i.e. kids doing well "don't need help" until they do is a much more common outcome - which is to say, you're smart enough to do very little work, but never actually learn how to do any work so once that no longer applies the reality-shock is pretty high.
Of course kids who are struggling might really just be stuck on one issue, and once they get past it they'll be fine - i.e. when I was in around Year 2, for whatever reason my reading-level reports for somewhat low. Post that of course, I shot ahead (I have no idea why or how, in that case the metric probably just was mis-measuring whatever my interest was at the time).
Which is an example of exactly the issue with the idea of "gifted kids" - they don't really exist. They're anomalies in measurement methodology, likely to be temporally localized as well. Because everyone knows some "gifted" child, but no one can point to the unique and amazing accomplishments of them as adults.
After all: plenty of gifted mathematicians, certainly people smarter then me. If they're working as mathematicians though, the lifetime earning potential of my average career in IT is going to be higher though.