48 postgraduate third-year orthopaedic residents completed arthroscopic knee surgery training using a simulator or traditional methods. The purpose of this trial was to determine if simulator-trained residents performed arthroscopic knee surgery in the operating room with greater thoroughness and accuracy in less time compared to residents trained using traditional methods. Orthopaedic residents trained on a realistic virtual-reality arthroscopic knee simulator performed statistically significantly better than the control group when analyzed using a checklist for rating procedural skills and probing skills. Training groups were similar with respect to visualization skills and time to complete the arthroscopic procedure.