Jie Ke versus AlphaGo: A ranking approach using decision making method for large-scale data with incomplete information

Abstract Ranking historical players in sports is challenging since some players have never played against each other. It is even more complex in Go because of AlphaGo, a project based on artificial intelligence, who became the world's number 1 after it defeated the 528th and the 4th human Go players. AlphaGo is ranked high in the current Go ranking system because it is undefeated. The objective of this paper is to propose a new ranking method for large-scale Go players by means of incomplete fuzzy pair-wise comparison matrix whose priority vector is derived using a cosine similarity measure. Using match results provided by Go4Go.net, experiments are designed to rank top Go players in the past 45 years and examine the change in ranking after AlphaGo faced off against Jie Ke. Furthermore, the proposed method was applied to rank all 1544 Go players available at Go4Go.net to illustrate its efficiency in handling large-scale data.