With its continued evolution, arthroscopic simulation has been increasingly recognized as a means for surgical skill training within graduate medical education. The purpose of this study was to ascertain whether a standardized arthroscopic simulation curriculum improved diagnostic shoulder arthroscopy operative performance, efficiency, and patient safety among orthopaedic trainees. Twenty-two orthopaedic surgery residents with varying levels of shoulder arthroscopy experience were consented and randomized into two groups, simulator (SIM) and standard practice (SP). All subjects were oriented to a virtual reality shoulder simulator, an anatomic checklist, and the Arthroscopic Surgery Skill Evaluation Tool (ASSET), a validated metric for assessing technical ability during arthroscopy. Video recording of expert diagnostic shoulder arthroscopy and relevant errors were also demonstrated, and individual case logs were queried to identify cumulative exposure to shoulder arthroscopy. At baseline testing, the attending surgeon established standard arthroscopic portals and all participants performed supervised, in vivo diagnostic arthroscopy with video recording, as well as the simulator-based diagnostic arthroscopy test. The SIM group subsequently received four 15-minute simulator-training sessions conducted by a single instructor over a 3-month period, while the SP group received routine arthroscopic exposure without further simulator training. After intervention, both groups were re-evaluated with supervised diagnostic shoulder arthroscopy and simulation testing. Two blinded, independent experts evaluated arthroscopic performance utilizing the 14-point checklist, ASSET score, and total time elapsed. Outcome measures were evaluated using the student t-test. Multivariate analysis was conducted to correlate simulator performance with number of arthroscopic cases. Following the intervention, subjects randomized to the SIM group had significantly better safety scores (p=0.005) during the diagnostic arthroscopy and performed better on the ASSET (p=0.061) when compared to the SP group; however, the latter difference only approached statistical significance. The SIM group also completed the simulation testing significantly faster (p=0.001) and was much more efficient in using the probe (p=0.001) during the simulation test following the intervention when compared to the SP group. Subjects in the SIM group were also more efficient in using the camera following the intervention when compared to the SP group; however, the difference between the groups was not statistically significant (p=0.070). Both groups demonstrated significant improvements on the ASSET during the diagnostic arthroscopy following the intervention when compared to their pre-intervention baseline assessments. In addition, the SIM group also performed the diagnostic arthroscopy significantly faster following the intervention (p=0.026). Finally, the SIM group demonstrated significant improvements from pre intervention to post intervention for the time required to complete the simulator assessment (p=0.030) and both probe (p=0.040) and camera (0.049) efficiency during the task, while none of the improvements observed in the SP over time were statistically significant for the simulator task assessment. We demonstrated that our SIM group was able to not only perform the designated tasks faster but also while demonstrating improved patient safety following high fidelity simulation training. To the best of our knowledge, the current study is the first to establish transfer validity, meaning surgical skills learned in a simulated environment can be reflected in the operating room theater.