Classical neurology teaching highlighted that investigations make a very small contribution to the diagnosis; we were therefore encouraged to focus on our clinical skills, and particularly the history. However, the development of increasingly sophisticated diagnostic techniques, particularly in imaging and genetics, has made us rely more on investigations in coming to a diagnosis. Several factors encourage indiscriminate testing, including the pressure for prompt diagnosis, the limited opportunity for follow up and risk-averse practice driven by increased litigation. Instead of focussing initially on the most likely diagnosis, and then using serial investigations parsimoniously, we feel pushed into running parallel investigations to cover all scenarios, including testing for very rare and unlikely entities. A single broad battery of tests at the outset helps to reassure us that ‘everything is done, and nothing forgotten’. Clinical Grand Rounds are very educational, but the presentation of rare and challenging cases mean the wrong lessons are learnt, with an appreciation of the clinical reasoning used losing out to a need to ‘do all the tests’. The resulting overuse of tests is now widely recognised and has prompted the …