Sixty-nine high school students practised on geometry subtopics using worked example-problem pairs. The free-choice group had a choice of their practice pairs, whereas the deliberate practice group was presented with practice tasks that covered all and only their weak areas as determined by a pretest. The results demonstrated a significant instructional group by level of expertise interaction with superior performance by the deliberate practice group over the free-choice group for more knowledgeable learners playing the major role in the significant interaction. On the basis of these findings, optimal strategies for organizing student practice in secondary classrooms are discussed in terms of cognitive load factors and expert-novice differences. It was concluded that a deliberate practice strategy over a free-choice strategy was desirable for more knowledgeable but not for less knowledgeable learners.