[dursk]

For any meaningful discussion on this subject to take place, we have to first specify what we mean by "great coders". The skillset required to win programming competitions is very different than what it takes to build maintainable & scalable production systems.

[nilkn]

Mathematics has a similar phenomenon. Not every great performer at the IMO (and similar) becomes a great researcher, although overall they certainly do better than the average. And even the most prodigious IMO performers rarely do any meaningful research until after completing a PhD later in life.

Terence Tao is an example of someone extremely prodigious at math competitions who also ended up being one of the world's best researchers. And he also didn't really do meaningful research until after his PhD.

[credit_guy]

I think "better than the average" is an understatement. I think they do much, much, much better than the average. While not very great IMO performer becomes great researcher (example: a friend of mine with 2 gold medals who didn't finish his Ph.D. because of mental illness), I think that most of them do become good researchers, if they choose a career in research. Of course, many go and join Wall Street or Sillicon Valley, so it's hard to say what their potential could have been.