Title: Alliance Prediction by Energy Minimization: Neutrality

Abstract: We extend the method of Axelrod and Bennetts for alliance prediction, based very loosely on the Spin-Glass theory of magnetism, to include the possibility of a neutral camp. We also explore the effect of physical separation between players. Using the second European war data, we demonstrate that either one of these extensions is sufficient to move Portugal out of the Axis camp, where it was previously found. If neutrality is allowed, then Portugal is found in that camp by itself. Without a neutral camp, Portugal is placed in the Allied camp if affinities decay with characteristic lengths of about 2000km. Most importantly, the consistency of results under such non-trivial alterations, provides incremental evidence for the robustness of the model as an empirical tool.