It's almost impossible to define fit, since most MIT applicants have done a fairly good job of researching, and after the initial round of poor applications are thrown out (am I right? I may have forgotten), MIT starts looking for what the blogs terms as "fit."
What the schools can have, however, is diversity within their respective student bodies. Each particular group or organization within MIT has its own demands for particular students. The math team needs a particular group of students, the sports teams need their particular groups of students (Caltech, in this respect, doesn't even care because its sports teams are recruited by those who actually take the time to attend the sports meetings, instead of any previous experience in the sports team per se), the music bands need their particular groups of students. These individual groups may have connections with the admissions office of MIT, and may ask the admissions office to take a certain number of students for their own respective purposes. This is how Ivy League admissions is done, and it's how an institution that wants excellence in everything, so to speak, does admissions.
But it's unlikely that each of the individual groups in MIT, all added together, demand a substantial proportion of the people who've made it past the initial stage. Then MIT goes to other factors to triage the applications into accept/deny, and we just don't know what they are. One could hypothesize that the same application accepted one year may be denied the next year even if his application did not indicate a likelihood to get into one of such individual groups at MIT. Still, I wonder what the 15% statistic really means. It could imply the number of students who made it past the initial cutoff.
That being said, that's unrelated to affirmative action, which is an entirely different issue.
==
Anyways, interesting link I found:
http://www.artofproblemsolving.com/F...c.php?t=140157