Lester Ingber I. Introduction: C 2 in Training and Computer Models A. Necessity of Comparing Computer Models to Exercise Data This project addresses the utility of establishing a mathematical approach to compare exercise data to large scale computer models whose underlying microscopic interactions among men and machines are driven by the natural laws of physics. In this study, the focus is to compare the Janus(T) wargame to National Training Center (NTC) data, since both systems then take into account human interactions. It also should be noted that "large scale" here refers to battalion level. (Army systems scale by factors of 3−5, from company to battalion to brigade to division to corps to army.) If these battalion level computer models can be favorably compared, and if consistency can be achieved between the hierarchy of large scale battalion level, larger scale corps level, and largest scale theater level computer models, then these higher echelon computer models also can be favorably compared. This could only enhance the value of training on these higher echelon computer models. 1 The requirement of depending more and more on combat computer models (including simulations and wargames) has been brought into sharper focus because of many circumstances, e.g.: (1) the nonexistence of ample data from previous wars and training operations, (2) the rapidly shortening time scale on which tactical decisions must be made, (3) the rapidly increasing scale at which men and machines are to be deployed, (4) the increasing awareness of new scenarios that are fundamentally different from historical experiences, (5) and the rapidly increasing expense of conducting training exercises. Furthermore, such computer models could be used to augment training. We now spend several million dollars to cycle each battalion through NTC. The training of these commanders could be greatly enhanced if inexpensive pre and post training wargames were provided that statistically replicate their training missions. Even, or rather especially, for the development of such training aids, proper analysis and modeling is required to quantitatively demonstrate that the computer models are good statistical representations of the training mission. However, the level of acceptance of computer models in major military battle-management and procurement decisions appears to be similar to the level of acceptance of computer simulations in physics in the 1960s. In physics, prior to the 1960s, theory and experiment formed a close bond to serve to understand nature. In the 1960s, academicians were fascinated with evolving computer technology, but very few people seriously accepted results from computer simulations as being on a par with good theory and good experiment. Now, of course, the situation is quite different. The requirement of understanding truly complex systems has placed computer simulation, together with theory and experiment, as an equal leg of a tripod of techniques used to investigate physical nature. The requirements necessary to bring combat c...