In the juvenile justice literature, deep-end interventions such as commitment to a confinement facility are reserved for the most severe delinquents but unfortunately have been shown to have negative consequences. The current study repurposes juvenile confinement within a criminal career context to empirically examine its role in homicide offending based on data from a sample of 445 male, adult habitual criminals. Poisson regression models indicated that juvenile confinementmeasured both dimensionally and categorically-predicted murder arrests despite controls for juvenile homicide offending, juvenile violent delinquency, juvenile felony adjudications, juvenile noncompliance violations, juvenile arrest charges, onset, age, three racial/ethnic classifications, career arrests, career violent index arrests, and career property index arrests. Receiver operating characteristics-area under the curve (ROC-AUC) graphs showed that juvenile confinement predicted murder significantly but modestly better than chance although career violent offending was the strongest predictor of murder perpetration.