Beyond Skill-Based Rating Systems: Analyzing And Evaluating Player Performance

Being competitive is a core element of human nature and rating systems are a great tool for satisfying the need of ranking players for the skills they demonstrate. But the games of today have left rating systems behind in their evolution, making it more and more unlikely that individual players are evaluated fairly for their performances. The goal of this study is to gain a deep understanding of different variations of rating systems and the environments they are used in, and to use that information to build and test a new feature for rating players. Using a skill-based rating system known as TrueSkill as a reference point, a dataset is gathered with the help of volunteers. The resulting implementation is based on idea of rewarding those individuals who performed well in a game but were on the losing side of the match. With this change to the rating system, we believe that it makes it worth for performing well even in situations that seem unwinnable, resulting less quitting from online team matches, and overall better player enjoyment. Based on the findings, we introduce an experimental performance analyzation system that tracks individual performances within a team setting.