Prodromal symptoms and other variables for a sample of 200 young people who had experienced a first-onset functional psychosis, were analyzed to examine their diagnostic efficiency and predictive power in relation to a diagnosis of schizophrenia. Two different techniques were utilized to generate optimal cut-off points for a number of prodromal symptoms, and optimal decision rules to maximize diagnostic efficiency. The product of the chance-corrected sensitivity and specificity, or the area under the QROC curve, was used to assess the predictive efficiency of a number of prodromal variables, DSM-III-R prodromal variables, pre-psychotic deterioration, pre-morbid functioning, and prodromal duration. The SPAN technique generated a decision rule that performed equivalently to the single variable 'duration of prodrome'. Implications of these results for future research are discussed.