We propose a methodology for analyzing viewer engagement during Overwatch League (OWL) broadcasts. On the one hand, we implemented image processing techniques to isolate key visual elements of the spectator interface in OWL videos, enabling an objective and granular analysis of the visual experience, qualified by measures of interest. On the other hand, we extracted audio descriptors from OWL videos of competitive events. We believe these audio descriptors are indicative of the excitement of the crowd and commentators, offering a subjective measure of the impact generated by the video broadcast of the encounter. A wrapper approach allowed us to characterize the relevance of the measures of interest by examining their relationship to the hype descriptors.