Patterns of Go Gaming by Ising Model

摘要

Go game gaming patterns are very hard to identify. The stochastic interaction during a Go game makes highly difficult the pattern recognition in Go gaming. We use the Ising model, a classic method in statistics physics, for modeling the stochastic interaction among spins that result in well identified patterns of phenomena in this discipline. An Ising energy function is defined; this function allows the formal translation of Go game dynamics: the use of rules and tactics to elaborate the complex Go strategies. The result of Go game simulations shows a close fit with real game scores during the evolution of all the game.