This paper explores the determinants of NBA player performance and market value using updated data from each players’ entire seasons (updated to 14th February 2025). Twenty-one variables—including Value (an overall intangible rating), Points Per Game (PPG), Player Efficiency Rating (PER), Win Shares (WS), Box Plus-Minus (BPM), Defensive Rating (DRtg), and True Shooting Percentage (TS%) alongside additional indicators such as Assist-to-Turnover Ratio (ATR), rebounds (TRB%), usage (AST%), offensive rating (ORtg), and games played rate (G/82%) were analysed. A multi-variable linear regression model was applied to evaluate how these metrics predict a player’s value. The results revealed that WS is most strongly correlate with players’ value. Notably, older, more experienced players with balanced offensive and defensive metrics commanded higher salaries. Visualizations, including heatmaps for correlation, radar charts for individual skill profiles, and box plots for outlier detection, provided additional insights into the relationships among variables. Overall, the findings underscore the importance of both advanced on-court metrics in shaping player market value. The study concludes with a discussion of limitations—such as not accounting for injuries or future potential—and suggests directions for future research, including more granular in-season data and more refined intangible and off-court metrics.
Chuhao Lin (Thu,) studied this question.