Sloan 2026 research papers
Posted: Sat Feb 14, 2026 1:33 am
Research paper abstract:
"Deep Reinforcement Learning for NBA Player Valuation: A Temporal Difference Approach with Shapley Attribution
This paper introduces a deep reinforcement learning framework for evaluating NBA players that learns context-dependent player value directly from game outcomes. Using temporal-difference learning with a distributional win-probability model, the approach estimates how actions and player presence influence expected outcomes across game states. We combine this with a neural Shapley value attribution method to fairly decompose team success into individual contributions while capturing interaction and synergy effects. Empirical results show improved predictive accuracy, greater stability than RAPM, and systematic identification of defensive value and player synergies that traditional metrics fail to capture."
Posters
"Scouting Anyone: Probabilistic Player Archetypes for Any League
This paper introduces a probabilistic framework for identifying basketball player archetypes that can be applied across leagues, including those without access to advanced data. While player segmentation is not new in basketball analytics, we propose a fundamentally different approach based on Archetypal Analysis, which identifies extreme and interpretable player profiles and models players as probabilistic mixtures of these archetypes, explicitly capturing hybrid roles and stylistic flexibility. "
"HoopEval: Individual Player Action Evaluation via Deep Reinforcement Learning
This paper presents HoopEval, a deep reinforcement learning framework for evaluating individual player actions in basketball using spatio-temporal tracking data. The approach models game dynamics and player interactions to estimate the value of both on-ball and off-ball decisions within their tactical context. By decomposing possession-level outcomes into fine-grained action evaluations, HoopEval provides interpretable measures of decision quality beyond traditional statistics. The results demonstrate its potential to support tactical analysis, player development, and data-driven coaching."
Don't think the papers are available yet, at least on Sloan site.
"Deep Reinforcement Learning for NBA Player Valuation: A Temporal Difference Approach with Shapley Attribution
This paper introduces a deep reinforcement learning framework for evaluating NBA players that learns context-dependent player value directly from game outcomes. Using temporal-difference learning with a distributional win-probability model, the approach estimates how actions and player presence influence expected outcomes across game states. We combine this with a neural Shapley value attribution method to fairly decompose team success into individual contributions while capturing interaction and synergy effects. Empirical results show improved predictive accuracy, greater stability than RAPM, and systematic identification of defensive value and player synergies that traditional metrics fail to capture."
Posters
"Scouting Anyone: Probabilistic Player Archetypes for Any League
This paper introduces a probabilistic framework for identifying basketball player archetypes that can be applied across leagues, including those without access to advanced data. While player segmentation is not new in basketball analytics, we propose a fundamentally different approach based on Archetypal Analysis, which identifies extreme and interpretable player profiles and models players as probabilistic mixtures of these archetypes, explicitly capturing hybrid roles and stylistic flexibility. "
"HoopEval: Individual Player Action Evaluation via Deep Reinforcement Learning
This paper presents HoopEval, a deep reinforcement learning framework for evaluating individual player actions in basketball using spatio-temporal tracking data. The approach models game dynamics and player interactions to estimate the value of both on-ball and off-ball decisions within their tactical context. By decomposing possession-level outcomes into fine-grained action evaluations, HoopEval provides interpretable measures of decision quality beyond traditional statistics. The results demonstrate its potential to support tactical analysis, player development, and data-driven coaching."
Don't think the papers are available yet, at least on Sloan site.