Learning grammar requires practice and practicing grammar can be boring. We examined whether an instructional game with intrinsically integrated game mechanics promotes this practice: compared to rote learning through a quiz. We did so “in the field.” Tens of thousands children visited, in their leisure time, a public website with tens of attractive online games for children during a 6-week-long period. Of these children, 11,949 picked voluntarily our grammar training intervention. Thereafter, unbeknown to them, they were assigned either to the (...) game or the quiz condition. By means of learning analytics, we examined variables related to participants’ persistence and performance. The results showed large participant drop-out before completing the first level in both conditions, confirming the boringness of the topic. More children completed at least one level in the game compared to the quiz. However, more children completed the intervention with the quiz. In the game, children answered fewer questions correctly and made more errors compared to the quiz. These findings suggest that even if a game initially catches user attention, it may not hold it. Plus, even if it is a minimalistic game with intrinsic integration of learning and playing, it may be distractive. We conclude that persistence in practicing grammar may be driven by other means than by a game’s shooting mechanics; for instance, by a desire to learn the topic and a feeling of achievement or by quizzing mechanics. (shrink)