At an August 1938 meeting of the British Association for the Advancement of Science, Professor Lewis F. Richardson attempted to use mathematics to predict the likelihood of war:
The professor reduced to beautiful differential equations general tendencies common to all nations — resentment of defiance, the suspicion that defense is concealed aggression, response to imports by exports, restraint on armaments by the difficulty of paying for them, and, last, grievances and their irrationality.
He concluded there was "no chance of war," which proved to be a somewhat inaccurate prediction.
The Alexandria Town Talk - Sep 27, 1938
offers some more info on what Richardson was up to:
Richardson viewed war instead in Tolstoyan fashion, as a massive phenomenon governed by forces akin to the forces of nature, over which individuals have little or no control. Accordingly, he ignored all those intricacies of diplomatic-strategic analysis usually pursued by political historians and turned his attention to quasi-mechanical and quantifiable processes which, he assumed, govern the dynamics of the international system of sovereign states.
Despite the eccentricity of his mathematical war-prediction model, Richardson was apparently quite influential in the history of mathematics. Wikipedia notes
that he did pioneering work in mathematical techniques of weather forecasting, as well as in the study of fractals.