What is far more likely is that Team A will build a good site and Team B will build a good site. Team A's will be slightly more springy to the touch and a litt

What is far more likely is that Team A will build a good site and Team B will build a good site. Team A's will be slightly more springy to the touch and a little more pizzazz. So, you pick Team A. Is the fact that you picked it a useful predictor of whether Team A or Team B would have succeeded in the marketplace, and to what extent? The usefulness of this type of selection is about 0%, and has the dramatically impeding side-effect of wasting Team B's time. Sure, some Y Combinator committee may think Team A's is better, but then, instead of dicking around with the contest, Team B should have just built the product anyway and gone off to sell it on their own. Time is money, and such "contests" assume that programmers' time isn't valuable, like it's a fancy form of relatively slow typing.