Because moderate and low level substance users are relatively common a "Prevention Paradox" is observed in that most incidents of harm occur in these groups rather than amongst frequent and heavy substance users. To extend consideration to prevention in younger age groups two studies of children and adolescents conducted in Victoria, Australia were reanalysed by recombining developmental, social and individual measures to form cumulative risk indexes for substance use. In a large cross-sectional student survey at around age 16 the majority of regular tobacco, alcohol, and cannabis use occurred in the moderate and low risk groups, hence the Prevention Paradox applied. However, the majority of illicit drug use occurred in the highest risk group (top 15%) and also the Prevention Paradox did generally not hold for drug use at younger ages. In a major longitudinal study risk factors at around age 11/12 years were used to predict substance use at age 17/18 years. The Prevention Paradox held for involvement in frequent smoking, heavy drinking as well as any use of cannabis, but was less clear for frequent cannabis use. It is concluded that universal prevention strategies are needed for late adolescent alcohol, tobacco and cannabis use and more targeted strategies for addressing harm related to early age drug use, frequent cannabis use and illegal drug use.