In this simulator study eleven pilots rated their motion perception during a series of decrab maneuvers of a twin-engine passenger aircraft. Platform yaw, sway, and roll motion were varied independently to examine their relative contribution to the pilots' judgements. In one set of conditions, the washout algorithms were bypassed so as to reproduce unfiltered aircraft motion. This was compared with washout-filtered motion in another set of conditions. Moreover, the effect of visual cues was studied by testing the unfiltered motion cues once under simulated VMC, and once under IMC. The results show that the simulation of heading alignment was positively affected by platform sway and roll, and also by the visual stimulus. Platform yaw was poorly recognized, and remained under the perceptual threshold in the presence of platform sway. Interestingly, unfiltered sway motion was perceived as too strong, even though the simulator workspace required downscaling to 70% of the actual aircraft motion. Finally, the subjective data was used to validate our human motion perception model. Due to the fact that the model does not yet account for the observed interaction between sway motion and the perceptual threshold for yaw, the model output did not quantitatively correlate with the magnitude ratings. However, a multiple regression analysis showed that, qualitatively, the model did predict the way pilots interpreted the platform motion. We conclude that better understanding of perceptual thresholds in a flight simulator setting is necessary to enable quantitative analysis of the effectiveness of ground-based motion cues.