As I write, a television pundit describes the London riots as a ‘random’ act. Random is a word that people tend to use randomly, without a definite plan or purpose. Medicine, which some have called a science, would offer a reasonably prominent spot to random and its many derivatives, such as randomization, in a word cloud of conversations at vivas and journal clubs. But science doesn't come naturally to most clinicians and neither do the finer points of randomized controlled trials – and those befuddled clinicians range from medical students and junior doctors to professors and journal editors. Applying received wisdom and an instinctive feel for doing the right thing (art) remain more natural skills than applying hard logic (science). That doesn't mean we execute either approach well, and indeed a balance of the two is probably the essence of good medical practice. Clinicians, it seems, are increasingly unlikely to win a Nobel Prize in Medicine. According to Hutan Ashrafian and colleagues, clinicians accounted for 79% of medical scientists to win a Nobel Prize in the first 30 years of the awards; a figure that has dropped to 26% in the last 30 (JRSM 2011;104:387–9). Focus is on basic biological mechanisms of medical discovery instead of translation of scientific discoveries to wider clinical use. Is this even a problem? Medicine has changed. Clinicians were once taught at length about the scientific method and the philosophy of science. Today, technological advances and our greater understanding of the human body offer us much else to fill the heads of students and trainees. We have guidelines and algorithms, polices and patient charters plenty enough to learn. Science has become a specialist subject. For most clinicians, there is no value in being able to distinguish a frequentist from a Bayesian. No benefit in understanding the science in the healing arts. Clinical medicine is a messy business, and Ashrafian and colleagues argue for special dispensation. Constructing a flawless scientific study to evaluate translational innovation is such an immense challenge that those who interpret Alfred Nobel's wishes should be mindful of the difficulties. In the meantime, some might view medicine as more art than science but many medical researchers consider the application of the scientific method to the messy world of clinical medicine a fascinating and essential challenge. To these gurus, a riot is no more random than the purpose of randomization in a clinical trial is a purely statistical construct. In case I've lost you, Ian Chalmers will help you understand how the landmark 1948 MRC trial of streptomycin used allocation based on random numbers to minimize allocation bias, enable like to be compared with like, and fulfil a clinical need (JRSM 2011;104:383–6).