Results indicated that although the tools achieved limited success in predicting sexual reoffending in the sample, more positive results were found for the prediction of broader forms of reoffending. None of the tools significantly predicted sexual reoffending, although the ERASOR structured professional judgment rating nearly reached significance; the Static-99 did not predict sexual or nonsexual reoffending in the sample of adolescents; and the YLS/CMI and the PCL:YV both significantly predicted nonsexual violence, any violence, and any reoffense. Clinicians should carefully acknowledge current limitations in their ability to predict adolescent sexual reoffending, and they should stay informed regarding new research findings. Future research should continue to examine the predictive validity and clinical utility of risk assessment tools with sexually abusive adolescents. In addition, to guide advances in tools, there is a need for further research on risk factors for adolescent sexual reoffending, including possible age-graded or development specific risk factors. Data were collected from 193 male adolescents who had been enrolled in a non-secure residential sex offender treatment program in a medium-sized Midwestern American city between 1992 and 2006. Tables and references