[eas]

The only problem is that these funds don't start out trying to be correlated with each other--the most successful strategies are those that no one else is doing, since you, tautologically, get the best prices on broadly undervalued assets. They don't want to hold all the same things, but finance is a small world at people catch on to successful strategies pretty quickly. It just pans out that way since they are built on analyzing historical relationships between assets and what has worked in the past, e.g., buy refiners when the crack spread widens, or high p/e/g ratio stocks. To be independent you'd have to either find that relationship no one knows about, or pick a strategy that hasn't worked in the past.

To put it in web terms, it's almost like these quant funds are just adapting to industry best practices. We see that people like social networking, crowdsourcing, "web 2.0" page layouts, etc., so we see lots of sites racing to add these features--because hey, that's what works, that's what users want. But what happens if people get sick of one of those features (or any other you pick)? The relationship between the market and that feature breaks down. All those sites who counted on that strategy will all fail (or adapt) at around the same time, and many will rush into the next hot area (Pointcast-style "push" technology, anyone?).

Quant funds are basically just advanced machine-learners; you could implement a black box-of-sorts on your website by looking at Comscore numbers/trends for different sites and plotting that out against the features/layouts/topics they use, and instantly adding some new widget to your sidebar or something when you see a positive relationship with traffic generation. (Techmeme is a great example of a web black box, BTW, always on the hottest tech trend.)