This link, containing the following screenshot. If interested, there's a ton of stuff on the web about game use for solving protein folding. Consider - could the same social networking approach, a game to collectively produce a productive algorithm work for beating the point spreads every NFL Sunday? It's a less structured problem, with greater noise, more stochastic, but it has a social value payoff if successful. Beating the Vegas bookies is a challenge, they are professionals, but those propagating the folding experiment are also skilled professionals.
If you want to look around, the lead author of the PNAS item is David Baker, at U. Washington, and some search return lists to consider are: here, here, here and here. Then one hit leads to another, etc.
______________UPDATE______________
A brief general interest introductory review, from Europe, and mentioning Foldit, is here in html, here as a pdf.
FURTHER: Baker lab, here.