Psychopathic traits assessed with expert-rated instruments are well-established predictors for both general and violent felony recidivism outcomes in adjudicated youth. However, it is not clear whether self-report assessments of psychopathic traits show similar predictive value. The current study evaluated both expert-rated (i.e., the Hare Psychopathy Checklist: Youth Version [PCL:YV]) and self-report measures (i.e., the Childhood Psychopathy Scale, Inventory of Callous Unemotional Traits, Antisocial Process Screening Device, and the Youth Psychopathic Traits Inventory) assessing youth psychopathic traits for predicting general and violent felony recidivism among formerly incarcerated high-risk male adolescents ( n = 275). Scores on both expert-rated and self-report instruments were significantly associated with a shorter time to felony rearrest into early adulthood (with an average follow-up window of 11 years). However, the expert-rated PCL:YV, but not self-report measures, provided significant utility for predicting violent felony arrest. These results suggest that in general, youth psychopathic traits, regardless of the specific instrument used, are predictors of general felony rearrest outcomes, while the expert-rated PCL:YV alone provides utility for predicting the most severe crimes (i.e., violent felonies). The strengths and limitations of the assessment modalities are discussed.