Computer science educators are increasingly adding components that compare theoretical predictions to empirical results. If we are interested in better integrating Human-Computer Interaction (HCI) concepts into a computer science curriculum, we might look at HCI lessons that draw upon the same set of practices. With this aim in mind, I present a lesson that uses Card, Moran and Newell's Keystroke Level Model and discuss the benefit of asking students to compare its theoretical predictions with empirical results from informal usability tests.