PriMera Scientific Engineering (ISSN: 2834-2550)

Research Article

Volume 6 Issue 6

Proposal for a New Evaluation Index for Human Fallibility in Shogi

Kazushi Mizuta*, Tetsuhiko Kinebuchi and Takeshi Ito

May 28, 2025

View PDF

Abstract

This paper proposes a new evaluation index for evaluating human fallibility in shogi. By analyzing the differences between the decision-making processes of AI and humans, we found that while humans rely on intuition and decisions made within a limited timeframe, AI often makes different decisions because it carries out exhaustive searches. In this study, we developed a new evaluation index that extracts features prone to human error using a policy network trained based on professional game records. This evaluation model uses logistic regression to predict the probability of making a mistake in the endgame. We have shown that the proposed indicator is effective through testing with test data. In the future, we plan to construct similar evaluation indicators for the opening and middle game phases and verify their effectiveness.

Keywords: Shogi AI; Human Fallibility; Policy Network

References

  1. Ito T. “Thinking of Computers vs. Professional Players -Contemporary and Future of Computer Shogi”. Journal of Information Processing Society of Japan 48.12 (2007): 4033-4040. in Japanese
  2. Imahashi H and Ito T. “A Proposal of a New Position Evaluation Index in Shogi that Close to the Human Sense of Value”. IPSJ SIG Technical Report GI-51.2 (2024): 1-9. in Japanese
  3. Yamaoka T and Kanou K. “How to Create Strong Shogi Software - Deep Learning Shogi AI Implemented in Python”. Mynavi Publishing Inc (2021). in Japanese
  4. ShogiHome. https://sunfish-shogi.github.io/electron-shogi/
  5. Kinebuchi T and Ito T. “A Method to Improve the Agreement Rate with Human Moves in Shogi by Considering the Flow”. Journal of Information Processing Society of Japan 58.9 (2010): 1549-1554. in Japanese
  6. Suisho 5. https://drive.google.com/file/d/1T-Go2KImMfKD_4m_j4fQFXrEfaGgAcS_/view
  7. ShogiGUI. http://shogigui.siganus.com
  8. What is the evaluation value 31111 output by the Shogi software?. https://yaneuraou.yaneu.com/2022/06/03/evaluation-value-31111/