It would be interesting to have a fund which was mostly an index fund, but avoided "losers" based on simple criteria. Picking overpriced losers is easier than picking winners. (I did that for the first dot-com boom, at "downside.com".) That's a concept worth testing against historical data.