Ralf Herbrich and Thore Graepel
We present a new Bayesian skill rating system which can be viewed as a generalisation of the Elo system used in Chess. The new system tracks the uncertainty about player skills, explicitly models draws, can deal with any number of competing entities and can infer individual skills from team results. Inference is performed by approximate message passing on a factor graph representation of the model. We present experimental evidence on the increased accuracy and convergence speed of the system compared to Elo and report on our experience with the new rating system running in a large-scale commercial online gaming service under the name of TrueSkill.
|Institution||Microsoft Research Ltd.|
Pierre Dangauthier, Ralf Herbrich, Tom Minka, and Thore Graepel. TrueSkill Through Time: Revisiting the History of Chess, MIT Press, 2008.
Ralf Herbrich, Tom Minka, and Thore Graepel. TrueSkill(TM): A Bayesian Skill Rating System, MIT Press, January 2007.