This paper examines the impact of changes in group composition on group performance, particularly within systems of dynamic groups. Building upon an existing analytical framework, we introduce new concepts such as group Base and Core, and define metrics including Total Member Engagement (TME), Solidity, and Density to quantify group dynamics over time. A case study using data from the Massively Multiplayer Online Game (MMOG) Travian demonstrates how these metrics correlate with performance. The study finds that performance fluctuations are strongly linked to revitalization and churn rates, and predictive models based on these metrics achieve a high accuracy ( \(R^2\) = 0.70). These findings highlight the importance of membership dynamics in understanding team effectiveness and suggest that incorporating communication networks could further enhance predictive accuracy.