The Sharpe Ratio

intermediate7 min read

Return per unit of risk — the most cited risk-adjusted metric, and its blind spots.

The Sharpe ratio is the most widely used risk-adjusted return measure. It asks not just “how much did you earn?” but “how much did you earn per unit of risk taken?” — calculated as excess return (above the risk-free rate) divided by the volatility (standard deviation) of returns.

The Sharpe ratio exists because raw returns are meaningless without the risk taken to get them — a 30% return from wild, white-knuckle swings is not the same achievement as a 30% return from a smooth ride. Sharpe puts them on a common scale: **return per unit of volatility.* A higher Sharpe means more reward for the bumpiness endured; it lets you fairly compare a calm strategy to a racy one. Roughly, a Sharpe under 1 is mediocre, around 1 is decent, and above 2 is excellent (and above 3, be suspicious — it may signal overfitting or hidden risk). But Sharpe has real blind spots: it treats all* volatility as bad — including big upside moves, which it unfairly penalises — and it assumes returns are roughly normal, so it underestimates strategies with rare catastrophic losses (fat tails). A naked-option seller can show a gorgeous Sharpe right up until the blow-up. So Sharpe is the best single risk-adjusted number to start with — and one you must never trust alone.

Formula (intuition) — (return − risk-free rate) ÷ volatility of returns; higher = more reward per unit of risk.
Rough scale — <1 mediocre, ~1 decent, >2 excellent, >3 be suspicious (overfit or hidden risk).
Blind spot 1 — penalises upside volatility too (which Sortino fixes, next lesson).
Blind spot 2 — assumes normal-ish returns, so it flatters fat-tailed strategies that hide rare catastrophic losses.

ExampleStrategy A returns 15% with low volatility → Sharpe ≈ 1.5. Strategy B returns 20% but with violent swings → Sharpe ≈ 0.7. B’s headline return is higher, but A delivered far more return per unit of risk — a better, more tradable engine. Sharpe revealed what the raw returns hid.

Key takeawayThe Sharpe ratio = excess return ÷ volatility — reward per unit of risk, letting you compare strategies fairly (>2 excellent, >3 suspicious). But it penalises upside volatility and underestimates fat-tail blow-up risk, so it’s the best risk-adjusted number to start with, never to trust alone.

FAQs

Can a Sharpe ratio be too good?

Yes — an unusually high Sharpe (say >3 for a discretionary or simple systematic strategy) is often a red flag for overfitting, a look-ahead bug, or hidden tail risk (like option selling that hasn’t met its catastrophe yet). Extraordinary risk-adjusted numbers deserve extraordinary scrutiny, not celebration.