Measuring outcomes in neurological practice is difficult. What should be measured? How should outcomes be corrected for the case mix and varying severity of a disease? If this is difficult in the context of a clinical trial, it is harder still in a quality improvement setting, and most challenging of all in the evaluation of a wider service. Attempts to measure the quality and outcome of neurological training share all these problems—along with a few extras… Even seemingly objective comparisons between training programmes, such as trainees’ pass rate in an exit exam, have limitations, owing to factors relating to the trainees themselves (do all programmes have trainees with equal ability?) and the validity of the assessment itself (is passing a knowledge exam an appropriate surrogate for neurological competence?). Equally the small numbers of trainees in some programmes make such comparisons even more fragile. Another way to assess a training programme is to ask …