During a 4 and 1/2 month period, 242 patients were referred to the Metropolitan Toronto Forensic Service by the court for a brief psychiatric assessment that involved being rated by a team of clinicians and two external coders regarding potential for future dangerous behavior. Since the original assessment, 18 patients from the sample have been charged with new offenses resulting in a reassessment. Because the new charges represent the first information available on the postassessment activity of patients, a preliminary analysis was conducted on the validity of the assessment process. Dangerous behavior was inferred from such charges as assault causing bodily harm, armed robbery, etc., and the absence of dangerous behavior was attributed to such offenses as theft and fraud. Affirmative predictions of dangerousness were found to be correct in seven of nine cases for one coder and in six out of eight cases for the second coder. The low number of incorrect affirmative predictions is especially striking, since earlier investigations in this area have contained a high rate of false affirmative predictions. Limitations of the study are noted. Tabular data and three references are provided.