Ricardian.
'Assume, and I'm not sure I think this is true, but arguendo, that there is some reified general intelligence that (a) predicts academic ability and (b) can be directly measured. This is the "let's pretend Cyril Burt wasn't a lying charlatan" position. That would mean that we could compare on this scale, which we may as well call g, two eighteen year olds, one of whom had been educated at Eton, the other of whom had been raised by wolves in the manner of Romulus and Remus. Suppose we determine that on our measure of intelligence, the latter is "more intelligent". How much more "intelligent" would they need to be in order to have a better chance of success in an MBChB programme than the former?
My answer is "they're disproportionate likely to fail, given realistic assumptions". Because the best predictor of success in exams is success in exams, and the idea that universities can fix in 66 weeks of teaching (in the case of the typical 3 year level 6 qualification) a deficit which has occurred over the 14 preceding years is farcical. Or, more crucially, can fix in the 22 weeks of the first year those self-same 14 years, such that the student passes the first year exams. BTECs provide Level 3 qualifications for people who can't take exams, but when they arrive in an environment where passing exams is the sine qua non of progression, they fail.'
The above is very much a reductio ad extremum argument.
I would say that schools like Eton are fantastically good at teaching students to dot the i's and cross the t's, to make sure an able but far from brilliant candidate can access the higher grades. I would say that an AABB from an ordinary school is probably the equivalent of AAAA from Eton, and that the former candidate would, on average, fare at least equally well in a university setting.
The reality is that, given small grade differences, it is maybe two years of teaching that need fixing, not 14. That can easily done in 66 weeks, assuming the candidate is motivated to do so. When I was at Cambridge (many years ago), I do remember the majority of the candidates from rough schools who had very ordinary A levels did just as well as those from the best private schools with the normal requirements (all A's as was then and a couple of S levels, or admittance via the entrance exam).
No, you cannot get there with a candidate raised by wolves, regardless of the 'g' score, but you can with carefully selected candidates with slightly worse results at A level.