Exploring Game Balance in the Scandinavian Fox Game with Monte-Carlo Tree Search

Abstract

This paper explores if Monte-Carlo Tree Search (MCTS) can perform well in Fox Game, a classic Scandinavian strategy game. MCTS is implemented using a cutoff in the simulation phase. The game state is then evaluated using a heuristic function that is formulated using theoretical arguments from its chess counterpart. MCTS is shown to perform on the same level as highly experienced human players, using limited computational resources. According to popular belief, as can be seen in online forums, the asymmetry in Fox Game leads to imbalances which favors the foxes. However the experiments in this paper show that, contrary to popular belief, it is the sheep that are favored, and quite heavily so. It is also shown that the game can be made more balanced by reducing the number of sheep at the start of the game from 20 to 18.