Análise multivariada aplicada a E-sports no cenário competitivo de League of Legends

Abstract

Games have been an integral part of human history, evolving from simple forms of entertainment to highly competitive and socially recognized sports. With technological advancements, digital modalities such as e-sports have gained prominence, particularly following the inclusion of Electronic Sports in the Olympic Games by the International Olympic Committee in 2023. This study explores the application of Multivariate Analysis techniques, specifically Principal Component Analysis (PCA) and cluster analysis, to identify patterns in gameplay styles among teams of League of Legends (LoL), a Multiplayer Online Battle Arena (MOBA) game. Data from major global LoL leagues in 2023 were utilized, categorizing variables into individual, collective, strategic, confrontation, and resource-related aspects. Using R software for data processing and analysis, PCA was employed to reduce dimensionality and create principal components that capture the variability in team performances. Cluster analysis, utilizing Euclidean distance and the Ward.D2 method, identified five distinct groups of gameplay styles. The results reveal specific characteristics for each cluster, such as prioritization of objectives, resource advantages, and ward control strategies, which can aid technical committees in defining strategies, analyzing opponents, and selecting players aligned with predominant gameplay styles. The conclusions emphasize the importance of statistical methods in optimizing decision-making within the competitive e-sports landscape, providing valuable tools for the development and enhancement of teams. This study contributes to a deeper understanding of the factors influencing team performance in LoL, offering insights that can be applied to improve efficiency and competitiveness in the realm of e-sports.