Introduction: There is paucity in the research on transfer validity of arthroscopic simulator training. The aim of this article is to determine whether skills derived from arthroscopic simulation are transferrable to the operating theatre and retained over time. Methods: A systematic review with rigorous criteria to identify the highest level of evidence available was carried out. The studies were critically appraised with narrative data synthesis. Results: Twenty-one studies on arthroscopic simulation were identified. Only two studies were randomised controlled trials. The first article demonstrated improved performance of basic knee arthroscopic tasks following a fixed period of training. The second article showed improved performance of arthroscopic tasks and no deterioration in the levels of skills following a period of six months. In addition, the two studies succeeded in demonstrating the importance of 3D motion analysis using computer simulators in the assessment of technical skills. Components of evaluation such as time to task completion, distance travelled by instruments and incidence of instruments collisions were associated with the highest validity and reliability of assessment. This systematic review highlighted the limitations of these two randomised studies. Discussion: Evidence from the two trials suggests that knee arthroscopy simulator training can result in improved performance. This review helped highlight the contribution of the two studies in terms of internal validity and consistency of using arthroscopic skills training. Further level I studies are however required to demonstrate the evidence for transfer and predictive validity of computer simulation as a training instrument.