Quick comparison answer
Short answer: in the sortino ratio vs sharpe ratio debate, Sharpe divides excess return by total volatility. Sortino divides it by downside deviation only. Sharpe punishes every swing, including profitable ones. Sortino punishes only the periods below your target return. For a trader with lumpy, lopsided results, Sortino is usually the more honest of the two risk and return metrics. For a smooth, symmetric return stream, both ratios tell roughly the same story, and Sharpe is the one more people recognise.
This page is for retail traders in forex, indices, metals and crypto CFDs who track their own investment performance.
How each ratio is built
The Sharpe ratio takes your average return, subtracts a risk-free rate, and divides by the standard deviation of returns. Standard deviation counts every move away from the mean. A month that gained 8% adds as much "risk" as a month that lost 8%. That is the whole point of the ratio, and also its most consequential flaw when traders measure risk-adjusted returns.
The risk-free input should not be stale. Many desks take it from the New York Fed SOFR reference rate page, a benchmark for a risk-free short-term rate. They then scale it to the period of the data.
The Sortino ratio keeps the same numerator and swaps the denominator for downside deviation: the root mean square of only those returns that fell below a target. The target can be the risk-free rate, zero, or any minimum acceptable return. Winning periods add nothing to the denominator, so it is the cleaner measure of downside risk.
Here is the calculation on six ordinary monthly returns, +3%, +1%, -2%, +4%, -1%, +2%, with a target of zero.
- Only two months fall below target: -2% and -1%. Square them: 4 and 1. Sum: 5.
- Divide by all six periods, not just the two losers: 5 / 6 = 0.83.
- Take the square root: downside deviation is roughly 0.9%.
- Standard deviation of the same six months is roughly 2.1%, more than double.
Mean monthly return is about 1.17%. Sharpe is 1.17 / 2.1, about 0.56 per month. Sortino is 1.17 / 0.9, about 1.3. Same trades, same profit, a ratio more than twice as high. One ambiguity: some spreadsheets divide by the number of losing periods instead of all periods. Pick one convention and keep it.
Side-by-side comparison table
| Feature | Sharpe ratio | Sortino ratio |
|---|---|---|
| Numerator | Return minus risk-free rate | Return minus target (risk-free or a chosen minimum) |
| Denominator | Standard deviation of all returns | Downside deviation of below-target returns only |
| Treats big winning periods as | Risk | Neutral |
| Best fit for | Symmetric, steady return streams | Skewed streams with rare large wins |
| Sensitivity to sample size | Moderate | High, because few periods feed the denominator |
The most consequential row is the denominator. Everything else flows from the fact that Sortino throws away roughly half of the data points.
Decision criteria
Ask these three questions before you pick a ratio for a given comparison.
- Is your return distribution symmetric? A tidy bell shape of monthly results favours Sharpe. A long right tail, common in trend-following, favours Sortino.
- How many periods do you have? Under 30, Sortino rests on a handful of losing observations and turns wobbly. Sharpe degrades more gracefully.
- Does the strategy take small frequent losses for rare large wins? Sharpe marks it down for the wins. Sortino does not.
If the answers split, compute both and treat the gap as information. A large gap tells you the volatility is lopsided, which is itself worth knowing for portfolio management.
When to choose each option
Choose the Sharpe ratio when:
- Your results are unremarkable in shape: many small wins, many small losses, few outliers. Scalping and mean reversion often look like this.
- You need a number a wide audience, including fund allocators, will accept without explanation.
Choose the Sortino ratio when:
- Your equity curve is stingy with losses and generous with wins, the classic trend-follower profile.
- You are judging a strategy against a hard floor such as a daily loss limit, where only the volatility that breaks the floor matters.
- Your target return is the monthly figure needed to cover spread and commission cost, not the risk-free rate.
Bluntly: compute both, lead with Sharpe, and use Sortino to explain the difference. Showing only the flattering one invites the awkward question of what the other says.
A worked example on a 10,000 account
Two traders each run a 10,000 account for a year. Both finish with 2,000 of net profit, a 20% return. With a 4% risk-free rate, the excess return is 16 percentage points.
Trader A runs a jittery intraday approach with symmetric monthly swings. Standard deviation is 16% and downside deviation is 14%.
- Sharpe: 16 / 16 = 1.0
- Sortino: 16 / 14 = roughly 1.14
Trader B runs a slower swing system. Most months are small losses or flat, and three months carry the whole year. Standard deviation is also 16%, because those three big months push it up. Downside deviation is only 8%.
- Sharpe: 16 / 16 = 1.0
- Sortino: 16 / 8 = 2.0
Identical Sharpe, and a Sortino gap of nearly two to one. Sharpe cannot see that B's volatility was mostly upside. Yet B's path was precarious in a different sense: a threadbare run of nine losing months is what pushes traders to abandon a working system. Our piece on the trading expectancy formula shows how to check whether such a run sits within the normal range. Pair it with the risk of ruin in trading calculation to see whether the account survives a longer drought.
Tradeoffs and caveats
Neither ratio is forgiving of bad inputs, and both have failure modes that seldom get mentioned.
- Small samples make Sortino brittle. Ten losing months out of sixty gives ten data points for the denominator. One extra bad month can halve the ratio.
- Sharpe treats a volatile winning streak as a defect. That is stubbornly wrong for many investment strategies built on momentum, and it explains why Sortino exists.
- Costs are often left out. Spread, commission and swap charges make a backtest look better than the account statement. Compute every ratio on net-of-cost returns, or the comparison is needlessly flattering.
- Neither ratio measures drawdown depth. A quietly stable strategy with a superb Sortino can still carry one drawdown deep enough to breach an evaluation rule. Risk tolerance deserves its own metric.
The honest position is that these are ranking tools. They scarcely tell you whether either option is good enough to trade.
Using the ratios inside a prop-firm evaluation
A prop evaluation changes what "downside" means. On a Pivex account the daily loss limit resets each day and the total drawdown is static rather than trailing. The real constraint is the largest loss you can absorb without crossing a line. That is a downside-only problem, so Sortino maps onto it more naturally than Sharpe.
A practical routine: set the Sortino target to zero, use daily returns, and watch the downside deviation figure. If it drifts toward a quarter of the daily loss limit, a breach is getting close even when the average day is profitable. Pivex sets out its daily and total loss limits in writing before you start, and its built-in assistant tracks the limits as you trade. The guide on what is a good win rate in trading explains why a high win rate with a low Sortino is a warning rather than a comfort.
Next step: pull your last 60 trading days from your platform and compute both ratios with a zero target. If Sortino flatters and Sharpe sulks, you have a skewed system that needs room to wait for its big months. A one-step Pivex evaluation with no time limit and no minimum trading days gives a patient system that room.
Related guides: trading expectancy formula and risk of ruin in trading.
Before participating, review Pivex's current program rules and decide whether a simulated proprietary trading evaluation fits your experience and risk limits. Participation does not guarantee a reward, income, or any particular trading result.