because a strategy that has never gone below MAR is precisely the strategy you want to celebrate, not the one for which the metric should explode. Rollinger and Hoffman (2013) spell this out and warn investors against tools that get it wrong. Our calculator uses the N convention.[10]

The intellectual lineage of downside risk starts with Harry Markowitz, who in his 1959 book on Portfolio Selection explicitly argued that semivariance — the variance of below-mean returns only — was a "more plausible" risk measure than full variance. He stuck with full variance only because the math was easier. Peter Fishburn extended this in 1977 with the Lower Partial Moment (LPM) framework, and Bawa & Lindenberg derived a CAPM under LPM in the same year. Sortino & van der Meer (1991) pulled this academic strand into a usable practitioner ratio. The whole tradition rests on the empirical fact that investors are loss-averse, not volatility-averse — a finding hammered home by prospect theory.[17, 6, 7]

Post-Modern Portfolio Theory and the History of Downside Risk

The phrase Post-Modern Portfolio Theory (PMPT) was coined by Brian Rom and Kathleen Ferguson in 1993. It denotes the family of portfolio-construction techniques that replace standard deviation with downside risk measures and asymmetric utility functions. The intellectual ancestor goes further back: A. D. Roy proposed the Safety First criterion in his 1952 Econometrica paper — minimize the probability of a return below a disaster level — published the same year as Markowitz's mean-variance framework but in the same journal issue. Roy argued that the typical investor cares first about the floor, not the average. The two papers planted opposing flags; modern portfolio theory took Markowitz's, and PMPT is essentially Roy's flag picked back up.[5, 8]

Sortino himself worked at the Pension Research Institute at San Francisco State University. His 1991 collaboration with van der Meer was titled simply "Downside Risk" and ran 5 pages in the Journal of Portfolio Management. The 1994 follow-up with Lee Price refined the formula and introduced the Fouse index, named after pension actuary William Fouse. In 2001 Sortino co-edited with Stephen Satchell the influential Managing Downside Risk in Financial Markets, and in 2009 published The Sortino Framework for Constructing Portfolios, which introduced the Desired Target Return™ as the canonical MAR.[1, 2, 3, 4]

Sortino in the 2026 Rate Cycle

When the risk-free rate sits near zero — as it did for most of the 2010s — the Sortino MAR question barely matters because zero, risk-free, and inflation are all clustered together. The 2022–2024 hiking cycle changed that. As of April 2026, the Federal Reserve has held the Fed Funds target at 3.50%–3.75% since the December 2025 cut, with the 3-month T-bill at roughly 3.68% and the 10-year Treasury in the 4.2%–4.3% range. The MAR you choose now significantly changes which strategies look acceptable.[15, 14, 21]

Concrete example. Take the same equity-bond 60/40 portfolio that posted a 9.5% annualized return with monthly Sortino 0.65 against MAR 6%, as in our default sample. Run the same series with MAR 0% and the Sortino jumps above 1.5 — looks fine. Run it with MAR 3.68% (current 3-month T-bill) and Sortino settles around 1.16 — still good. Run it with MAR 8% (the post-1990 S&P 500 nominal average) and Sortino drops below 0.4 — poor. The data did not change. The judgment changed entirely with the goalpost.

The forward outlook reinforces the case for a thoughtful MAR. Vanguard's December 2025 Economic and Market Outlook for 2026 projects U.S. GDP growth of about 2.25%, core inflation easing toward 2.6%, and a Fed Funds path landing near 3.5% by year-end — implying a risk-free rate that hovers in the mid-3s rather than reverting to the post-2008 zero floor. At the same time, Damodaran's 2025 ERP study reports a U.S. implied equity risk premium of 4.33% as of January 1, 2025 — almost exactly its 1960–2024 average, suggesting equities are neither obviously cheap nor obviously expensive on a risk-premium basis. Together these point to a defensible 2026 equity MAR in the 7.5%–8.5% band: roughly the risk-free rate plus the implied ERP.[24, 22, 23]

The lesson: as risk-free rates rise, every risk-bearing strategy must reach higher to look "good." A 60/40 that was excellent in 2010 may be merely adequate in 2026. The Sortino ratio surfaces this honestly — Sharpe does too, but Sortino is more sensitive in skewed distributions. Schwab's investor education recommends recomputing risk-adjusted ratios annually whenever the risk-free rate has moved more than 100 basis points.[13]

Limitations of the Sortino Ratio

The Sortino ratio inherits most of Sharpe's limitations and adds a few of its own. First, sample-size sensitivity. Downside deviation is computed only from below-MAR observations, so its statistical reliability depends on having a meaningful number of them. CFA Institute guidance recommends at least 36–60 monthly observations. With fewer below-MAR points, the metric is dominated by individual outliers; remove one observation and the answer can change by 30%.[9]

Second, the MAR-comparison problem. Two funds reporting Sortino ratios using different MARs are not comparable; the metric is meaningful only relative to a fixed goal. This makes Sortino harder to use for cross-strategy benchmarking than Sharpe (which always uses the risk-free rate by convention). Always disclose the MAR.

Third, return manipulation. Goetzmann, Ingersoll, Spiegel, and Welch (2007) showed that virtually any reward-to-risk ratio — Sharpe, Sortino, Information Ratio — can be artificially inflated by managers who write out-of-the-money options or smooth illiquid marks. Sortino is no exception. The fix is to require independent valuation of illiquid positions, audit reported returns, and look at multiple ratios in conjunction with skewness and kurtosis.[18]

Fourth, the DD = 0 case. If every observation in your sample is at or above MAR, downside deviation is zero and Sortino is mathematically undefined (division by zero). Some tools render it as +∞, which is meaningless. Our calculator treats this as "undefined" and surfaces it explicitly. The honest interpretation: the strategy has had no material drawdown against your goal in this window — congratulate yourself, but the metric cannot rank further upside.

Two modern responses are worth knowing about. On the academic side, Kroll and Marchioni (2024) introduced Sortino(γ), a modified ratio that adjusts the MAR threshold so the resulting ranking is consistent with first- and second-order stochastic dominance — addressing the otherwise awkward fact that standard Sortino can rank a stochastically dominated portfolio above the dominating one. On the regulatory side, advisers communicating Sortino numbers to U.S. clients fall under SEC Rule 206(4)-1; the April 2024 Marketing Rule risk alert from the SEC Division of Examinations cites lack of fair-and-balanced treatment of material risks, omission of relevant time periods, and inadequate disclosure of inputs as common deficiencies — all squarely applicable to risk-adjusted ratios that hinge on a chosen MAR and lookback window.[27, 25]

Sortino vs Treynor, Calmar, and Information Ratio

Sortino is one tool in a larger family of risk-adjusted return measures, each isolating a different facet of risk. Knowing when to deploy each is part of being a careful evaluator.

The **Treynor ratio** divides excess return by beta — the systematic-risk component of total volatility. It answers "how much excess return per unit of market exposure?" Useful for comparing equity strategies that already hold significant index exposure, and for ranking long-only mutual funds against their beta. Less useful for hedge funds whose beta is unstable or near zero.

The **Calmar ratio** divides annualized return by maximum drawdown — the worst peak-to-trough decline. It directly answers the question every retail investor actually cares about: "what is the worst loss this strategy has put me through, and how big is the reward in compensation?" Calmar is hugely popular in CTA and trend-following analysis, and is the most stress-test-aligned ratio of the four. It is, however, a single-point statistic — one bad month can dominate the denominator forever.

The **Information ratio** divides active return (portfolio − benchmark) by tracking error (standard deviation of active return). It answers "how much excess return per unit of bet against the index?" — the canonical metric for evaluating active managers paid to deviate from benchmarks. Sortino, by contrast, judges absolute risk-adjusted return relative to a goal, not relative to a benchmark.

The pragmatic choice: report all four when ranking strategies. Sharpe and Sortino agree on the broad picture; Treynor isolates market risk; Calmar emphasizes worst-case experience; Information Ratio quantifies skill against a benchmark. Each catches a flavor of risk the others miss.

Sortino in Real Portfolios

Across the institutional landscape, Sortino has carved out a stable role: the second-most-reported risk-adjusted return metric after Sharpe, and the canonical metric for hedge fund performance attribution. Morningstar Direct computes Sortino on a trailing 36-month window using the monthly risk-free rate as MAR; this surface in the Risk & Rating attributes of nearly every fund profile.[12]

Hedge fund managers report Sortino prominently because their strategies are systematically punished by Sharpe. Long/short equity, global macro, market neutral, and trend-following styles routinely produce skewed return distributions; their Sortino numbers are typically 30–50% higher than their Sharpe numbers, which materially affects allocator preference. Industry rule of thumb: Sortino > 1 acceptable, > 2 excellent, > 3 exceptional. The Medallion Fund's reported Sortino is meaningfully higher than its already-extraordinary Sharpe — a hint that even legendary track records benefit from the framing.

Pension funds use Sortino with a custom MAR equal to their actuarial assumed return — typically 6.5% to 7.5% in the U.S. context. The metric becomes a direct readout of how often, and how badly, the fund missed its required return. The 2008 and 2020 drawdowns appear starkly in such Sortino numbers, even when the underlying Sharpe looked superficially recovered.

Retail investors benefit most from Sortino with MAR = 0 (capital preservation) or MAR = inflation (real return). The first is honest about the asymmetry of pain — losing 10% in a month hurts more than gaining 10% feels good. The second is honest about purchasing power — beating zero is irrelevant if inflation is 3% and your portfolio earned 2%. In an era when annual T-bill yields again exceed inflation, both framings can lead to similar judgments, but they ask different questions. The 2025 CFA Institute Research Foundation study by Pham, Cui, and Ruthbah on the 60/40 portfolio across markets and decades reinforces the case for downside-aware metrics: they find that diversifying into alternative assets often raised drawdowns and reduced Sharpe efficiency relative to the textbook 60/40, with non-U.S. markets such as Japan exhibiting deeper drawdowns that an absolute risk-adjusted measure like Sortino would have flagged earlier than Sharpe alone.[23]

Frequently Asked Questions

Common questions investors and analysts raise about computing and using the Sortino ratio: