In their recent article in CHEST (February 2011), Fraser et al1 demonstrated that students who have just listened to a mitral regurgitation on a high-cost, high-fidelity patient manikin simulator (“Harvey”) can recognize mitral regurgitation in a real patient more accurately than can students who heard other abnormal heart sounds on Harvey during training. We wish to issue a cautionary note. The evidence presented does not constitute an endorsement of simulator training, except as compared with no training. If, as we suspect, the students had little or no exposure to murmurs prior to the study (in their 8 h with standardized [healthy] patients and 2 days on the ward), then we might expect the students who heard the simulated mitral regurgitation would diagnose mitral regurgitation the next time they heard any murmur and, consequently, do better than students in the comparison groups who heard some other abnormality. Consistent with this interpretation, students who heard aortic regurgitation did not do significantly better than students who heard no murmur. The authors also describe a pilot study in which students who had practiced on aortic stenosis did worse on a case of mitral regurgitation than students who had heard no abnormalities.