Recommendation System for Defining Competitive Teams in eSports: Experimental Analysis of CBLOL
Resumo
Roster building in esports is a complex strategic challenge that can be optimized using data-driven methodologies. This study proposes a novel approach for assembling League of Legends teams in the CBLOL, integrating performance analysis and machine learning techniques. The method accounts for role-specific performance metrics and introduces a network-based metric inspired by the Erdős Number to assess a player’s competitive experience. The core objective is to build a recommendation system that supports roster selection by identifying players likely to contribute to team success. Case studies show the methodology’s application in team selection, When predicting team performance in the CBLOL 2024 Split 2, the model yielded a Pearson correlation of 0.89 between predicted and actual rankings, highlighting its effectiveness in competitive roster building.
Palavras-chave:
esports, recommendation, machine learning, League of Legends, feature selection
Referências
Breiman, L. (2001). Random forests. Machine Learning, 45(1):5–32.
Esports, L. (2024). Lol esports welcomes faker to hall of legends. Accessed: 2025-01-05.
Fandom (2024). Leaguepedia - api documentation.
ge.globo (2023a). Cblol já teve supertimes antes da kabum, e eles fracassaram.
ge.globo (2023b). Final cblol 2023: Loud é melhor time da história do brasil, dizem robo e tin.
Hoffman, P. (1998). The Man Who Loved Only Numbers: The Story of Paul Erd˝os and the Search for Mathematical Truth. Hyperion, New York.
Hosmer, D. W. e Lemeshow, S. (2000). Applied logistic regression.
Kearns, M., Littman, M. L., e Singh, S. (2001). Graphical models for game theory. In Proceedings of the Seventeenth Conference on Uncertainty in Artificial Intelligence (UAI2001), pages 253–260, San Francisco, CA, USA. Morgan Kaufmann Publishers Inc.
Kim, J., Dunkel, W., e Lee, B. (2024). Operationalizing general mechanical skill in timepressure action esports. Journal of Electronic Gaming and Esports, 2(1):11.
La Cava, W., Bauer, C., Moore, J. H., e Pendergrass, S. A. (2019). Interpretation of machine learning predictions for patient outcomes in electronic health records. arXiv preprint arXiv:1903.12074.
League of Legends (2024). Como jogar.
Mais Esports (2024). Cblol 2024: Liga teve o terceiro maior pico de audiência de finais.
Moore, E. F. (1959). The shortest path through a maze. Proceedings of the International Symposium on the Theory of Switching, pages 285–292.
Nagorsky, E. eWiemeyer, J. (2020). The structure of performance and training in esports. PLOS ONE, 15(8):e0237584.
OECD (2008). Handbook on Constructing Composite Indicators: Methodology and User Guide.
TechTudo (2020). Cblol 2021: Entenda seleção dos times e regras do sistema de franquias esports.
Xu, E. L., Qian, X., Liu, T., e Cui, S. (2016). Detection of cooperative interactions in logistic regression models. arXiv preprint arXiv:1602.03963.
Zar, J. H. (2005). Spearman rank correlation. Encyclopedia of Biostatistics.
Esports, L. (2024). Lol esports welcomes faker to hall of legends. Accessed: 2025-01-05.
Fandom (2024). Leaguepedia - api documentation.
ge.globo (2023a). Cblol já teve supertimes antes da kabum, e eles fracassaram.
ge.globo (2023b). Final cblol 2023: Loud é melhor time da história do brasil, dizem robo e tin.
Hoffman, P. (1998). The Man Who Loved Only Numbers: The Story of Paul Erd˝os and the Search for Mathematical Truth. Hyperion, New York.
Hosmer, D. W. e Lemeshow, S. (2000). Applied logistic regression.
Kearns, M., Littman, M. L., e Singh, S. (2001). Graphical models for game theory. In Proceedings of the Seventeenth Conference on Uncertainty in Artificial Intelligence (UAI2001), pages 253–260, San Francisco, CA, USA. Morgan Kaufmann Publishers Inc.
Kim, J., Dunkel, W., e Lee, B. (2024). Operationalizing general mechanical skill in timepressure action esports. Journal of Electronic Gaming and Esports, 2(1):11.
La Cava, W., Bauer, C., Moore, J. H., e Pendergrass, S. A. (2019). Interpretation of machine learning predictions for patient outcomes in electronic health records. arXiv preprint arXiv:1903.12074.
League of Legends (2024). Como jogar.
Mais Esports (2024). Cblol 2024: Liga teve o terceiro maior pico de audiência de finais.
Moore, E. F. (1959). The shortest path through a maze. Proceedings of the International Symposium on the Theory of Switching, pages 285–292.
Nagorsky, E. eWiemeyer, J. (2020). The structure of performance and training in esports. PLOS ONE, 15(8):e0237584.
OECD (2008). Handbook on Constructing Composite Indicators: Methodology and User Guide.
TechTudo (2020). Cblol 2021: Entenda seleção dos times e regras do sistema de franquias esports.
Xu, E. L., Qian, X., Liu, T., e Cui, S. (2016). Detection of cooperative interactions in logistic regression models. arXiv preprint arXiv:1602.03963.
Zar, J. H. (2005). Spearman rank correlation. Encyclopedia of Biostatistics.
Publicado
30/09/2025
Como Citar
ZECCA, Vinícius Antonio Ramos; CONTRERAS, Rodrigo Colnago.
Recommendation System for Defining Competitive Teams in eSports: Experimental Analysis of CBLOL. In: TRILHA DE COMPUTAÇÃO – ARTIGOS CURTOS - SIMPÓSIO BRASILEIRO DE JOGOS E ENTRETENIMENTO DIGITAL (SBGAMES), 14. , 2025, Salvador/BA.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2025
.
p. 19-25.
DOI: https://doi.org/10.5753/sbgames_estendido.2025.7214.
