I would be interested to learn of anyone's experience using mastery-based (or "mastery-level") grading in a Theory of Computation course. Usually this requires—at a minimum— a detailed taxonomy of competencies, and what it means to demonstrate mastery of each competency.