Systems for the quantitation of injury severity are crucial to the study of injury epidemiology, and to comparative evaluations of health care delivery in different regions, patient mixes, and environments. Existing anatomic injury severity scores suffer from subjectivity, and may be computed only at relatively high man-hour expense. In this paper we present an injury severity ranking system termed the Anatomic Index, derived from objective, observed probabilities of mortality, and computable from the same HICDA diagnostic codes widely available in hospitals and through software in hospital computer systems. Mathematic underpinnings of the method, theoretical consequences of underprediction and overprediction of mortality, and advantages of the new Index are presented.