Tail Ratio and Gain-to-Pain Ratio

The Sharpe ratio boils a return stream down to one number, but it treats a good surprise and a bad surprise as symmetric — real strategies rarely are. The tail ratio and the gain-to-pain ratio are two quick, distribution-shape checks that ask a more direct question: when this strategy has a bad day, how bad is it compared to a good day, and does the accumulated upside outweigh the accumulated pain?

The tail ratio compares the size of extreme positive returns to extreme negative returns, typically the 95th percentile of daily returns divided by the absolute value of the 5th percentile. A tail ratio above 1 means the best days are bigger than the worst days are painful; a ratio below 1 flags a strategy with an unpleasant left tail — frequent small gains funding rare large losses, the classic short-volatility shape.

The gain-to-pain ratio works on the whole history rather than just the tails: sum every positive monthly (or daily) return and divide by the absolute sum of every negative return. A gain-to-pain ratio of 2 means the strategy has made twice as much from its winning periods as it has lost from its losing ones, cumulatively.

Suppose a strategy's monthly returns over two years sum to +18 on winning months and −6 on losing months (in percentage points). Gain-to-pain = 18 / 6 = 3.0 — a healthy figure, since anything above roughly 1.5–2 is generally considered strong. Separately, if its best day was +2.1% and its 5th-percentile day was −1.4%, the tail ratio is 2.1 / 1.4 ≈ 1.5, suggesting the daily distribution isn't obviously skewed toward blowup risk.

Both ratios are cheap, distribution-shape complements to Sharpe: they don't require assuming returns are normally distributed, and they surface the asymmetric-risk strategies — options-selling, some carry trades — that a single volatility-adjusted number can quietly flatter.