Chess Playing Task

A Chess Playing Task is a game-playing task for playing a chess game.

• Context:
  • It can range from being a Human-Only Chess Playing Task to being a Human-Machine Chess Playing Task to being an Machine-Only Chess Playing Task.
  • …
• Example(s):
  • Kasparov/Deep Blue Chess Match.
  • ...
• Counter-Example(s):
  • Tic-Tac-Toe Playing Task.
  • Checkers Playing Task.
  • Go Game Playing Task.
• See: Domain-Specific Intelligence Task, Rules of Chess.

---

---

References

2024

• (Wikipedia, 2024) ⇒ https://en.wikipedia.org/wiki/Chess#Organized_competition Retrieved:2024-3-12.
  • Contemporary chess is an organized sport with structured international and national leagues, tournaments, and congresses. Thousands of chess tournaments, matches, and festivals are held around the world every year catering to players of all levels.

    Tournaments with a small number of players may use the round-robin format, in which every player plays one game against every other player. For a large number of players, the Swiss system may be used, in which each player is paired against an opponent who has the same (or as similar as possible) score in each round. In either case, a player's score is usually calculated as 1 point for each game won and one-half point for each game drawn. Variations such as "football scoring" (3 points for a win, 1 point for a draw) may be used by tournament organizers, but ratings are always calculated on the basis of standard scoring. A player's score may be reported as total score out of games played (e.g. 5½/8), points for versus points against (e.g. 5½–2½), or by number of wins, losses and draws (e.g. +4−1=3).

    The term "match" refers not to an individual game, but to either a series of games between two players, or a team competition in which each player of one team plays one game against a player of the other team.

2024

• (Ruoss et al., 2024) ⇒ [[::Anian Ruoss]], [[:: Grégoire Delétang]], [[::Sourabh Medapati]], [[::Jordi Grau-Moya]], [[::Li Kevin Wenliang]], [[::Elliot Catt]], [[::John Reid]], and [[::Tim Genewein]]. ([[::2024]]). “Grandmaster-Level Chess Without Search.” doi:10.48550/arXiv.2402.04494
  • QUOTE:
    • The paper introduces a novel approach for achieving grandmaster-level chess play without traditional search algorithms.
    • The paper utilizes a transformer model with 270 million parameters trained on 10 million chess games annotated with action-values by Stockfish 16.
    • The paper achieves a blitz Elo rating of 2895 on Lichess, demonstrating high-level performance against human players.
    • The paper compares the model's performance with AlphaZero and GPT-3.5-turbo-instruct, showing its superiority without the need for Monte Carlo tree search (MCTS).
    • The paper emphasizes the importance of model and dataset size, indicating strong chess capabilities emerge with sufficient scale.
    • The paper conducts extensive ablation studies to validate the significance of scale and specific design choices for the model's success.
    • The paper contributes insights into the potential of supervised learning in complex tasks like chess, traditionally dominated by search-based methods.

---