The Sortino Ratio and Downside Deviation: Measuring the Risk That Actually Hurts in 2026
The Sortino ratio measures return per unit of downside risk. It takes the return you earned above a target you set, then divides it by a measure of how badly and how often you fell short of that target. Results at or above the target contribute nothing to the denominator. Only the shortfall counts.
That one design choice is the whole point. The Sharpe ratio, which most traders reach for first, divides excess return by total standard deviation, and standard deviation treats a large winning day as risk. Two accounts can post identical returns, and the one with occasional large green days scores worse on Sharpe than the one that grinds. If you have ever felt that a risk statistic was punishing you for a good week, that is exactly what happened.
This guide covers what the Sortino ratio is and how it is built, what downside deviation actually measures, how to choose the target return that anchors the whole calculation, where the ratio misleads, and why the metric maps onto a funded account better than the alternatives.
Key takeaways
- Separate the two kinds of surprise. The Sortino ratio scores return against downside deviation only, so upside variation stops counting against you.
- Set the target deliberately. The minimum acceptable return anchors the entire calculation, and changing it changes the answer more than most traders expect.
- Distrust a high number on a short sample. If a period contains very few results below the target, the denominator is tiny and the ratio inflates without meaning anything.
- Match the metric to the rule. Daily loss limits and drawdown allowances are one-sided constraints, so a one-sided risk measure describes your real problem better than a two-sided one.
- Use it alongside drawdown, not instead of it. The Sortino ratio says nothing about the order results arrived in, and order is what ends accounts.
What this guide covers
- What the Sortino ratio measures
- Downside deviation, step by step
- Choosing your minimum acceptable return
- Where the Sortino ratio misleads
- Why it fits a funded account
What the Sortino ratio measures
The Sortino ratio is average return above a chosen target, divided by downside deviation below that same target. Written plainly: take your average result, subtract the minimum acceptable return, then divide by a volatility figure calculated using only the results that fell short. A higher number means you produced more return for each unit of shortfall risk you took on.
CFA Institute's reference note on the measure describes it as return per unit of downside risk, calculated by subtracting the minimum acceptable return from the mean return and dividing by the semi-standard deviation. The full write-up is available as a CFA Institute paper on the Sortino ratio, and it is worth reading if you want the formal treatment.
How it differs from the Sharpe ratio
Both ratios put excess return on top and a volatility measure on the bottom. The difference is entirely in what goes into the bottom. Sharpe uses total standard deviation, which counts every deviation from the mean in either direction. Sortino uses downside deviation, which counts only deviations below the target.
| Aspect | Sharpe ratio | Sortino ratio |
|---|---|---|
| Numerator | Return above the risk-free rate | Return above a chosen target, the minimum acceptable return |
| Denominator | Total standard deviation of returns | Downside deviation, calculated only from results below the target |
| Treats a large winning day as | Risk, because it is a deviation | Nothing, because it is at or above the target |
| Best suited to | Comparing broadly symmetric return streams | Strategies with uneven or skewed results |
| Main weakness | Penalizes desirable variation | Unstable when few results fall below the target |
Both measures average outcomes and ignore the sequence in which results arrived. Neither one describes path risk.
Why skew makes the difference matter
If your results are roughly symmetric, Sharpe and Sortino will rank strategies similarly and the choice barely matters. The gap opens up when returns are skewed. A strategy that produces many small wins and occasional large losses looks better under Sharpe than it deserves, because the small wins keep total volatility low. A strategy that produces many small losses and occasional large wins looks worse under Sharpe than it deserves, for the mirror-image reason. Sortino corrects the second distortion directly and exposes the first, since the large losses land squarely in the denominator.
Downside deviation, step by step
Downside deviation measures how far your results fell below a target, on average, treating everything at or above the target as a zero. It is the engine of the Sortino ratio and the part most traders get wrong, so it is worth walking through slowly.
The four steps
- Pick a target. This is the minimum acceptable return, often shortened to MAR. It can be zero, it can be a risk-free rate, or it can be a number you have decided is the least you will accept from a period.
- Compute the shortfall for each period. For every result, subtract the target. If the answer is positive, record a zero. If it is negative, keep the negative number.
- Square, average, take the root. Square each of those values, average them across all periods including the zeros, and take the square root. That is downside deviation.
- Divide. Take your average return minus the target and divide by the downside deviation. That is the Sortino ratio.
The step people skip is including the zeros in the average. If you average only over the losing periods, you are computing something else and the number will be too large. The count in the denominator is the total number of periods, not the number of shortfalls.
The same fifteen daily results, scored two ways. Total volatility treats the green bars as risk. Downside deviation discards everything at or above the target and measures only the shortfall below it.
Divides excess return by total standard deviation. A strategy with large winning days is scored as riskier for having them.
Divides excess return over the target by downside deviation. Only results below the target contribute to the denominator.
tradefundrr.comA small worked example
Suppose a target of zero and five daily results of plus two, minus one, plus three, minus two and plus one, expressed in units of your account's daily risk. The shortfalls are zero, minus one, zero, minus two and zero. Squared, that is zero, one, zero, four and zero. The average across all five periods is one, and the square root is one. Average return is 0.6, so the Sortino ratio is 0.6 divided by 1, or 0.6. This is an illustrative example built to show the arithmetic, not a result from any account.
Choosing your minimum acceptable return
The minimum acceptable return is not a technical detail. It is the assumption that defines what counts as a bad day, and moving it moves your Sortino ratio substantially. Three common choices, each with a different meaning.
Zero
A target of zero says any losing period is a shortfall and any flat or winning period is not. This is the most common choice for active traders and the easiest to interpret. It answers a clean question: how much return am I generating per unit of losing-day volatility?
The risk-free rate
Using a short-term risk-free rate as the target says that failing to beat cash counts as falling short. This makes the Sortino ratio comparable to Sharpe and is the convention in portfolio analysis. For a day trader in a simulated account it is usually the wrong frame, because you are not comparing your session against a savings alternative.
A rule-derived target
The most useful choice for a funded trader is often a target drawn from your own account rules. If your program's daily loss limit is a known dollar figure, setting the target at zero and separately tracking how often and how deeply you approach that limit tells you something the ratio alone will not. Some traders set the target at a small positive number equal to the daily gain they need to make progress, which turns flat days into small shortfalls and produces a stricter score.
Whatever you choose, choose once and keep it. A Sortino ratio computed against a target you change between periods is not comparable to itself, and comparing your number to someone else's is meaningless unless you both used the same target and the same period length.
Where the Sortino ratio misleads
The Sortino ratio has three failure modes, and all three are common enough that you should check for them before trusting a number.
Too few shortfalls
If a sample period contains very few results below the target, the denominator becomes small and the ratio becomes very large. A trader who had one mildly negative day in a strong month can produce a spectacular Sortino ratio that says almost nothing about the strategy. The rule of thumb is simple: if you cannot count at least a couple of dozen shortfall observations, treat the number as directional at best. This is the same small-sample problem that makes Monte Carlo testing useful, since resampling exposes how fragile a result is.
It ignores sequence entirely
The Sortino ratio is built from a set of returns and does not care what order they arrived in. Ten losing days scattered across a quarter and ten losing days in a row produce an identical ratio. In a funded account those are not remotely the same event, because a run of consecutive losses is what exhausts a drawdown allowance. Pair the ratio with a look at your actual equity path and your worst consecutive stretch. Our note on drawdown in percentage versus dollars covers why the path matters more than the average.
Backtested figures carry their own limitations
A Sortino ratio computed on simulated or hypothetical results is not evidence of what an account will do. The CFTC's advertising regulations for commodity trading advisors require a specific cautionary statement alongside hypothetical performance for exactly this reason, and the reasoning behind that requirement is set out in the CFTC's release on amendments to those advertising regulations. Hypothetical results are designed with the benefit of hindsight and do not represent actual trading. That caution applies to your own backtest as much as to anyone else's marketing.
Why it fits a funded account
The Sortino ratio maps onto a funded account better than Sharpe does because a funded account's rules are one-sided. Nothing in your account terms penalizes a large green day. The daily loss limit only measures losses. The drawdown allowance only measures losses. Payout eligibility depends on meeting written conditions, not on producing smooth returns. A risk statistic that treats upside variation as a problem is answering a question your program never asked.
What to track alongside it
- Sortino ratio against a fixed target you chose in advance and do not change.
- Worst consecutive losing run in sessions and in dollars, because that is what spends the drawdown allowance.
- Distance to the daily loss limit on your worst days, not just whether you crossed it.
- Expectancy per trade, so you know whether the edge is real before you judge how smoothly it delivers. Our guide to expectancy covers the calculation.
How the loss limit interacts with the metric
Programs differ in how a daily loss limit is enforced. Some run a hard limit where the first crossing closes the account. Others run a soft limit where crossing ends the trading day only and the account continues into the next session, with no warning tally attached. Under a soft limit every crossing still consumes drawdown allowance, and the drawdown is what eventually ends things. Confirm which model applies to your program in your own written account terms, because it changes what a bad day in your Sortino sample actually costs you.
Where to build the habit
Computing a Sortino ratio takes a spreadsheet and a few weeks of honest record keeping. Building the record is the hard part, and a simulated funded account is a reasonable place to build it, because the environment is live market data with published rules, a published drawdown allowance, an 80/20 profit split where the trader keeps 80 percent, and a defined path to a payout if you follow the rules. It is not a substitute for live trading. It is where you find out whether your downside is as controlled as you assume.
Frequently asked questions
What is the Sortino ratio in simple terms?
It is return per unit of downside risk. You take your average return above a target you set, then divide by a volatility measure built only from the results that fell below that target. Anything at or above the target counts as zero risk, so upside variation does not reduce the score.
How is the Sortino ratio different from the Sharpe ratio?
Only the denominator differs. Sharpe divides by total standard deviation, which counts deviations in both directions, so a large winning day increases measured risk. Sortino divides by downside deviation, which counts only results below the target. On symmetric return streams the two rank strategies similarly; on skewed ones they diverge.
What is a good Sortino ratio for a trader?
There is no universal threshold, and any specific number quoted without a stated target return and period length is not comparable. What matters more is consistency of method: compute it the same way, against the same target, over a large enough sample, and watch whether your own number is improving or deteriorating.
What should I use as the minimum acceptable return?
Zero is the most common and clearest choice for active traders, because it makes any losing period a shortfall. A risk-free rate is the portfolio-analysis convention but is usually the wrong frame for day trading. Whatever you pick, keep it fixed, or your numbers will not be comparable across periods.
Does the Sortino ratio account for my drawdown in a funded account?
No, and this is its most important limitation. The ratio is built from a set of returns and ignores the order they arrived in, while drawdown is entirely about order. Ten losing days scattered across a quarter and ten in a row score identically, but only one of them threatens a drawdown allowance.
Can I use the Sortino ratio to pass an evaluation?
It is a diagnostic, not a target. Evaluations are passed by meeting the written conditions of the program, which usually means a profit objective without breaching the daily loss limit or the drawdown allowance. A healthy Sortino ratio suggests your losing days are contained, which correlates with staying inside those rules, but the ratio itself is not a criterion.
How many observations do I need before the number means anything?
Enough shortfall observations to make the denominator stable, which in practice means at least a couple of dozen results below your target rather than a couple of dozen results total. A period with one small losing day will produce an impressively high ratio that carries almost no information.
Should I trust a Sortino ratio from a backtest?
Treat it as a hypothesis, not a result. Hypothetical and simulated performance is designed with the benefit of hindsight and does not represent actual trading, which is why regulators require a cautionary statement alongside it. A backtested ratio tells you the idea was not obviously broken on that data, and no more than that.
What to do with this
Export your last sixty sessions, set a target of zero, and compute the Sortino ratio in a spreadsheet. It is four columns. Then compute the Sharpe ratio on the same data and look at the gap. If Sortino is much higher, your losses are more contained than your total volatility suggests. If the two are close, your results are roughly symmetric and the extra metric is not buying you much.
Then look at something neither number shows you: your longest consecutive losing run. That is the figure a daily loss limit and a drawdown allowance actually respond to. The Sortino ratio tells you how deep your bad days go. Your equity path tells you whether they arrived close enough together to matter.
Measure your downside against real rules
TradeFundrr publishes the daily loss limit, drawdown allowance, position rules and 80/20 split for every simulated program, so the risk you are measuring is the risk the account actually enforces.
Get Funded →