Our current ability to predict neurologic outcome based on intrapartum information is very limited. Adding information on neurologic state during the first days of life adds powerfully to the ability to predict long-term outcome. Further enhancement of predictive abilities will require putting together into single studies the kinds of information on predictors that is now reported in scattered and variously selected small series. Although prediction of outcome requires looking forward from information available in the delivery room to include observations on neonatal encephalopathy and its biochemical and other correlates, an understanding of causation requires looking backward from immediate perinatal and delivery room events to consider the contribution of prenatal factors, known and unknown. Many characteristics commonly labeled as indicators of "neonatal hypoxia," for example, may be the first evidence in the infant of earlier-established abnormality. Birth events appear to contribute only a small proportion of CP and much less of other chronic neurologic disabilities. There is a possibility that the addition of new agents to our therapeutic armamentarium may be capable of reducing that number still further. Prediction and causation as related to chronic neurologic disability gain fresh relevance in view of the need to prepare for clinical trials of new therapies in perinatal medicine.