The routine recommendation to women to count fetal movements daily during late pregnancy for the prevention of antepartum late fetal death in normally formed singletons has been evaluated. 68,000 women were randomly allocated within thirty-three pairs of clusters either to a policy of routine counting or to standard care, which might involve selective use of formal counting or informal noting of movements. Antepartum death rates for normally formed singletons were similar in the two groups, regardless of cause of prior risk status. Despite the counting policy, most of these fetuses were dead by the time the mothers received medical attention. The study does not rule out a beneficial effect, but at best, the policy would have to be used by about 1250 women to prevent 1 unexplained antepartum late fetal death, and an adverse effect is just as likely. In addition, formal routine counting would use considerable extra resources.