A guided missile system is a very complex assembly of interacting sub-systems, and rarely fits into a neat mathematical framework; but the need for a sound theoretical background is perhaps even more acute in this than in other fields, because of the almost prohibitive cost of protracted experimentation and cut-and-try methods. A theoretical model—even an approximate one—can greatly reduce the amount of experimental work necessary to prove a system, since the less efficient arrangements can be eliminated without a shot being fired. The use of simulators and computers, although indispensable, is in no way a substitute for this theoretical understanding: for the computer solutions can be obtained only for specific cases, and are of limited use unless they can be generalised to apply to other situations.