Jie Ke versus AlphaGo: A ranking approach using decision making method for large-scale data with incomplete information
Level 5 - mechanism / opinion, no new human data
By design analogy: methodological and computational decision-modeling study evaluated on historical match data.
OpenAlex W2734806994 · doi:10.1016/j.ejor.2017.07.030
What was done
The authors developed a ranking method for large-scale Go datasets with missing matchups using an incomplete fuzzy pairwise comparison matrix, deriving priority vectors via cosine similarity. They tested the model on historical match data from Go4Go.net spanning 45 years, assessing ranking changes after AlphaGo played Ke Jie (Jie Ke) and applying the framework to rank 1,544 Go players.
What was found
The abstract presents no specific numerical results, ranking outcomes, or quantitative performance metrics; it states only that the method was applied to rank top players over 45 years, analyze changes following the AlphaGo vs. Ke Jie matches, and successfully compute rankings for 1,544 players.
Why it matters
It provides a multi-criteria decision-making framework to address incomplete head-to-head comparison matrices when ranking competitors, accommodating both human players and artificial intelligence agents.
Limits
The abstract reports no numerical accuracy comparisons against standard rating systems (such as Elo or Glicko). The approach relies entirely on recorded historical match data from a single database (Go4Go.net), and validation criteria are not specified.
Cited by
- supports The AI program AlphaGo defeated the world champion in the board game Go.