We compare predictions from a conventional protocol-based approach to risk assessment with those based on a machine-learning approach. We first show that the conventional predictions are less accurate than, and have similar rates of negative prediction error as, a simple Bayes classifier that makes use of only the base failure rate. Machine-learning algorithms based on the underlying risk assessment questionnaire do better under the assumption that negative prediction errors are more costly than positive prediction errors. Machine-learning models based on two-year criminal histories do even better. Indeed, adding the protocol-based features to the criminal histories adds little to the predictive adequacy of the model. We suggest using the predictions based on criminal histories to prioritize incoming calls for service, and devising a more sensitive instrument to distinguish true from false positives that result from this initial screening.