The frequentist properties of a procedure proposed by Begg & Mehta (1979) for the comparison of two treatments in the context of a clinical trial are compared to those of the Bayes procedure and the procedure proposed by Anscombe (1963). While the last two procedures are quite effective, the first prescribes premature termination of the experimental phase which leads to a greatly increased chance of assigning the inferior treatment to the remaining patients in the clinical trial and results in the mistreatment of an unnecessarily large number of patients.