Numerous real-time air traffic control simulation studies have been conducted in the United Kingdom; this paper summarises what has been learned from them. Realistic assessments of the validity of real-time simulation as a method of solving air traffic control problems are now possible in the light of experience, and its future use can be planned in accordance with its known limitations. Real-time simulation cannot answer every kind of question with equal validity, and its relevance for assessing capacity, workload or stress is dubious. Measures in simulation may be chosen primarily for their practical convenience and because they can provide answers expressed in system concepts, but the technique can be used to show discrepancies between subjective and objective evidence, and to suggest why they have arisen. Real-time simulation is both indispensable and over-used in air traffic control.