Crime and disorder are concentrated among a relatively few offenders, victims, and places. These regularities have given rise to proposals for selective sentencing, situational crime prevention, and problem-oriented policing, all of which may be effective if people and places run high risks due to relatively permanent characteristics. However, recent evidence suggests another reason for differences in risks: one crime may temporarily increase the risks of further crimes. If this phenomenon is an important cause of repeat incidents, long-term measures that change permanent characteristics may be less effective than temporary measures that only protect high-risk people and places for a short time. A method for distinguishing the relative importance of long-run and short-run factors was developed and tested on a sample of fast food restaurants in San Antonio, Texas. For this sample, permanent characteristics are approximately five times as important as temporary characteristics in explaining repeat offenses, suggesting that, for this sample at least, long-term problem-solving is an appropriate response. Figures, tables, references