Social Network Analysis is widely used to study interaction structures across diverse domains, yet a persistent methodological concern is how to define and weight edges in a network. Conventional approaches often rely on interaction frequency, which may overlook differences in the impact or significance of individual interactions. This study addresses this issue by proposing a shift in network construction from quantity-based to quality-based weighting. Using football passing interactions as a concrete empirical example, we integrate data-driven pass values derived from Expected Threat (xT) to construct a contribution-based Valued Passing Network (VPN) framework. We compare it with the Conventional Passing Network (CPN), which often overlooks the varying difficulty and impact of passes by focusing solely on their frequency. Using event data from the FIFA World Cup 2022, the VPN is quantitatively compared with the Conventional Passing Network (CPN) by examining correlations between network metrics and benchmarked player ratings. The results show stronger correlations with VPN’s out-degree centrality (r = 0.42) and total degree centrality (r = 0.39). These advantages are particularly evident when analysed by position, especially for forwards (r = 0.44) and midfielders (r = 0.51). These differences were further verified as statistically significant through permutation testing. A case study of the World Cup Final match further illustrates how the VPN provides deeper insights into player contributions and team coordination quality in practice than the CPN. Altogether, this study demonstrates that incorporating data-driven pass value into network models can enhance the assessment of player performance and team dynamic coordination, while also ensuring a balance between analytical depth and practical clarity.
Ma et al. (Sat,) studied this question.