There have long been claims that compensations for noneconomic damages are random because tort law does not provide clear guidance regarding these compensations. I investigate, in both settled and tried medical malpractice cases, whether noneconomic damage payments are arbitrary and what determines the probability and size of these payments. I find that payments for noneconomic damages are not completely random. They vary, in predictable ways, with observable characteristics of the case. The data suggest similar patterns in non-medical malpractice cases. I end by discussing the implications of my findings for the debate on the efficiency and rationale of noneconomic damage compensation.