How does deep conceptual change occur when students play well-designed educational games? To answer this question, we present a case study in the form of a microgenetic analysis of a student's processes of knowledge construction as he played a conceptually-integrated digital game (SURGE Next) designed to support learning about Newtonian mechanics. Grounded in the Knowledge In Pieces framework of conceptual change (A. diSessa, 1993), we analyze the processes through which the student, Jamal, developed an expert-like understanding of deflections, a phenomenon that has been previously identified as challenging to understand for novice physics learners. We also explore the key characteristics of SURGE Next supporting these conceptual change processes. Our analysis shows that Jamal's learning involved iterative refinement of his conceptual understanding through distributed encoding (A. diSessa, 1993). That is, as Jamal advanced through the game levels in SURGE Next, he developed a progressively more distributed sense of mechanism (A. diSessa, 1993) and was able to identify and operationalize the roles of the direction and magnitude of an object's initial (or previous) velocity in determining the velocity resulting from the application of a new impulse. We also discuss the methodological and design implications of our findings for future research on digital games for learning. [An earlier version of this article was presented at the National Association of Research in Science Teaching Conference, 2013.]