In a previous guide, we covered the Sharpe ratio as a fundamental risk-adjusted performance metric for crypto bot evaluation. The Sharpe ratio has one significant flaw: it treats all return volatility as bad, including large positive returns. A strategy that sometimes generates large winning months is penalized by Sharpe ratio the same as a strategy that sometimes generates large losing months — even though these two situations are fundamentally different from a trader's perspective. The Sortino ratio, developed by Frank Sortino, fixes this by only penalizing downside volatility — volatility that is below a target return threshold.
For crypto bot trading, where upside "volatility" (large winning months during strong trends) is actually desirable, the Sortino ratio provides a more accurate risk-adjusted picture than Sharpe. This guide covers Sortino ratio calculation, interpretation, comparison with Sharpe, target values, and how to use it in DennTech strategy evaluation. For the Sharpe ratio foundation, see our Sharpe ratio guide.
Sortino Ratio Formula
$$SR_{Sortino} = rac{R_p - R_T}{\sigma_{downside}}$$Where:
- R_p = Portfolio return (annualized)
- R_T = Target return (often 0 or risk-free rate)
- σ_downside = Downside deviation — standard deviation of returns that fall below the target return, annualized
Key difference from Sharpe: the denominator uses only negative return deviations from the target. Returns above the target are completely excluded from the denominator calculation — they do not penalize the ratio at all.
How to Calculate Downside Deviation
- For each period's return, calculate the deviation below the target:
min(return - target, 0) - Square each of these negative deviations (positive deviations contribute zero)
- Average the squared values
- Take the square root → this is the periodic downside deviation
- Annualize by multiplying by √(periods per year)
Example with daily returns and target of 0%:
Returns: +1.2%, -0.8%, +2.1%, -0.3%, +0.9%, -1.5% Below-target deviations: 0, -0.8%, 0, -0.3%, 0, -1.5% Squared: 0, 0.000064, 0, 0.000009, 0, 0.000225 Average: (0.000064 + 0.000009 + 0.000225) / 6 = 0.0000497 Downside dev (daily) = sqrt(0.0000497) = 0.00705 = 0.705% Annualized downside dev = 0.705% × sqrt(365) = 13.47%
Sortino Ratio Interpretation
| Sortino Ratio | Interpretation |
|---|---|
| Below 0 | Strategy loses money — unacceptable |
| 0 to 1.0 | Poor downside risk-adjusted return |
| 1.0 to 2.0 | Acceptable to good performance |
| 2.0 to 4.0 | Strong downside-adjusted performance |
| Above 4.0 | Excellent (or potential backtest overfitting) |
Because Sortino only penalizes downside volatility, Sortino ratios are generally higher than Sharpe ratios for the same strategy — typically 1.5–2× higher for strategies with favorable win profiles. Be aware of this when comparing Sortino ratios across different sources that might report either metric.
When Sortino and Sharpe Diverge — And Why It Matters
The divergence between Sortino and Sharpe ratio is most pronounced for strategies with asymmetric return distributions — where large positive returns are more common or larger than large negative returns. Two scenarios:
- Trend-following strategies during strong bull markets: These generate some very large positive months when trends are strong. Sharpe penalizes these large wins; Sortino does not. A MACD trend strategy on BTC during a bull phase may have Sharpe of 1.2 but Sortino of 2.4 — accurately reflecting that the "volatility" is mostly upside. See our MACD guide.
- Mean reversion strategies in volatile bear markets: These generate consistent small wins but may have large losses during trend breaks. Sharpe and Sortino diverge less here because the volatility is more symmetric. See our RSI guide for context.
When Sortino is significantly higher than Sharpe, it indicates your strategy's positive volatility is greater than its negative volatility — a desirable characteristic. When Sortino and Sharpe are similar, return volatility is symmetric — the strategy wins and loses with similar magnitude variability.
Using Sortino in DennTech Strategy Evaluation
DennTech's performance dashboard shows both Sharpe and Sortino ratios alongside MDD, profit factor, and win rate. When evaluating strategies, follow this hierarchy:
- Is the strategy profitable? (Total return > 0, profit factor > 1.0)
- Is the drawdown manageable? (MDD < 20–25% for most strategies)
- What is the Sortino ratio? (>1.5 is good, >2.0 is strong)
- Is Sortino significantly higher than Sharpe? (Indicates favorable upside asymmetry)
- How does Sortino hold up across the out-of-sample test period from our backtesting guide?
A strategy with Sortino > 2.0, MDD < 15%, and profit factor > 1.5 represents an excellent risk profile for live deployment with appropriate position sizing from our position sizing guide.
The Calmar Ratio: Another Alternative
The Calmar ratio divides annual return by maximum drawdown — a simpler metric that directly answers "how much return do I get per unit of my worst drawdown?"
Calmar Ratio = Annual Return / Maximum Drawdown
Example: 30% annual return, 12% MDD → Calmar = 2.5. A Calmar above 1.0 means the annual return exceeds the maximum drawdown — a minimum quality threshold for most bot strategies. Above 2.0 is strong performance.
Frequently Asked Questions
- Which ratio should I use — Sharpe or Sortino?
- For crypto bot evaluation, Sortino is generally more meaningful because it correctly treats large winning months as desirable (not as "risk"). However, both ratios provide complementary information. Use Sortino as your primary risk-adjusted metric, and compare the Sharpe-to-Sortino ratio to understand your strategy's return asymmetry. A ratio of Sortino/Sharpe > 1.5 suggests your strategy has meaningfully more upside volatility than downside volatility — a good sign. See our Sharpe ratio guide for the Sharpe component.
- What target return (R_T) should I use in the Sortino formula?
- For simplicity, use 0 (zero) as the target return — this means any negative return is treated as downside risk. More sophisticated traders use the annualized risk-free rate (currently ~4–5% for USD) as the target, which means only returns below the risk-free rate count as downside. DennTech's Sortino calculation uses 0 as the default target for simplicity and consistency. You can check the documentation for the exact calculation parameters used.
- My strategy has a high Sortino but high MDD — should I trust the Sortino?
- A high Sortino with a high MDD is a warning sign — it likely means the strategy has very large winning periods but also very large drawdown periods. The Sortino is high because the large wins dominate the upside, but the large drawdowns are still dangerous to your account. Always evaluate Sortino alongside MDD, not in isolation. A strategy with Sortino 3.0 but 45% MDD is not appropriate for most bot traders — the drawdown would be psychologically and financially difficult to sustain. See our MDD guide for context. View live metrics at the live demo or start at the pricing page.
For the complete risk metric framework: Sharpe ratio, Sortino ratio (this guide), MDD, risk/reward ratio, and circuit breakers.