Informational Analysis of Go, Part 2: Zuozi and Huanqitou

Abstract

This paper explores the second nature of the Go game (Go rules), based on its long history that played out in 9 × 9 board in BC 2356, which later developed into the 19 × 19 with several variants of its rules. An open-source implementation of AlphaZero, namely LeelaZero, as the artificial intelligence (AI) player, is adopted for data collection through self-play experiments. A particular emphasis on two ancient rules of Go, namely as the Huanqitao and Zaozi were conducted. The results are then analyzed and assessed via game refinement theory and motions in mind concepts. An informational analysis indicates that the Go game has been searching for fair and exciting game playing mechanisms that had evolved from the educational needs to competitive settings. Also, the change in the ancient rules to the komi rule mainly emphasized the dominance of such rule change in term of competitive pairing and fairness, while preserving the challenge and excitement of the 19 × 19 Go board game.