The present study evaluated the convention to design training environments by giving access to easy strategies first and progressing to more difficult and efficient strategies thereafter. An experiment was conducted focusing on training in a simplified editing task. This task could be performed with an easy (mouse-based) strategy as well as with a more efficient (script-based) strategy. Two learning environments were compared, based on the order of the introduction of the following 2 strategies to participants: an easy-first program and a difficult-first program. The results highlight 2 interesting patterns. First, initial training in an easy strategy impaired the acquisition of a more efficient strategy. Second, learning the easy strategy first reduced between-subjects variability. It helped poor performers but resulted in a lower proportion of high-level performers.