Analyzing E-Sports match win and losses through feature modeling

Abstract

While sports data analysis initially relied primarily on personal intuition and experience, it has rapidly evolved into a data- and numbers-driven analysis. It has become increasingly common for modern sports analytics to collect and analyze large amounts of data and use statistical models and machine learning algorithms to predict game outcomes.This approach makes more accurate and consistent analysis available, helps identify complex patterns, and increases the reliability of predictions by accounting for the many variables in a sporting event. In the swiftly growing field of E-Sports, analyzing game data for match strategy and winning has become more and more important. However, even League of Legends, which has the largest gaming market, only includes limited data such as kills, deaths, and gold in its win analysis model. In this paper, we focus on how game win analysis can be made more sophisticated through selecting and analyzing factors not previously considered. These results provide a statistical indication of the importance of game factors, which will provide useful guidance for building winning strategies in e-sports.