A couple of thoughts (this is one of the things I study):
1. McGill is a very good university. Some would say the second best in Canada (to U Toronto), but that depends on how you define "good." Is it acceptance rate, college entrance exam scores, Nobel Laureates, research funding, and so on. My main job at my university right now is helping to get the highest possible score on the every-six-year national research ratings. The stuff that places are doing to game the system is pretty amazing (we are not, btw). You see those QS ratings? I'm one of the raters for that. I'm not allowed to rate my university, but I certainly know who our competition is. Same thing when I used to do US News and World Report ratings when I was at Rutgers. There are folks who call up friends at other universities and agree to a "you wash my hand" sort of deal. So two big problems with ratings: one, it is arbitrary and depends upon what you pick to look at and how much to count it; two is that the universities work to make themselves look as good as possible (Why do you think universities use the highest score you have in each of the three areas if you take the test more than once? Hint: It ain't because they are being nice to you.)
2. RU has traditionally used class rank more than SAT scores because it results in a more diverse university. Since everybody has a valedictorian, and since NJ schools are highly segregated, this works to the University's advantage in getting diverse classes. Anybody who thinks that the SC rulings have gotten rid of universities taking race and ethnicity into account --- well, there is this bridge in Brooklyn. The Cal universities have come the closest to doing this (hence huge numbers of Asian kids).
3. The kid who got turned down with the perfect scores isn't telling the full story. I have no idea what that story is, but if the rest of his record looked like his academic record, he would have been accepted to at least a couple of Ivies. My guess? One of his teachers who he used for a recommendation said something along the lines of "His focus on academics is admirable, but ...." and then a subtle knife went in his back. Doesn't take much, and Ivy admissions folks are great at picking up small signals. If all of the elites turned him down, there's a spot on his blotter somewhere.
4. Universities try to meet several goals in admissions. One is that they want the best classes possible in terms of likelihood of academic success. Two is that they want a diverse student body -- in terms of ethnic background, economic background, field of study, getting some good athletes (in non-revenue generating sports, etc.). The Ivies look at each application individually. RU pretty much mass processes them with a formula using class rank and test scores, with class rank more important (unless they have changed recently).
5. The winner on the thread in my view is Hudson, with the idea of "Why don't we take kids who are really likely to develop strong ties with the University?" I know of no school that actively takes that into account, although some will say, "We want (insert school name) people to come here," whatever that means. The question then becomes, "How do we find such people in the admissions process?" Is loyalty something you can select for, or is it developed while at the university? Super interesting idea leading to some interesting questions. This would be studiable!
EDIT addition: On studying for the SATs. The benefit of studying in a fashion that the overwhelming majority of kids might do will in fact, get you somewhere around 50-100 points per section (of the three sections). BUT, that might make the difference in an application. I work with kids on test-taking skills, and some kids are stunningly bad at this. Thus, I can usually help those kids a lot. The tests are timed, so you have to not only be able to work, you have to work at pace and strategically. Can you go up from 1760 to 2400? It's possible, but my guess is that the 1760 was an aberration, and we don't know how many tries it took to get the 2400. So why do you see test prep companies claiming massive gains? Because of the regression to the mean effect. My boy did super well the first time he took the SATs. I asked him if he wanted to take them a second time. His response: "I'm booking these and going to the Mets game the next time the test is given." My daughter's first scores were disappointing to her. So we worked together, and the second time she did much better. In part, that was because she had a bad day the first time she took them and a good day the second. The kids who go to test prep school are much more likely to have had a negative error (technical term for a bad day) made on their test the first time they took it. Kids who are super pleased with their scores don't go to test prep school (typically). Thus, when the test prep schools say their average gain is 4X and when ETS says in their studies, the average gain is only X, they are both telling the truth. ETS uses experimental designs that control for the regression effect; prep schools use whoever walks in the door.
