It's not completely useful, nor is it completely useless. As long as it is possible to be a "good" investor, in the sense of making lots of money, it can still be somewhat useful. If you continue to bet on people who made successful investments in the past, you might get burned by the people who merely got lucky, and you will make money when the person really was a smart investor. Past performance thing is not a convincing argument against this. A valid argument would be that the ratio of those groups is not favorable enough to overcome the management fee. I'd be interested to see if the Buffet bet you mention is related to this, but my work filter blocks that site.