Boys show a stronger preference for digital entertainment games than girls. For this reason, it may be that game-based learning is more acceptable to boys than to girls. Yet game-based learning might improve the performance of both boys and girls, depending upon the instructional design. In a quasi-experimental study with a secret-trail game, effects were examined on students' subject-matter knowledge. Analysis of covariance revealed that both boys and girls of the game intervention group showed a higher test performance, compared to students of the control group. However, different game activities mediated this effect of the secret-trail game on performance: girls seemed to profit more from searching the Internet to complete assignments and boys from competing with others. The performance of both boys and girls was negatively influenced by technical problems. The results are discussed within the framework of gender-inclusive game design.