I'm pretty pleased with the results of USAMO expansion. When we pushed for it, some naysayers opposed on the grounds that the students wouldn't be ready for the USAMO and it would discourage them. Kiran Kedlaya plotted AMC indices against USAMO scores for 2005, and we found that if we cut out the top 12 and the state qualifiers, there was nearly no correlation. So there was evidence that at the AMC index margins there were plenty of students ready for the USAMO.

We saw in the stats this year that there are still a lot of students taking the test and getting 0-3. But the percentage of students in that range wasn't nearly as high as might be expected, particularly given that students near that AIME edge didn't find out about the expansion until late in the game (and thereby had little incentive to shift their year-long study habits to include more exposure to Olympiad problems). I imagine that next year, many more of the students near that margin will be much more well prepared than this year, if only because they now have an extra problem worth of wiggle room to play with on the AIME, and can thereby spend more time studying Olympiad-level problems (which I bet will help them with their AIME anyway).

Speaking of that AIME problem wiggle room, I understand from the AMC that there are a significant number of MOPpers who have that extra problem of wiggle room to thank for qualifying for the USAMO. (A number of those are avid AoPSers, too, so we're happy!)

In sum, I think the expansion has been a success, and I'm excited to see what next year will bring. (Moreover, hopefully more students being interested in olympiad-style problem solving will also cause more olympiad-style contests to pop up, so we won't have so many students taking the USAMO not understanding what's required for a rigorous proof, and getting 1's on problems they think they have a 7 on. At the very least, there are programs to help students learn how to avoid this fate )