Clinicians who embark on clinical trials take on a commitment over and above the care of individual patients in their charge and hope to gather knowledge which can help both their patients collectively and society in general. Most are aware that such trials will involve them in considerable time and effort in preparation, execution, and analysis and their patients in extra visits, blood tests, and other assessments. It seems to me, however, that many clinicians undertaking trials are unaware that the procedures necessary when managing patients in groups are quite different from those appropriate to individual patients. I believe that a failure to appreciate this underlies many of the serious flaws in published trials. Such flaws are not merely questions of minor statistical inference to interpretation. They are often major errors in basic trial design or nonsensical manoeuvres in analysis. The result is usually to invalidate the conclusions of the trial (sometimes even before they are read) and thus to waste the time, effort, and energy of those who took part in it. When published, such trials serve to swell the medical literature with information which is at best uninformative and at worst misleading and make it harder to separate the wheat from the chaff.