A major task of sports analytics is player evaluation. Previous methods commonly measured the impact of players' actions on desirable outcomes (e.g., goals or winning) without considering the risk induced by stochastic game dynamics. In this paper, we design an uncertainty-aware Reinforcement Learning (RL) framework to learn a risk-sensitive player evaluation metric from stochastic game dynamics. To embed the risk of a player's movements into the distribution of action-values, we model their 1) aleatoric uncertainty, which represents the intrinsic stochasticity in a sports game, and 2) epistemic uncertainty, which is due to a model's insufficient knowledge regarding Out-of-Distribution (OoD) samples. We demonstrate how a distributional Bellman operator and a feature-space density model can capture these uncertainties. Based on such uncertainty estimation, we propose a Risk-sensitive Game Impact Metric (RiGIM) that measures players' performance over a season by conditioning on a specific confidence level. Empirical evaluation, based on over 9M play-by-play ice hockey and soccer events, shows that RiGIM correlates highly with standard success measures and has a consistent risk sensitivity.

A major task of sports analytics is player evaluation. Previous methods commonly measured the impact of players' actions on desirable outcomes (e.g., goals or winning) without considering the risk induced by stochastic game dynamics. In this paper, we design an uncertainty-aware Reinforcement Learning (RL) framework to learn a risk-sensitive player evaluation metric from stochastic game dynamics. To embed the risk of a player's movements into the distribution of action-values, we model their 1) aleatoric uncertainty, which represents the intrinsic stochasticity in a sports game, and 2) epistemic uncertainty, which is due to a model's insufficient knowledge regarding Out-of-Distribution (OoD) samples. We demonstrate how a distributional Bellman operator and a feature-space density model can capture these uncertainties.

We treat athletes as if they are real-life superheroes that overcome physical challenges to achieve greatness in their respective sports. Today's athletes are physically faster, stronger and more agile than the generation before, but something is wrong. Some recent news includes the NBA expanding its mental health programme for its players and the NFL changing its rules and procedures to better protect its stars from concussions. The focus of any individual or team sport is to maximise player performance. In our sports culture, we are obsessed with team and player statistics using traditional measures in each sport.

We've seen strong efforts from almost every signature franchise in 2016. The success and quality of the products have raised the anticipation for next year's crop of releases, and it's also caused the sports gaming community to become even more demanding. While the thirst for sports gaming titles shows no sign of tailing off, the fire that burns in our guts for the best products available is still alive. After spending countless hours with just about every sports game released in 2016, I've selected the standout games and modes in a variety of categories. All three of these games delivered ultra-realistic renders, though the dynamics and structure of their sports create a diverse circumstance. Most often, player faces aren't visible in an American football game, but EA Sports' Madden 17 had as many scanned-in head models as the series has ever had.