Action Valuation in Sports: A Survey
Artur Xarles, Sergio Escalera, Thomas B. Moeslund, Albert Clapés
TL;DR
This survey addresses the need for a cohesive understanding of Action Valuation (AV) in sports, detailing how actions are scored by their contribution to future outcomes and outlining core challenges such as data scarcity and the lack of standard evaluation. It presents a nine-dimension taxonomy covering data, methodologies (EB, MDP, RL), architectural modeling, targeted outcomes, credit assignment horizons, action types, and player-awareness, and surveys methodological trends, datasets, and evaluation strategies. The authors identify critical gaps—most notably the public availability of rich AV datasets (combining ED, OTD, and VD) and the absence of a standardized benchmarking framework—and propose concrete directions, including multi-agent RL for simultaneous action valuation and increased exploration of off-ball actions and player-specific valuation. Collectively, the work clarifies how AV can enhance performance analysis, game understanding, and optimization of decision-making in real-world sports settings, with implications for scouting, coaching, and analytics research.
Abstract
Action Valuation (AV) has emerged as a key topic in Sports Analytics, offering valuable insights by assigning scores to individual actions based on their contribution to desired outcomes. Despite a few surveys addressing related concepts such as Player Valuation, there is no comprehensive review dedicated to an in-depth analysis of AV across different sports. In this survey, we introduce a taxonomy with nine dimensions related to the AV task, encompassing data, methodological approaches, evaluation techniques, and practical applications. Through this analysis, we aim to identify the essential characteristics of effective AV methods, highlight existing gaps in research, and propose future directions for advancing the field.