Thore Graepel is a Principal Researcher at Microsoft Research and Professor of Machine Learning at University College London. His research interests revolve around large-scale applications of machine learning including skill estimation, user modelling, recommender systems and game AI. Thore is one of the co-inventors of the TrueSkill algorithm for ranking and matchmaking used in Xbox Live, and has also contributed to various aspects of user modelling and recommendation in Microsoft's services. More recently, his work on estimating personality and other user attributes from social networking data has received a lot of attention.
In multiplayer online gaming one of the key questions is how we can match players together that are a good fit and hence will have a great gaming experience together. Different aspects such as existing social connections, personality, quality of connection, and skill can play a role. In this session I will focus on the skill dimension and explain how the TrueSkill system works out the individual skills of players from the observed game outcomes. I will discuss the impact on the quality of matchmaking and give some good practice advice for competitive matchmaking. Time permitting, I will discuss some future ideas for matchmaking including personality based matches