Current theories of skill learning suggest that novices learn optimally in a simplified environment. This information can be incorporated in simulator designs. Our purpose was to assess whether basic visuospatial training is beneficial for performance on an arthroscopy model. One group of trainees practiced three visuomotor tasks while the other group was not given this opportunity. Both groups then performed three different surgical tasks on a simulated arthroscopy model. Practice with the visuomotor tasks enhanced performance on two of the tasks on the arthroscopy model. The basic navigational skills learned through practice transferred to the performance of arthroscopic surgery tasks and these skills should be included in the design of a comprehensive arthroscopy simulator.