Depth, balancing, and limits of the Elo model

Cauwet, Marie-Liesse; Teytaud, Olivier; Cazenave, Tristan; Saffidine, Abdallah; Liang, Hua-Min; Yen, Shi-Jim; Lin, Hung‐Hsuan; Wu, I‐Chen

doi:10.1109/cig.2015.7317964

Cited by 2 publications

(1 citation statement)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In contrast to the previous work (Sephton, Cowling, and Slaven 2015), our empirical results in the next subsection show that strengths can be adjusted across a wide range over 800 Elo rating with the threshold ratio 0.1 and the interval of in . Thus, our approach is very suitable for games that are considered to have very high depth (Cauwet et al 2015).…”

Section: Our Approachmentioning

confidence: 99%

On Strength Adjustment for MCTS-Based Programs

Liu

et al. 2019

AAAI

Self Cite

View full text Add to dashboard Cite

This paper proposes an approach to strength adjustment for MCTS-based game-playing programs. In this approach, we use a softmax policy with a strength index z to choose moves. Most importantly, we filter low quality moves by excluding those that have a lower simulation count than a pre-defined threshold ratio of the maximum simulation count. We perform a theoretical analysis, reaching the result that the adjusted policy is guaranteed to choose moves exceeding a lower bound in strength by using a threshold ratio. The approach is applied to the Go program ELF OpenGo. The experiment results show that z is highly correlated to the empirical strength; namely, given a threshold ratio 0.1, z is linearly related to the Elo rating with regression error 47.95 Elo where −2≤z ≤2. Meanwhile, the covered strength range is about 800 Elo ratings in the interval of z in [−2,2]. With the ease of strength adjustment using z, we present two methods to adjust strength and predict opponents’ strengths dynamically. To our knowledge, this result is state-of-the-art in terms of the range of strengths in Elo rating while maintaining a controllable relationship between the strength and a strength index.

show abstract

Section: Our Approachmentioning

confidence: 99%