Combining antenatal risk factors to generate an overall probability of risk to the fetus in an individual case presents many problems. A computer simulation using Bayes theorem provides a means for evaluating different approaches to this problem. Some of the findings of this simulation are unexpected, in particular: (1) the distribution of probabilities in a series of cases is highly skewed; and (2) the relationship between numbers of risk factors and overall risk followed a sigmoid curve. These observations are of considerable potential significance in the clinical prediction of fetal risk.<<ETX>>