Visualizing Performance Dynamics of Soccer Players in Personalized Networks

    Activity: Talk and presentationAcademic presentationAcademic


    Research on soccer performance and its related factors has mostly been conducted 1) at one or a few points in time, 2) on a group level, and 3) as a causal-chain of monodimensional predictor and outcome variables. This approach has not yet led to a clear set of variables that can explain soccer performance, and omitted the fact that performance factors can interact in individual-specific, complex ways. In accordance with recent developments we argue that soccer performance should be approached as 1) dynamic, 2) individual-specific, and 3) emerging from (mutual) interactions between multidimensional factors (Den Hartigh et al., 2016; Glazier, 2017; Neumann et al., 2021; Unnithan et al., 2012). The purpose of this research is to present a novel analytical method, that is, time-varying vector-autoregressive (TV-VAR) modeling, that may capture the interactions and changes between the performance-related factors of individual soccer players (Bringmann et al., 2018). Therefore, we measured important psychological (e.g., confidence, motivation) and physiological (e.g., load, recovery) factors of youth male players from a professional soccer club on a daily basis across one season. We assessed the temporal dynamics (i.e., autoregressive and cross-lagged effects) of those factors and visualized the findings in changing network graphs. Results show that the performance-related factors are influencing themselves and other factors with sometimes changing effects over time. For instance, the findings of one player revealed that self-confidence was a central element in the network as it influenced many other factors, including the performance of the player. For a second player, a different network structure was found with multiple factors influencing each other, such as the recovery, self-confidence, mood, and performance. The findings confirm the assumption that the performance networks are highly individual-specific. Hence, TV-VAR models offer a tool to capture the dynamics, individual-specificity, and multidimensional interaction of soccer performance. In that sense, the models may allow researchers and practitioners to know which knob to turn for which player and to detect when relations between factors change, which may be indicative of a change to a higher or lower level in performance (Hill et al., 2020; Scheffer et al., 2018).
    Event titleWorld Congress on Science and Soccer 2022
    Event typeConference
    LocationCoimbra, PortugalShow on map
    Degree of RecognitionInternational


    • WCSS 2022
    • Sports
    • Football
    • Soccer
    • Networks
    • Dynamic modeling
    • time series
    • Multidisciplinarity
    • Interactions
    • Persoanlized Approach
    • Performance