After 2 years of working on a design research project looking at the development of student thinking about negative numbers in the context of a computer game, I arrived at a critical problem. The game was designed such that the principles of working with adding and subtracting positive and negative numbers were integrated into the game play so that students could only achieve best scores across the game levels by utilizing these principles to make decisions during play. In as little as two sessions of 30 min of play, almost all students were able to master the basic principles as demonstrated by their success across all four game levels. When it came to measuring learning gains on a traditional pencil and paper posttest, however, students showed significant but limited gains. It was hypothesized that the factor limiting posttest scores was that students´ learning had been encoded in such a way as to reflect the purposes and goals of the game and that this encoding didn´t readily transfer to a traditional worksheet filled with addition and subtraction problems. In order to test this hypothesis, we borrowed a theoretical framework to see if students´ game playing could be conceptualized as a form of 'preparation for future learning.” The results give strong support for this conceptualization of transfer in supporting game-based learning and suggest interesting possibilities for bridging game-based learning activities with the reality of more traditional school-based models of assessment.