# Fat Tails **Domain:** Risk and statistics **Doc Type:** Developed Concept Node **Maturity:** Developed **Collection:** [[collections/Gamification|Gamification Collection]] ## Definition A **fat-tailed** distribution assigns materially more probability to extreme outcomes than a thin-tailed model such as the normal distribution. In such domains, rare events can dominate totals, sample averages converge slowly, and recent calm provides weak assurance about the next extreme. ## Significance War and mass killing require tail-aware reasoning. Counting ordinary years can understate catastrophic risk, while one extreme event can alter the historical aggregate. Policy should therefore separate frequency from severity and avoid extrapolating safety from short intervals. ## Sources - Pasquale Cirillo and Nassim Nicholas Taleb, [Statistical properties and tail risk of violent conflicts](https://arxiv.org/abs/1505.04722) ## Routes [[wiki/Extreme Value Theory|Extreme Value Theory]] · [[wiki/Pasquale Cirillo and Nassim Nicholas Taleb|Cirillo and Taleb]] · [[wiki/Democide|Democide]]