In each of the 3 years of the program's operation, youths ages 10-17 were randomly assigned to an experimental group or to a no-treatment control group. Arrest records from 1969 to 1975 from the Hawaii Family Court were analyzed for the 335 experimentals and 218 controls to determine the number of youths arrested for at least 1 major offense in the 3-year study period which began with their initial year of participation. Results showed that the arrest rate among the 50 youths arrested for at least 1 major offense in the year before entering the Buddy System was 223 percent lower than for controls with a prior year offense. This suggests that the Buddy System approach is worth continuing with youngsters after they have been arrested for a major criminal offense. However, the arrest rate for those who had been arrested for only juvenile status or minor offenses, or who had never been arrested at all, was higher after spending more than 1 year in the Buddy System. It is speculated that this resulted from Buddy System liaisons with youths who had more serious offense records. Consequently, caution is urged in developing programs for the prevention of major criminal offenses for delinquents with no recent record of such arrests. It is concluded that because the potential effectiveness of delinquency prevention programs is strongly affected by base rates, such programs should be evaluated differently, depending on whether major criminal offenses are considered separately or lumped together with juvenile status and minor offenses. A data table and 11 references are provided.