An extensive database of exam scores is studied to determine the effects of a grading policy that drops the lowest exam score. We find evidence that some students engage in strategic behavior, either by understudying for one of the exams or missing one altogether, but the vast majority of students show no evidence of strategic behavior. We also find evidence that many students "satisfice", showing how a large percentage of students passed up an expected improvement in their course grade. We find that the probability a student will choose to complete an optional final exam is inversely related to their grade going into the final. Further, the likelihood of a student completing the final exam rises with the spread between prior exam scores and falls with the points needed to raise their course grade.