In competitive sports and games, rating systems play an often underrated role. Rating systems allow for competitors to measure their own skill, incentivize competitive performances, and is crucial to providing balanced match-ups between competitors. In this paper, we present a novel Bayesian rating system for contests with many participants. This system can be viewed as an extension of the popular Glicko rating system to multiple players, and is widely applicable to popular coding websites such as emph{Kaggle}, emph{LeetCode}, emph{Codeforces}, and emph{TopCoder}. The simplicity of our system allows us to show theoretical bounds for properties such as outlier robustness, running time, among others. In particular, we show that the system encourages emph{truthful play}: that is, intentional losses by a competitor will never raise their rating. Experimentally, the rating system outperforms existing systems in terms of accuracy, and is faster than existing systems by up to an order of magnitude.

The Web Conference is announcing latest news and developments biweekly or on a monthly basis. We respect The General Data Protection Regulation 2016/679.