I've tried to compare it more accurately.

It's very hard to evaluate selectivity. It's not just the raw number of people participating. It seems that large majority of serious ACM ICPC participants (both contestants and their coaches) are practising on Topcoder, and for the ICPC the best college CS students are recruited much the same as best highschool math students for IMO.

I don't know if Linus Torvalds would necessarily do great on this sort of thing - his talents are primarily within software design, and his persistence as the unifying force behind Linux. (And are you sure you'd recruit a 22 years old Linus Torvalds who just started writing a Unix clone?). It's also the case that 'programming contest' is a bit of misnomer - the winning is primarily about applied mathematics - just as 'computer science' is a misnomer.

In any case, its highly dubious that understanding of QM sequence is as selective as any contest. I get it fully that Copenhagen is clunky whereas MWI doesn't have the collapse, and that collapse fits in very badly. That's not at all the issue. However badly something fits, you can only throw it away when you figured out how to do without it. Also, commonly, the wavefunction, the collapse, and other internals, are seen as mechanisms of prediction which may, or may not, have anything to do with "how universe does it" (even if the question of "how universe does it" is meaningful, it may still be the case that internals of the theory have nothing to do with that, as the internals are massively based upon our convenience). And worse still, MWI is in many very important ways lacking.

Selectivity, in the relevant sense, is more than just a question of how many people are granted something.

Of course. There's the number of potential participants, self selection, and so on.

How many people are not on that site, but could rank highly if they chose to try? I'm guessing it's far more than the number of people who have never taken part in the IMO, but who could get a gold medal if they did.

IMO is a highschool event, and 'taking part' in terms of actually winning entails a lot of very specific training instead of education.

(The IMO is more prestigious among mathematicians than topcoder is among programmers. And countries actively recruit their best mathematicians for the IMO. Nobody in the Finnish government thought it would be a good idea to convince and train Linus Torvalds to take part in an internet programming competition, so I doubt Linus Torvalds is on topcoder.)

Nobody can recruit Grigori Perelman for IMO, either.

There's ACM ICPC, which is roughly the programming equivalent of IMO . Finalists have huge overlap with TC. edit: more current . Of course, TC lacks the prestige of ACM ICPC , but on the other hand it is not a school event.

There certainly are things as selective or more than the IMO (for example, the Fields medal), but I don't think topcoder is one of them, and I'm not convinced about "plenty". (Plenty for what purpose?)

Plenty for the purpose of coming across that volume of technical brilliance and noting and elevating it to its rightful place by now. Less facetiously: a lot of people know everything that was presented in the QM paper, and of those pretty much everyone either considers MWI to be an open question, an irrelevant question, or the like.

edit: made clearer with quotations.