I wonder if Colton has ever compared the Mock compilations to the actual draft results to see how accurate of a predictor the mocks are.
I suspect it holds true for about 6 picks then starts to diverge and is only slightly better than a random guess by 25.
I think I did that in 2011. I can't remember the details of what I found out, but I'm sure your general sense was correct--the early picks on the mock drafts are much more accurate than the later picks.