While clinical grades are pretty arbitrary, they're still probably the 2nd most impt factor.
1. Step 1
2. Clinical grades
3. LOR (names > content)
4. Research
5. AOA
6. Program reputation
7 et al, Step 2, Pre-clinical grades, ECs etc
You're correct. And thats how the directors themselves admit to evaluating applicants.
This ongoing argument is useless as
most people aren't going to turn down a top 20 school for a random unranked school. Many of the people following this thread will be lucky to get into one school.
There's little to no difference between similarly ranked schools.
The advantage that big name research schools offer is (usually) access to big name faculty in most fields for mentorship and LORs. They also offer access to a large variety of clinical research. For competitive specialties that is pure gold, and will make or break you on the rank list.
PDs at the best programs can evaluate 50 applicants per position. The only way to differentiate applicants at the top of their game is to arbitrarily give additional points for marginal things, like pre clinical grades, reputation, quality of LOR, LOR faculty reputation, success of previous grads they had, etc. Is that LOR an 8 or a 10? They also consider likelihood of the candidate actually ranking the program high enough to match. The interview can be very revealing, but if you're only interviewing ~100 of 600+ applicants, you have to do everything you can to get in the door.
But, again, most people are not facing the (obvious) choice of Penn vs. Eastern Virginia, and most are not candidates for the best residencies in competitive fields. You get in where you get in, eval cost:location:reputation, do your best, and the chips fall where they fall on match day.