Transcription of TrueSkill 2: An improved Bayesian skill rating system
{{id}} {{{paragraph}}}
Example: confidence
TrueSkill 2: An improved Bayesian skill rating system
TrueSkill 2: An improved Bayesian skill rating systemTom MinkaMicrosoft ResearchRyan ClevenThe CoalitionYordan ZaykovMicrosoft ResearchMarch 22, 2018 AbstractOnline multiplayer games, such asGears of WarandHalo, use skill -based matchmakingto give players fair and enjoyable matches. They depend on a skill rating system to inferaccurate player skills from historical data. TrueSkill is a popular and effective skill ratingsystem, working from only the winner and loser of each game. This paper presents anextension to TrueSkill that incorporates additional information that is readily available inonline shooters, such as player experience, membership in a squad, the number of kills aplayer scored, tendency to quit, and skill in other game modes. This extension, which wecall TrueSkill2, is shown to significantly improve the accuracy of skill ratings computedfromHalo 5matches. TrueSkill2 predicts historical match outcomes with 68% accuracy,compared to 52% accuracy for IntroductionWhen a player wants to play an online multiplayer game, such asHaloorGears of War, theyjoin a queue of waiting players, and a matchmaking service decides who they will play with.
Bayesian inference in this generative model gives the optimal skill ratings under the assumptions. ... automatic parameter estimation over a batch of historical data. TrueSkill2 operates in two modes: an online mode that only propagates skill ratings forward in time, and a batch mode
Loading..
Tags:
Information
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document: