BACKGROUND Simulation-based training provides minimal feedback and relies heavily on self-assessment. Research has shown medical trainees are poor self-assessors. The purpose of this study was to examine trainees' ability to self-assess technical skills using a simulation-trainer. METHODS Twenty-one medical students performed 10 repetitions of a simulated task. After each repetition they estimated their time and errors made. These were compared with the simulator data. RESULTS Task time (P < 0.0001) and errors made (P < 0.0001) improved with repetition. Both self-assessment curves reflected their actual performance curves (P < 0.0001). Self-assessment of time did not improve in accuracy (P = 0.26) but error estimation did (P = 0.01) when compared with actual performance. CONCLUSIONS Novices demonstrated improved skill acquisition using simulation. Their estimates of performance and accuracy of error estimation improved with repetition. Clearly, practice enhances technical skill self-assessment. These results support the notion of self-directed skills training and could have significant implications for residency training programs.