Quant
The Sharpe ratio, in plain English
Return alone is a vanity metric. "Up 30%" tells you nothing about how much white-knuckle risk you took to get there — a calm, steady climb and a heart-attack rollercoaster can end at the same place. The Sharpe ratio is the number that separates them: how much return you earned per unit of risk.
The idea in one line
Take your return above the risk-free rate (what you'd earn doing nothing), and divide it by your volatility — how much the returns bounce around:
Sharpe = (return − risk-free) ÷ volatility
Higher is better. Two ways to raise it: make more return for the same bumpiness, or make the same return with less bumpiness. That second one is why a smooth curve beats a jagged one even when they finish level.
What counts as good?
As a rough field guide (annualised):
- Below 1 — the returns don't clearly justify the risk.
- Around 1 — solid.
- 2 or above — very good; be suspicious and double-check it's real.
- 3+ — excellent, and a strong hint to look for a bug or overfit before you believe it.
Don't ask "how much did it make?" Ask "how much did it make for how much stress?" That ratio — return over risk — is the Sharpe.
The fine print
- It punishes all volatility, including the upside kind — a strategy that occasionally jumps up gets dinged too. (The Sortino ratio fixes this by only counting downside moves.)
- It assumes returns are well-behaved. Strategies with rare, catastrophic losses can show a flattering Sharpe right up until the day they don't.
- It's comparable, not absolute — most useful for ranking strategies tested the same way over the same period.
Used with those caveats, it's the single most useful number for answering the question that actually matters: not "did it win?", but "was it worth it?"