Ahh, that is a good direction for the site to go in I think.

My thinking is that YC's problem is mainly figuring out how likely people are to MSPW and how fun they will be to work with.

I am trying to think out how the co-founder problem differs from the YC problem, but I'm not sure which assumptions to use:

1) The productivity of the startup is mainly the sum of the traits of each individual. Plus, to a lesser extent, the synergies between them.

2) The productivity of the startup is mainly an emergent phenomenon that is difficult to predict from looking at either of the two co-founders alone. Plus, to a lessor extent, the traits of each individual. (For example, the output of someone working with Steve Jobs is different than their output in other circumstances.)

Whether it's more of case 1 or more of case 2 matters when figuring out in which direction to take the community, because in each case you are going to want a different systemic design to bring out the different salient factors. (Of course it could actually be case 1 for some startups and case 2 for other startups, which would make the problem even more difficult.)

"1) The productivity of the startup is mainly the sum of the traits of each individual. Plus, to a lesser extent, the synergies between them."

I would argue the synergies between cofounders to be the single most valuable entity in a partnership. You may have complementary skills but if you don't come together and are able to do incredible things that are spurred on by eachother. Spark is key - this linear trait argument isn't correct at all. Never compromise without spark.

You don't know by chatting with someone - best to work on something non trivial together and figure out if you get on.

This seems like it could be an interesting feature idea -- maybe I could include a way for people to solve a problem together after they introduce themselves?