Third report in a series describing the results of a major longitudinal study into young offending in New Zealand. This particular report posits the question “To what extent is it possible to identify potential young offenders at 10 years of age?”. The results show some prediction is possible but that the practical utility of the findings is low. At best, the results can be used to provide broad guidelines concerning those children who have high or low risks of offending. Paper is heavy on statistics and equations and provides some new statistical approaches to the problem of criminological prediction. Extensive analysis within the report indicated that the two approaches offering the most effective means of identifying potential delinquents were: the unweighted points score system and the same system applied to the population partitioned into subgroups defined on the basis of race and socioeconomic status. Within the second approach, the report found that the non-European subgroup appeared to offer the most favourable situation for prediction.