Critics of charter schools in New York City, America's largest school district, often allege that charters score better on standardized tests, on average, than traditional public schools because charters "cream-skim" (i.e., attract) the brightest, most motivated, students. Yet this accusation neglects the fact that not all traditional public schools are open admission. Some, such as Stuyvesant High School and the Bronx High School of Science, use entrance exams and other demanding criteria to recruit the best students. New York is also home to 98 traditional selective public middle schools. If admitting students with better academic credentials were the key to better test scores, New York's traditional selective public schools (which pursue such practices) would easily outperform its charter schools (whose students are admitted by lottery and are more likely to be poor and nonwhite). But this is not the case. This report compares aggregate test scores in math and English in New York's 73 charter middle schools with those of its 98 traditional selective middle schools. (There are no traditional selective elementary schools, and state exams in math and English are not administered after eighth grade, the final year of middle school.) It finds that students at charters score equally well in math as (though worse in English than) students in traditional selective schools. However, when an apples-to-apples comparison is made by comparing students only from similar racial and socioeconomic backgrounds, charters shine more brightly: their students score better in math than, and just as well in English as, those in traditional selective schools.