The use of computers for cue generation and the safety inherent in simulators permits new “tasks” which are impossible in real flight. We are investigating the potential efficacy of using a particularly striking non-standard task, namely flying at zero-altitude and even below ground. Our hypothesis was that increased experience with optically violent displays during training would improve performance during the testing phase of maintaining low, above ground flight. In the training phase half of the subjects maintained zero altitude while the other half flew as low as possible without crashing. Contrary to our hypothesis the low-trained subjects had a lower standard deviation and mean altitude than the zero-trained subjects in testing. Although our hypothesis was not supported, the concept of exploring novel ways of using flight simulators deserves further attention.