A backtest that reports "32% annual return" and stops there is telling you almost nothing. The question that matters is how much risk you took to get it, and there's no single answer — risk means different things depending on what would actually hurt you. The Sharpe, Sortino, and Calmar ratios are three answers to "was this return worth the risk?", and they disagree often enough that picking the right one changes which strategy you'd choose. None of this is investment advice.
The Sharpe ratio: return per unit of total volatility
The Sharpe ratio is the default everyone reaches for. It's the strategy's excess return (return above the risk-free rate) divided by the standard deviation of its returns. Higher is better; a Sharpe above 1 is generally considered good for a retail strategy, above 2 is excellent and worth double-checking for overfitting.
Its strength is universality — it's the lingua franca of strategy comparison, so reporting it lets others benchmark you. Its weakness is that it treats all volatility as bad, including upside volatility. A strategy that occasionally spikes 15% in a good month is penalized exactly as if it had dropped 15%, which is backwards from how an investor actually experiences risk.
The Sharpe ratio's core assumption is that a 10% gain and a 10% loss are equally "risky" because both move you away from the mean. But no investor lies awake worrying about unexpected gains. That mismatch is exactly what the Sortino ratio sets out to fix.
The Sortino ratio: punishing only the downside
The Sortino ratio is the Sharpe ratio's more honest cousin. It uses the same numerator (excess return) but divides by downside deviation — the standard deviation of only the negative returns. Upside volatility no longer counts against you.
For most retail strategies, Sortino is the more meaningful number. You don't care that a strategy is "volatile" if all the volatility is in your favor; you care about how much it can lose. A strategy with big, lumpy gains and small, controlled losses will show a much better Sortino than Sharpe — and that gap is information, telling you the strategy's volatility is mostly the good kind.
The catch is comparability: fewer people report Sortino, and the choice of the minimum acceptable return (the threshold below which a return counts as "downside") affects the result, so you have to state it.
The Calmar ratio: return per unit of worst pain
The Calmar ratio takes a completely different view of risk. It divides the annualized return by the maximum drawdown — the largest peak-to-trough decline over the period. It ignores day-to-day volatility entirely and asks one question: how much did you earn relative to the worst loss you had to stomach?
This is the metric that maps most directly to whether you'd actually stay invested. A strategy with a stellar Sharpe can still have a 60% max drawdown buried in it, and almost no real person holds through a 60% drawdown without panic-selling at the bottom. Calmar surfaces that. It's especially useful for leveraged or trend-following strategies where the average looks fine but the worst case is brutal.
The Calmar ratio is only as meaningful as your test period is representative. If your backtest never includes a real crash, the maximum drawdown is artificially small and the Calmar ratio is flattering. Always check what market regimes your test actually covered before trusting it.
How to use all three
Don't pick one — report several, because the gaps between them are where the truth lives.
A practical approach: lead with Sharpe so others can benchmark you, include Sortino to show whether your volatility is mostly upside, and check Calmar to confirm the worst-case drawdown is survivable. When a strategy looks great on Sharpe but poor on Calmar, that's a warning that its smooth-looking returns hide a tail risk you'd hate to live through. When Sortino is much higher than Sharpe, that's a reassuring sign the volatility is the kind you want.
The number that matters most is the one tied to the risk you'd actually fail to tolerate. For most people building strategies they intend to hold, that's the drawdown — so don't let a beautiful Sharpe ratio talk you past a Calmar ratio that's quietly telling you the truth.
Risk-adjusted return metrics exist because raw return is a salesman's number. Learn what each ratio penalizes, report more than one, and you'll stop being fooled by strategies whose only real skill is hiding their downside.
Originally published at pickuma.com. Subscribe to the RSS or follow @pickuma.bsky.social for new reviews.
Top comments (0)