There is no gold-standard measurement for 'fetal asphyxia'; methods of fetal surveillance must therefore be evaluated in the context of a management program that includes clinical intervention when a diagnosis of 'fetal compromise' is made. The use of non-randomized comparison groups is prone to major bias which can only be avoided satisfactorily by random assignment of alternative monitoring policies. To avoid being misled by random errors, randomized controlled trials must be surprisingly large. The hypothesis, generated from the results of four small trials (total of 2,000 women), that more intensive intrapartum monitoring reduces the risk of neonatal seizures, was tested and sustained in the larger Dublin trial (13,000 women). Early-onset neonatal seizures now provide the best validated epidemiological index of obstetrically-preventable intrapartum asphyxia at or after term. Paediatric follow-up at age 4 failed to identify any beneficial effect of intensive monitoring on cerebral palsy, despite the protective effect on neonatal seizures. The results of the Dublin trial are consistent with those of comparable trials, and this enhances the generalizability of the study's results.