This study compared an understudied subgroup of sex offenders, those who had arrests for nonsexual domestic violence (n = 174), with nonviolent sex offenders (n = 467) and sex offenders arrested for nonsexual violent crimes against nonfamily members (n = 205). The domestic batterer subgroup had significantly higher sexual and general recidivism within the first 7 years after supervision. The four standardized risk assessment scales, Rapid Risk Assessment for Sex Offender Recidivism, Static-99, Structured Anchored Clinical Judgment-Minimum Version, and Static-2002, showed significant and similar performance for predicting the nonviolent subgroup's sexual recidivism after supervision, but these scales had less predictive accuracy for the other two subgroups. Two risk factors, general criminality and self-reported use of substances before committing sex crimes, had different predictive accuracy on sexual recidivism across these three subgroups. Implications for improving risk assessment and supervision of sex offenders are discussed.