In this article, the author talks about the release of the most comprehensive study of SAT exams. The headline on the Web site of the College Board, the maker of the test, was, "SAT Studies Show Test's Strength in Predicting College Success." At the same time, a headline on the Web site of the group FairTest, a 23-year-old, nonprofit watchdog organization, presented a rather different conclusion: "College Board Reports Confirm: New SAT is Not Significantly Better Predictor Than Old Exam." In FairTest's view, the old test was not a very useful predictor, either. The original SAT contained only math and verbal sections; the revised test added a writing section in 2005, including a 25-minute essay. Numbers always seem to carry the weight of authority, but if nine major publishers could deem the manuscript of the first Harry Potter book unworthy of publication, then is it reasonable to expect that an SAT rater can reliably make the fine distinctions called for when grading the writing test? In this article, the author argues that the process of grading the SAT's essay portion is so demonstrably subjective that the results are useless.