Non-homogeneous random walks on Elo-rating systems
2022
Thesis, Stochastic Processes, Rating Algorithms, Humboldt University of Berlin
The Elo rating system, a paired comparison model, is widely used in sports, games, and various other more abstract applications.
We identify the Elo rating of a single player over time as a spatially non-homogeneous random walk and provide a recurrence theorem for the rating process. Further, we discuss a series of practical questions and improvements regarding the implementation of Elo-type rating systems.
BSc Thesis at Humboldt-Universität zu Berlin, Department of Mathematics, Applied Financial Mathematics & Applied Stochastic Analysis, supervised by Prof. Dr. Dörte Kreher