In criminological prognoses, the decisionmaker can erroneously predict (1) failure for the individual who succeeds, or (2) success for the individual who fails. These two kinds of error apply to all decisions or estimates irrespective of the means of derivation. The first kind of error is usually termed a 'false positive,' or 'overprediction.' Overpredictions tend to increase in proportion to the decrease in the number of individuals who fit the 'fail' category. False positives cannot be avoided. At present, statistical predictions, and by inference, all clinical predictions, produce a large number of false positives. The precision of prediction methods can probably be improved significantly by investing effort in three areas: the basic data, methods of input (coding) of basic data to analytical systems, and the analytical systems. Although the prediction of recidivism has been subject to a wide range of error, much greater accuracy has been obtained in the predictions of decisions of authorities in the criminal justice system (e.g., judges, parole officers, and probation officers). Modern decision theory and related practical methods can assist in decisionmaking where the objectives are clearly stated. The main issue today is not how to make predictions, but rather why to predict and when. Five footnotes are included.