# Kurtosis

*Risk & Portfolio — Finicade finance glossary*

Kurtosis measures how much of a distribution's variance comes from rare extreme observations rather than ordinary ones. A normal distribution has kurtosis of 3, and financial returns routinely show far more — the statistical fingerprint of fat tails. High kurtosis is why models calibrated on typical days understate crisis losses: the average is unchanged, but the probability of a five-sigma move is orders of magnitude higher than the bell curve allows.

**Also known as:** excess kurtosis, leptokurtic, fourth moment

**Related terms:** [Fat Tails](https://finicade.com/glossary/fat-tails), [Skewness](https://finicade.com/glossary/skewness), [Normal Distribution](https://finicade.com/glossary/normal-distribution), [Value at Risk (VaR)](https://finicade.com/glossary/value-at-risk), [Tail Risk](https://finicade.com/glossary/tail-risk)

**Taught in:** Risk Arena — When VaR Fails: Fat Tails & Blowups

Source: https://finicade.com/glossary/kurtosis