Measuring the indirect effects on team performance from teammate interactions, or team synergy, has proven to be an elusive concept in sports analytics. In order to shed further light on this topic, we apply advanced statistical techniques designed to capture player complementarities, or "network" relationships, between teammates on Major League Baseball teams. Using wins-above-replacement metrics (WAR ) over the 1998-2016 seasons and spatial factor models embodying the strength of teammates` on-the-field interactions, we show that roughly 40 percent of the unexplained variation in team performance by WAR can be explained by team synergy. By building a set of novel individual player metrics which control for a player`s effect on his teammates, we are then able to develop some "rules-of-thumb" for team synergy that can be used to guide roster construction.
© Copyright 2019 Journal of Sports Analytics. IOS Press. Alle Rechte vorbehalten.