> In other words, if all of the managers of the 2,862 funds hadn’t bothered to try to pick stocks at all — if they had merely flipped coins — they would, as a g

> In other words, if all of the managers of the 2,862 funds hadn’t bothered to try to pick stocks at all — if they had merely flipped coins — they would, as a group, probably have produced better numbers. Instead of two funds at the end of five years, basic probability theory tells us there should have been three. Um. You are trying to tell me that 2 instead of 3 is enough to reject the null hypothesis? I rather doubt it.