Sharpe Ratio vs Sortino Ratio: Comparing Portfolio Risk and Return

Sharpe ratio vs Sortino ratio is a choice about how to measure variability around a portfolio's returns. Sharpe compares average excess return with the volatility of those excess returns. Sortino compares average return above a stated target with the size and frequency of shortfalls below that target. Positive returns can count as Sortino shortfalls if they fall below a positive target.

You might see a strong ending gain on a portfolio chart and still want to know how uneven the returns were. These ratios add context, but their values depend on the measurement window, return frequency, and calculation method. Neither tells you which portfolio will perform better next.

All numbers below are hypothetical. They describe a calculation, with no current stock prices, fund results, or recommended return target. Sources checked October 28, 2026.

Sharpe ratio vs Sortino ratio at a glance

Input or question	Historical Sharpe ratio	Historical Sortino ratio
Return reference	Usually a risk-free return for each observation period	A disclosed target, also called minimum acceptable return or MAR
Numerator	Arithmetic mean of portfolio-minus-reference returns	Arithmetic mean portfolio return minus target
Denominator	Standard deviation of excess returns	Downside deviation relative to the target
Which observations affect the denominator?	Variation above and below the mean	Shortfalls below the target; other observations contribute zero under the full-sample method below
Can the ratio be negative?	Yes, when average excess return is negative	Yes, when average return falls below the target

In his 1994 explanation of the Sharpe ratio, William Sharpe defines the historical measure using average differential return and its standard deviation. For the usual risk-free comparison, calculate each period's portfolio return minus its risk-free return before taking the mean and standard deviation. If that reference return varies, using portfolio volatility alone can give a different result.

Sortino uses a target return and downside deviation. The target sets the shortfall threshold. A 1% monthly return falls short of a 2% monthly target by one percentage point even though the portfolio gained value. CFA Institute's discussion of the Sortino ratio explains the target's role and why variations in the calculation matter.

Calculate both ratios from the same monthly returns

Suppose you have four monthly returns: −2%, 1%, 3%, and 6%. Use an assumed constant monthly risk-free return of 0% for Sharpe and a monthly target of 0% for Sortino. Setting both to zero isolates the denominator difference; it does not suggest that zero is the right reference for another analysis.

The arithmetic mean monthly return is:

(−2% + 1% + 3% + 6%) ÷ 4 = 2%

That is an average of four observations. It is neither the compounded four-month gain nor a CAGR.

For Sharpe, use the sample standard deviation of the monthly excess returns. Here the excess returns equal the portfolio returns because the assumed reference is zero. Subtract the 2% mean from each return: −4, −1, 1, and 4 percentage points. Square those differences, add them, divide by 4 − 1, then take the square root:

Monthly sample standard deviation = √[(16 + 1 + 1 + 16) ÷ 3] = 3.3665 percentage points

Monthly Sharpe ratio = 2% ÷ 3.3665% ≈ 0.59

For Sortino, use the full-sample downside deviation convention: subtract the target from each return, replace positive differences with zero, square the remaining differences, average across all four observations, and take the square root.

Month	Return	Shortfall relative to 0% target	Squared shortfall, using decimal returns
1	−2%	−2%	0.0004
2	1%	0%	0
3	3%	0%	0
4	6%	0%	0

Monthly downside deviation = √(0.0004 ÷ 4) = 0.01 = 1%

Monthly Sortino ratio = (2% − 0%) ÷ 1% = 2.00

The same returns produce 0.59 and 2.00 because we measured different forms of variability. You cannot treat Sortino's larger number as proof that it describes a better portfolio. These are two measures of one sample, and four months is a teaching example, not enough evidence for a durable performance judgment.

Downside deviation is not the standard deviation of losing months

For the method above, write the downside deviation formula as:

Downside deviation = √[sum of min(return − target, 0)² ÷ number of all observations]

Keep returns and the target in the same units and frequency. If you enter 2% as 0.02, enter a 1% target as 0.01.

In our example, dividing by only the one losing month would produce 2% downside deviation and a Sortino ratio of 1.00. Dividing by all four months produces 1% and 2.00. Omitting the zero entries changes the answer.

Nor should you run a standard-deviation function on just the losing months. With one losing month, a sample standard deviation is undefined. That calculation also loses the distance from the stated target. The Red Rock Capital paper hosted by CME Group sets out the full-sample method and retains the zero entries.

A different target changes which observations count. With a 2% monthly target, our four shortfalls become −4%, −1%, 0%, and 0%. Downside deviation becomes about 2.06%. Since the mean return equals the target, the numerator and Sortino ratio are zero. Disclose the target whenever you report the ratio.

Can you compare published Sharpe and Sortino ratios?

Compare like with like: one portfolio's Sharpe with another portfolio's Sharpe, or Sortino with Sortino. Match these inputs first:

Dates and observation frequency. A monthly series and a daily series capture different movements. Use the same start and end dates.
Return basis and costs. Check price return versus total return, dividend treatment, fees, and currency. For an account with deposits or withdrawals, first use a return method that handles those cash flows.
Reference and target. Identify the risk-free series for Sharpe and the target for Sortino. Match their periods to the portfolio returns.
Statistical convention. Check sample versus population standard deviation and the downside denominator. Our Sharpe example uses sample standard deviation; our downside calculation divides by all observations.
Annualization. Our answers are monthly ratios. Do not compare them with annualized figures. Sharpe discusses the assumptions behind scaling across periods; CFA Institute warns that scaling monthly downside deviation can differ from calculating it from annual returns. Follow the provider's stated method.

A negative Sharpe ratio says that the portfolio's arithmetic average return fell below its reference over the measured periods. A negative Sortino says the average fell below its target. Avoid applying a simple ranking rule to negative ratios: holding a negative numerator constant, a larger denominator makes the ratio less negative even though measured variability increased.

If the denominator is zero, report the ratio as undefined under this method. A sample with no below-target observations does not establish that future losses are impossible. There is also no universal “good” ratio that removes the need to inspect the dates, assumptions, and return path.

Use the ratios alongside a historical portfolio chart

Sharpe and Sortino summarize periodic returns. Neither preserves their sequence. Reordering the same four monthly returns leaves our ratios unchanged, but it can change the peak-to-trough decline. Inspect portfolio drawdown alongside a risk-adjusted summary.

In Equity Replay, use Compare portfolios to examine illustrative allocations over the same available dates. Read the path, ending change, and reference line together. The app's historical chart provides context for the questions these ratios address; this guide does not add a Sharpe or Sortino calculation to the app.

Equity Replay is an educational historical simulator, not investment advice, a broker, or a record of trades. It applies entered allocation weights once at the start of the window and follows available historical close-price changes. It does not model dividends, fees, taxes, deposits, withdrawals, buy-and-hold execution, or rebalancing. Read the replay methodology before comparing its output with a fund's reported risk statistics.

Record the source and assumptions next to any ratio you calculate outside the app. A label such as “monthly, price-only, January through December, zero target, full-sample downside deviation” gives another reader enough information to question or reproduce the calculation. A bare number does not.