Ord–Hammond Catastrophe Hypothesis

The Ord–Hammond Catastrophe Hypothesis proposes that the probability distribution of global catastrophic events follows a power-law structure: catastrophes affecting larger fractions of the global population are exponentially less likely, but the tail of the distribution is fat enough that civilization-scale events remain non-negligible. Named after Toby Ord and Jim Hammond, the hypothesis challenges the assumption that catastrophic risks can be managed by estimating a single expected value. Instead, it suggests that the tail properties of the distribution — the correlation between different risk domains and the potential for cascading amplification — dominate the total risk landscape.

The hypothesis has implications for global catastrophic risk assessment: if catastrophic events are power-law distributed, then historical frequency is a systematically biased estimator of future probability, because the largest events are by definition outside historical experience. The hypothesis also implies that risk reduction efforts should prioritize tail resilience — the capacity to survive events larger than any that have occurred — over optimization against historically observed hazards.