The accuracy of different subtest combinations in predicting Verbal, Performance, and Full Scale IQ, scores on the revised edition of the Wechsler Adult Intelligence Scale (WAIS-R) was examined. A best subset multiple regression technique with sample sizes ranging from 1,062 to 1,084 individuals was used. All possible subtest combinations were analyzed to determine the ones that showed the three best and the single worst multiple Rs with IQ scores from the complete WAIS-R. All short forms provided accurate estimates of Full Scale scores, and predictive accuracy increased slightly as more subtests were used. The data suggest that statistically searching for a “best” short form is largely futile. Short forms should be selected on the basis of their efficiency to provide the information required.