The present study examined a group of 447 Dutch forensic psychiatric patients on conditional release (CR). After a brief overview of the Dutch CR system for forensic psychiatric patients is presented, two sets of factors were studied: factors associated with behavioral experts' recommendations on CR readiness (prevalence rates of recommendations in contrast to or consistent with the judge's decision on CR, written vs. oral recommendations and consensus among parties); and characteristics of forensic psychiatric patients on CR imposed consistent with or contrary to behavioral experts' recommendations (i.e., demographic variables, psychopathology, judicial background and outcomes of violent behavior). Patients on CR consistent with or contrary to the behavioral experts' recommendations did not differ in demographic background variables and psychopathology. Patients on CR contrary to the behavioral experts' recommendations, however, had more prior convictions, and their index offense more often included a sex offense and a property offense (with or without violence) as compared to patients on CR consistent with the behavioral experts' recommendations. In addition, decisions contrary to the behavioral experts' recommendations were more often reached after a 3‐month continuation adjournment and after an appeal. Also, patients on CR contrary to the behavioral experts' recommendations were less often found to have been on trial leave before their CR. Finally, recidivism rates were found to be much higher for those on CR contrary to, as compared with those on CR consistent with the behavioral experts' recommendations. However, when corrected for other well‐known static predictors of recidivism, this significant difference was no longer prevalent. The implications for evaluators' CR readiness reports and lessons that may be learned from the Dutch forensic psychiatric CR system are discussed. Copyright © 2016 John Wiley & Sons, Ltd.