It is almost universally agreed that more frequent formative assessment (homework, clicker questions, practice tests, etc.) leads to better student performance and generally better course evaluations. There is, however, only anecdotal evidence that the same would be true for more frequent summative assessment (exams). There maybe many arguments against giving more exams, including the general "pain" associated with examinations, as well as reduced teaching time, since classroom sessions are dedicated to exams rather than lecturing. We present evidence that increasing the number of exams in fact does lead to better learning success, less cheating and guessing on homework, and better student course evaluations. (Contains 6 figures.)