I think people are psychologically averse to acknowledging this level of "filtering" even though it's mostly true, because it suggests an uncomfortable level of

I think people are psychologically averse to acknowledging this level of "filtering" even though it's mostly true, because it suggests an uncomfortable level of inequality. However I do think that there might be just a little too much focus on contest-math when it comes to identifying the country's best mathematicians. It's easily seen that the best contest mathematicians usually become the best professional mathematicians. In aggregate, though, most contest-math participants come from a very particular demographic (i.e. they live in big cities in wealthy areas, have very supportive/tiger parents (this is the key), go to magnet schools, obviously practice a lot, etc.). These people are still probably much more likely than average to become professionals, but it filters out a lot of bright people from less wealthy families or areas, or in regions where culturally math competitions are barely even a thing.