Computer-assisted drill & practice is investigated in the context of a short-term, novel mathematical task. A between groups design is used to statistically compare the efficiency of four methods (three response-sensitive and one response-insensitive) for structuring drill & practice in base eight arithmetic. In addition, a descriptive analysis of both task structure and suitability of the task for individualization is performed. While no statistical differences among the treatment groups were observed, substantial information concerning the properties of the task was obtained.