Sharpe Ratio vs Sortino Ratio: What Each One Measures and When They Disagree

Want the daily routine? Get the free 15-Minute Swing Trader’s Checklist. Get it free →

Two risk-adjusted return ratios can rank the same 2 strategies in opposite order. Here is why, worked through on 12 hypothetical months.

Published: 16 October 2026 · Last updated: 16 October 2026 · By Spencer Li, CFTe

---

Strategy A made money in 11 months out of 12. Strategy B lost money in 6 of them.

Ask the Sharpe ratio which one is better, and it picks A. Ask the Sortino ratio, and it picks B by almost 3 times.

The Sharpe ratio divides your average return above a safe rate by the standard deviation of all your returns, so every swing counts as risk, whether it went up or went down. The Sortino ratio divides the same excess return by the downside deviation, which counts only the returns that fall below a target you choose.

When your returns are lopsided, the two will rank strategies differently. Read them together. The gap between them tells you the shape of your risk, and the shape is what you actually sit through.

Below are both formulas, the 2 strategies worked out line by line, a quick rule for reading the gap, and the 2 things neither ratio can see.

What is the Sharpe ratio, in plain words?

William Sharpe introduced it in 1966 to compare mutual funds. He called it the “reward-to-variability ratio”. As he wrote in 1994, “the measure has gained considerable popularity, the name has not.”

The formula:

Sharpe ratio = (average return − risk-free rate) ÷ standard deviation of returns

The top line is the reward: what you earned above a safe alternative like cash. The bottom line is the variability: how far your returns wander from their own average.

Do note that it measures the wander in both directions. A month 5% better than your average raises the standard deviation exactly as much as a month 5% worse. To the Sharpe ratio, a pleasant surprise is risk too.

In the examples below I set the risk-free rate to 0 to keep the arithmetic clean. When you compare your own numbers, use the same rate for everything you compare.

What is the Sortino ratio, and what does it change?

The Sortino ratio keeps the top line and swaps the bottom one. Instead of the standard deviation, it divides by the downside deviation: how far your returns fall short of a target you pick, often 0.

Sortino ratio = (average return − target return) ÷ target downside deviation

The idea comes from Frank Sortino’s work in the early 1980s. The clearest public walkthrough I know is a short paper by Thomas Rollinger and Scott Hoffman, “Sortino: A ‘Sharper’ Ratio”. It defines the downside deviation as the root-mean-square of each month’s shortfall below the target, “where all returns above the target return are treated as underperformance of 0.”

That last part trips people up. The good months are set to 0, but they stay in the count. The authors underline it: “we do not ‘throw away’ the 0% values.” Taking the standard deviation of the losing months alone is a common shortcut, and their paper shows why it gives the wrong number.

In practice, this means a big up month costs a strategy nothing on the Sortino ratio. Only shortfall counts as risk.

What happens when Sharpe and Sortino disagree?

Here are 2 hypothetical strategies over the same 12 months. The returns are invented to show the effect, and kept to round numbers.

Sharpe ranks A first, narrowly. Sortino ranks B first, at 2.7 times A’s figure.

The reason sits in the bottom line of each formula. B’s standard deviation is high because of its 3 big up months of 4%, 5% and 6%, and Sharpe counts those swings as risk even though they are the best thing that happened to the account all year. Sortino ignores them and sees only B’s small losses, none worse than 1.5%.

A looks calm to Sharpe because 11 of its 12 months are almost identical. Sortino sees the one month that hurt: a 6% drop that wiped out 4 months of gains.

Which one would you rather hold? Asked honestly, most traders pick B. Its worst month was 1.5%, and A’s was 4 times that. A’s pattern, small steady gains with a rare large loss, is the pattern of selling insurance, and it looks safest on paper right up to the month the big claim comes in.

[INFOGRAPHIC: Strategy A and Strategy B as 12 monthly bars side by side, with each strategy’s Sharpe and Sortino ratio printed underneath]

Good Wobble, Bad Wobble: how do you read the two together?

I call this Good Wobble, Bad Wobble. Every strategy wobbles. Good wobble is the swing above your target, and bad wobble is the swing below it. Sharpe counts both. Sortino counts only the bad.

That makes the gap between the 2 ratios a reading on the shape of your returns. There is a simple benchmark for it.

If your monthly returns are spread evenly on both sides of 0, the Sortino ratio comes out at about 1.4 times the Sharpe ratio. That 1.4 is the square root of 2, and I checked it with 200,000 simulated months. With a positive average, the multiple climbs higher still.

A ratio below 1.4 is your cue to go and look at the worst month before you trust the smooth line.

Both ratios only describe how a strategy has behaved, and what makes a strategy behave well in the first place is the routine behind every trade: a stop fixed before entry and a position sized to it.

Our free 15-Minute Swing Trader’s Checklist walks through that routine before every trade. Get the free checklist →

Why can 12 months of returns not settle the question?

Sharpe gave the test himself. In his 1994 paper he wrote: “The t-statistic will equal the Sharpe Ratio times the square root of T (the number of returns used for the calculation).”

Apply that to Strategy A. A monthly Sharpe of 0.40 times the square root of 12 gives a t-statistic of 1.40. The usual rough bar for a result being more than luck is about 2, and 1.40 does not clear it.

To reach 2 at a monthly Sharpe of 0.40, you would need about 25 months of returns. At a monthly Sharpe of 0.20, you would need 100 months, which is more than 8 years.

It is the same lesson as trade counts, where we give a system 100–200 trades before judging it. A ratio worked out from a short sample is a guess with 2 decimal places. The full argument is in How Many Trades Do You Need? The Law of Large Numbers and Sample Size.

Do note the √12 step too. Multiplying a monthly ratio by √12 assumes each month has nothing to do with the last. Sharpe flagged that returns “may be serially correlated”, and when they are, the annual figure overstates or understates the truth.

What can neither ratio see?

Both ratios are built from an average and a spread. Sharpe was open about the limit. Comparisons built only on those 2 numbers “do not take into account possible differences among portfolios in other moments”, which in plain words means the size and shape of the tails.

Put 3 things beside either ratio before you trust it.

The worst month and the max drawdown. A’s −6% month is the single number that explains the whole table above. Drawdown has its own section in The Performance Numbers That Matter.

The order of the months. Take Strategy B and move its 6 losing months together so they land in a row. Both ratios stay exactly the same, because neither cares about order. But the max drawdown goes from 1.5% to 5.4%, and a 6-month losing run is a very different thing to sit through than 6 small losses scattered across the year between your winning months.

How many trades sit behind each month. A month built on 2 trades and a month built on 30 trades count the same in both formulas. Your journal knows the difference, so keep the trading journal next to the ratios.

How do you work out both ratios from your own returns?

You need monthly returns as percentages, which a decent journal already gives you. Then:

1. Pick a target and keep it. Use 0, or the cash rate. Whatever you pick, use the same one for everything you compare.
2. Work out the Sharpe ratio. Take the average monthly return, subtract the target, then divide by the standard deviation. Sharpe’s own paper gives the spreadsheet version for 60 months of excess returns: AVERAGE(C1:C60)/STDEV(C1:C60).
3. Work out the downside deviation. For each month, take the shortfall below the target, or 0 if the month beat it. Square each one, average them over every month including the zeros, then take the square root.
4. Work out the Sortino ratio. Divide the same top line as step 2 by the downside deviation.
5. Annualise both the same way. Multiply each monthly ratio by √12, and remember the independence assumption.
6. Write the worst month and the max drawdown next to them. Then work out Sortino ÷ Sharpe and read it against 1.4.

Step 6 is the one most people skip. It is also the one that would have caught Strategy A.

A ratio is a summary of a shape. Before you trust the summary, look at the shape.

Sharpe and Sortino ratio FAQ

What is a good Sharpe ratio?

There is no cut-off that holds across every market, period and method. A ratio only means something next to another ratio worked out the same way: same period length, same risk-free rate and a similar number of months.

Is a higher Sortino ratio always better?

No. A high Sortino ratio built on very few losing months rests on very few data points. Strategy A’s entire downside deviation comes from a single month.

Can a Sharpe ratio be negative?

Yes. If your average return is below the risk-free rate, the top line of the formula is negative, and so is the ratio.

Why does my Sortino ratio differ from an online calculator’s?

Check 2 things: the target return, and whether the tool drops the above-target months from the count. Dropping them is the shortcut the Red Rock paper shows to be incorrect.

Should I use daily, weekly or monthly returns?

Monthly is the common choice for a trader’s own account. Sharpe’s paper recommends fairly short periods such as monthly, then annualising for comparison. Whatever you choose, use the same period for every strategy you compare.

Go deeper

This post sits under our guide to money and risk management. If you would like the routine that fixes your stop and size before every trade, the free 15-Minute Swing Trader’s Checklist is the place to start. Enter your email and we’ll send it over:

We use your email address to send you the checklist, then emails about trading and our courses. Unsubscribe in one click, any time. Privacy policy.

---

About the author. Spencer Li is the founder of Synapse Trading and a Certified Financial Technician (CFTe). His trade log is public: every trade since April 2024, dated, losses left in.

Educational content only, not personalised financial advice. Trading carries risk, including loss of capital. Past performance is not indicative of future results. Synapse Trading is not licensed or regulated by MAS.

Want the daily routine? Enter your email for the free 15-Minute Swing Trader's Checklist.

Free 5-page PDF, then occasional trading emails. Education only, not financial advice. Unsubscribe anytime. Kit stores your email in the US: see our privacy policy.

Want to learn the full method? The Trading Mastery Program teaches the 4 strategies I have traded since 2011. It suits beginners and experienced traders.

Rather follow the trades as they happen? Daily Trading Signals is our Telegram channel: my trade ideas with entry, stop and exit, and every trade is public.

Educational content only, not personalised financial advice. Trading carries risk, including loss of capital. Past performance is not indicative of future results. Synapse Trading is not licensed or regulated by MAS.