We apply a new bootstrap statistical technique to examine the performance of the U.S. open-end, domestic equity mutual fund industry over the 1975 to 2002 period. A bootstrap approach is necessary because the cross section of mutual fund alphas has a complex nonnormal distribution due to heterogeneous risk-taking by funds as well as nonnormalities in individual fund alpha distributions. Our bootstrap approach uncovers findings that differ from many past studies. Specifically, we find that a sizable minority of managers pick stocks well enough to more than cover their costs. Moreover, the superior alphas of these managers persist.