Estimating football player skills is a key task in sports analytics. This paper introduces several extensions to the widely used Expected Possession Value (EPV) model to address the selection problem challenge. First, we assign greater weight to events occurring immediately before a shot (decay effect). Second, our model more accurately accounts for possession risk by incorporating both the decay effect and effective playing time. Third, we assess individual player abilities in winning aerial and ground duels. Using the extended EPV model, we develop a Pass-Carry Reward (PCR) metric that characterizes a player’s skill in improving team possession. Furthermore, we predict values of this metric for various football players in the upcoming season, particularly considering the strengths of their teams and opponents by utilizing a modified Glicko-2 rating system.