Dynamická evaluace síly týmů v anglické Premier League s využitím přístupu založeného na score-driven rankingu

This thesis investigates the structural and dynamic drivers of team strength in the English Premier League from 1993 to 2025, seeking to quantify the influence of financial disparities versus unobserved sporting momentum and historical prestige. To address the lack of historical financial records, the Multiple Imputation by Chained Equations (MICE) algorithm is applied. The complete longitudinal panel is analyzed using dynamic Generalized Autoregressive Score (GAS) models combined with the Plackett-Luce distribution for rank-ordered data, supplemented by Ordinary Least Squares (OLS) regression focused on squad market value. To ensure the reliability of these findings, the empirical analysis contrasts the full longitudinal history with a modern 20-year subsample, providing a methodological robustness check that mitigates the potential influence of imputed data from the league’s early years. The empirical results confirm that the relative market value of a squad acts as the absolute dominant predictor of sporting success. When this metric is controlled for, the long-term autoregressive memory of the model significantly weakens. This reveals a profound structural reality: modern capital invested in players has largely overwritten historical inertia, demonstrating that the perceived prestige of elite clubs is inseparable from their overwhelming squad valuations. However, short-term, reactive sporting momentum remains a crucial driver of outcomes. Furthermore, the analysis identifies structural market inefficiencies in modern squad construction. The models demonstrate that maintaining an inflated roster actively penalizes overall team strength, while uncovering a striking "age paradox": the transfer market systematically undervalues the on-pitch utility of older players.