Training simulations provide opportunity for users to repeat actions until mastering a task or learning objective. Once a mistake has been identified, however, difficulties may arise in replicating the exact sequence of previous actions, particularly in a complex, open-ended environment. We present an investigation of student learning, reflection, and metacognition when using a 3D simulation with built-in functionality that allows students to redo their actions by focusing on the point of failure instead of starting from the beginning. Interviews and observations were used to gather qualitative data, finding that starting at the point of failure within a complex, multiplayer simulation holds advantages and presents obstacles for student learning.

Two key findings were identified during this study. The first key finding relating to the influence of reflection while redoing on learning was that participants used reflection with varying degrees of complexity as a scaffolding mechanism. The second key finding was that the students' abilities to orient themselves in the problem space contributed to the amount of contextual information they needed to be successful-in this case, either starting from the beginning or from the point of failure.