I suspect that both Cambridge and Oxford get applicants sufficiently good that, having eliminated the definite yesses and the definite noes, admissions could be done randomly and arbitrarily without its effect being very noticeable. But that's just the cynic in me.
The interesting thing to analyse is whether Oxford and Cambridge would consistently pick the same people or if the difference in the system causes systematic differences in who is picked. Since practically nobody can apply for both, there's no way of testing unfortunately.