Back to library
Black SwanMental model

Thick tail (and thin tail)

Толстый хвост · Тонкий хвост · Fat tails · Тяжёлые хвосты

In thick-tailed distributions, extreme events are more frequent and more extreme than the bell curve predicts. The tail does most of the work.

Plain explanation

Thin tails (Gaussian) describe heights, blood pressure, measurement noise. Thick tails describe market returns, wealth, pandemic spread. In thick-tail land, the «average year» tells you nothing - one rare year contains 90% of the action.

Why it matters

Risk models that assume thin tails (most pre-2008 financial models did) systematically miss the events that actually matter.

Practical example

A 1-in-1000-year financial event happens every decade. The «1-in-1000» was computed from the thin-tail assumption.

How to use
  1. 1Always ask «is this domain thin-tailed or thick-tailed?» before applying any statistical reasoning.
Read the original book

This part of the knowledge base is inspired by the book. Go to the Ukrainian edition to explore the concept in depth.

Source notes
  • · Меточка по Талебу - fat tails / thin tails