BACKGROUND: Virtual reality simulation is effective in training the novice to perform basic laparoscopic skills. METHODS: Using the Minimally Invasive Surgery Training--Virtual Reality (MIST-VR) trainer, 27 honors high school students were tested at the easy level, prospectively randomized to eight training sessions at the easy (group A, n = 14) or medium (group B, n = 13) level, then retested at the easy level. RESULTS: Both groups were statistically similar at baseline. All scores improved significantly (50.1% to 81.3%) over the period of training (p < 0.05). Although the group A scores were significantly better than the group B scores throughout training (p < 0.05), on final testing at the easy level, group B surpassed group A for all the tasks except TransferPlace (p = 0.054). CONCLUSIONS: Virtual simulation is an effective laparoscopic training method for the novice, providing significant improvement in skill levels over a relatively short period. More challenging training seems to predict greater improvement over time and better final skill levels.