if they have chosen insufficient numbers of patients for trial to demonstrate benefit, or lack of it, then that in itself is a failure- the simplest kind of error when statisticians and epidemiologists are easy to find to ensure such a mistake is not made