Completely disagree that 'most of the time, the rankings pan out'. If that were true, the majority of five star players would become all Americans and NBA draft picks, which is not the case. I do agree that in aggregate the rankings are statistically meaningful in that the likelihood of picking a random five star player and having that player become an all conference or AA is much higher than if you were to pick a random three star player.
The problem with your statement and your view of recruiting in general is that it doesn't allow for asymmetric information, i.e. it assumes that no one has a better source of information than the recruiting rankings. While that's generally true for us as fans, Matt Painter has shown that his 'Plan A' players consistently perform at a high level, regardless of ranking. I agree that having Matt target and land more highly ranked players is a positive development, but assuming that a random four of five star player is going to be better than a guy that MP tabs a Plan A option is misguided.