Bayesian inference for some Lanchester combat laws

In this paper, we illustrate Bayesian statistical inference for some simple stochastic versions of the Lanchester linear and square laws for combat. We show that, given battle data, Bayesian methods may be used both to predict which side will win the battle and to choose between alternative models. We illustrate our approach with some simulated tank battle data. It is shown that for these data, there is little difference between the models considered.