This paper examines whether people benefit more from playing a commercial off-the-shelf game in pairs rather than in solitary mode. The basic idea behind this didactic method is that there is a serious risk that solitary game play yields insufficient articulation and explanation for learning to take place. Participants in the experimental condition played a strategy game in collaborative mode (pairs). Solitary play formed the control condition. During game play data were gathered about engagement (ie, flow). Also, the dialogues of the pairs were recorded. After game completion participants individually completed a knowledge test. For solitary players this ended the session. Collaborating pairs could discuss test answers (without receiving experimenter feedback) and give a final group answer. Collaboration was found not to affect game engagement and also did not affect individual knowledge test scores. The collaboration presumably did not advance the players' individual knowledge because the game dialogues mainly dealt with superficial game features such as move proposals. The collaborating players benefitted significantly from the opportunity to discuss test scores. The discussion revolves around game selection and game didactics (including scripted collaboration and debriefing) as routes for future studies to follow in ways of improving game utilisation in school.