We evaluated a virtual reality shoulder arthroscopy simulator using a standardized skills-assessment algorithm in 3 specific groups with various degrees of surgical expertise. The simulator (Mentice Corp, Gothenberg, Sweden) consists of a computer-based, dual-force feedback system with video monitor. Modeled structures include cartilage, labrum, ligaments, biceps tendon, and rotator cuff. The study included 3 groups of volunteers: group 1, medical students interviewing for orthopaedic residency (n = 35); group 2, orthopaedic residents interviewing for sports medicine fellowship (n = 22); and group 3, experienced faculty at a shoulder surgery course (n = 21). Data were collected anonymously and subjects completed a standardized test protocol designed to assess accuracy and efficiency. Subjects used the probe to "touch" a sphere that appeared at various locations within the joint (11 positions total). The sphere changed location immediately on contact with the tip of the probe. The following parameters were calculated by the computer: time (from touching the first ball until touching the eleventh ball), path ratio (percent of measured path length relative to the ideal path), collisions (number of times the probe / arthroscope contacted any tissue), and injuries (collisions beyond a threshold force). Test time and path ratio differed significantly as a function of surgical experience. There was no significant difference in probe collisions between the groups. Arthroscope collisions and injuries averaged 2 or less in all of the groups. There was significant correlation between path ratio and time to complete the test in groups 1 and 2 (r =.527 and r =.827, respectively, P <.001), but not in group 3 (r =.376, P =.10). There was essentially normal distribution of time performance in groups 1 and 2. Time was shorter and more consistent in group 3, suggesting greater consistency in the experienced surgeons. These data suggest that this arthroscopy simulator facilitates discrimination of arthroscopic skills. Computer-based simulation technology provides a major opportunity for surgical skills development without morbidity and operating room inefficiency.