AlphaZero as a playtest tool
DOI:
https://doi.org/10.5902/2448190485269Keywords:
Software testing, Tabletop games, Machine learning, Neural networkAbstract
The tabletop gaming market has maintained rapid growth in recent
years, reaching tens of billions of dollars. Balancing these games is a market
demand and a challenging discipline that requires significant analytical skill
from the game designer. This skill is built through the conduct and observation of
hundreds of playtests in test groups, in a error-prone process, due to the difficulty
in finding individuals willing to play prototypes repeatedly, with only minor rule
iterations. Additionally, not every game designer keeps a complete record of the
tests and can handle the cause-and-effect relationships that small rule changes
have on the outcomes. This work investigates AlphaZero as a computational
intelligence technique to alleviate the demand for human players in the game
creation and testing process. While existing research in the field seeks a more
efficient agent, this study applies the algorithm to generate a dataset to help
the game designer identify points of imbalance and explore creativity. At the
current stage, a version of AlphaZero is implemented for self-training. Finally,
we discuss areas in which the method can assist the game design process using
the data generated by the algorithm during training.
Downloads
References
BECKER, A.; GÖRLICH, D. What is Game Balancing? - An Examination of Concepts. ParadigmPlus, v. 1, n. 1, p. 22–41, abr. 2020. ISSN 2711-4627.
BOARDGAMEGEEK. SPIEL’22 Preview. 2022. Disponı́vel em: ⟨https://boardgamegeek.com/geekpreview/55/spiel-22-preview?sort=hot⟩. Acesso em: 7 Setembro de 2023.
BOARDGAMEGEEK. 2023. Disponı́vel em: ⟨https://boardgamegeek.com/⟩. Acesso em: 7 Setembro de 2023.
BRITANNICA, T. E. of E. Go. 2023. [Online; acesso em 3 setembro 2023]. Disponı́vel em: ⟨https://www.britannica.com/topic/go-game⟩.
COULOM, R. Efficient selectivity and backup operators in monte-carlo tree search. In: SPRINGER. International conference on computers and games. [S.l.], 2006. p. 72–83.
FöRSTER, R. AlphaZero from Scratch. 2023. [Online; acesso em 12 setembro 2023]. Disponı́vel em: ⟨https://github.com/foersterrobert/AlphaZeroFromScratch⟩.
HE, K. et al. Deep residual learning for image recognition. 2015.
KOCSIS, L.; SZEPESVÁRI, C. Bandit based monte-carlo planning. In: SPRINGER. European conference on machine learning. [S.l.], 2006. p. 282–293.
MARCELO, A.; PESCUITE, J. Design de jogos: Fundamentos. [S.l.: s.n.], 2009.
ROMERO, B.; SCHREIBER, I. Game Balance. 1st edition. ed. Boca Raton: CRC Press, 2021. ISBN 978-1-4987-9957-7 978-1-03-203400-3.
SILVER, D. et al. Mastering the game of Go with deep neural networks and tree search. Nature, v. 529, n. 7587, p. 484–489, jan. 2016. ISSN 0028-0836, 1476-4687.
SILVER, D. et al. Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm. 2017. Disponı́vel em: ⟨https://doi.org/10.48550/arXiv.1712.01815⟩.
SILVER, D. et al. A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play. Science, v. 362, n. 6419, p. 1140–1144, dez. 2018. ISSN 0036-8075, 1095-9203.
ŚWIECHOWSKI, M. et al. Monte carlo tree search: a review of recent modifications and applications. Artificial Intelligence Review, Springer Science and Business Media LLC, v. 56, n. 3, p. 2497–2562, jul 2022. Disponı́vel em: ⟨https://doi.org/10.1007% 2Fs10462-022-10228-y⟩.
TENSORFLOW, G. Get started with TensorFlow.js. 2023? [Online; acesso em 4 setembro 2023]. Disponı́vel em: ⟨https://www.tensorflow.org/js/tutorials⟩.
TEUBER, K. Colonizadores de Catan. 1995.
TRZEWICZEK, I. I play-tested it 100 times. [S.l.], 2017.
WOODS, S. Eurogames: The design, culture and play of modern European board games. [S.l.]: McFarland, 2012.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2023 Revista ComInG - Communications and Innovations Gazette
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Os manuscritos aceitos e publicados são de propriedade da revista ComInG.
Os originais deverão ser acompanhados de documentos de transferência de direitos autorais contendo assinatura dos autores.
A carta de direitos autorais deve ser enviada para o e-mail coming@inf.ufsm.br
É vedada a submissão integral ou parcial do manuscrito a qualquer outro periódico. A responsabilidade do conteúdo dos artigos é exclusiva dos autores.
É vedada a tradução para outro idioma sem a autorização escrita do Editor ouvida a Comissão Editorial.
ENGLISH
Manuscripts accepted and published are the property of the journal ComInG.
The originals must be accompanied by documentation of copyright transfer containing the signature of the authors.
You may not submit full or partial manuscript to another journal. The responsibility of the article's content is exclusive of the authors.
You may not translating into another language without the written permission of the Editor after consultation with the Editorial Board.