Where the spacing effect promotes longer intervals between facts that need to be memorized, the testing effect argues for intervals that are short enough to recall the facts. As the ease by which facts are memorized differs greatly between students, an individual assessment of how well certain facts are represented in memory is required to successfully balance spacing and testing effects. We present a model that adapts itself to the abilities of the student, and show in a real-world experiment that this model outperforms other approaches to spacing.