Humans in vehicles can sense accelerations and therefore theoretically deduce the path of motion, even without vision available. However, human perception is limited, causing spatial disorientation such as that during the fatal airplane crash of John F. Kennedy Jr. in 1999. Meanwhile, mathematical models have been developed to predict spatial disorientation, and have been used with success in many experimental settings. These models are easy to understand mathematically, using basic concepts such as derivatives. However, recent research has uncovered certain mysteries, in which the models fail to match human perception. For this talk, we look at three mysteries, puzzling about motion in three dimensions.