Dynamic Performance Networks in Sports

Research in the field of sports performance has mostly been conducted 1) at one or a few points in time, 2) on a group level, 3) monodisciplinary, and 4) as a causal-chain of fixed predictor and outcomes variables. In the present study we argue that sports performance should be approached as 1) dynamic, 2) individual-specific, 3) multidimensional, and 4) emerging from interactions between multiple determinants (Den Hartigh et al., 2016; Glazier, 2017; Neumann et al., 2021). 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 individual-specific dynamic networks of (changing) interactions within and between the multidimensional performance determinants (Bringmann et al., 2018). Therefore, we measured important psychological (e.g., confidence, motivation) and physiological (e.g., load, recovery) variables of youth male players from a professional football club on a daily basis across one season. We assessed the temporal dynamics (i.e., autoregressive and cross-lagged effects) of those variables and visualized the findings in changing network graphs. Results show individual-specific interactions within and between the performance determinants with sometimes changing effects over time. Hence, TV-VAR models may be a suitable method to approach performance regarding the four above-stated elements. The models may allow researchers and practitioners to know which knob to turn for which athlete and to detect when relations between variables change to explain and predict performance transitions (Hill et al., 2020). In that sense, any kind of sports may benefit from the method when longitudinal measurements are in place.