Foldit is a game where you try to maximize the stability of a protein by pulling it, twisting the orientation of its amino acid sidechains, and wiggling it. Most of that work can be automated, leaving the actual gameplay to work through trial and error. The idea is that, eventually, games like this could harness the power of humanity to solve problems that can't be solved by brute force computation alone. It's a good idea, but I'm not sure about the execution.
The main problem is that the game's tutorial isn't very clear; it wasn't very clear to me what I was doing most of the time (and I've taken classes on proteins!). It holds you by the hand and tells you what to do pretty exactly, guiding you through various puzzles, but because you have no idea what you are doing you blindly follow the instructions and are unlikely to be able to apply the principles to other situations. But at least it has a tutorial.
Aside from the tutorial, there are some competitions, where you play against other players to get the highest score. This part is slightly more fun, but again, mainly trial and error, with no real strategy except to nudge and pull a protein here or there and wiggle it and see if your score goes up. But, on the other hand, at least this strikes me as more reflective of problems in real life than the artificial problems in most games.
The game's a fairly large download, so I'd only recommend it to those who are sure that pulling strands of amino acids and watching them move around in 3D is something they want to try out. It does get oddly addictive after a while.