AbstractThe paper discusses the difficulties caused by the non-repeatable nature of all operational research studies. The author regrets the anecdotal nature of much of our research and makes a plea for more documentation of case studies rather than word of mouth folklore. Simulation is shown to be a powerful tool for experimental replication and is especially valuable in that it offers a better approach than mathematical modelling. In a final section, doubt is cast on the usefulness of much of the teaching of O.R.