I thought it just might be modestly more accurate. So for example if someone had a 300% return one contest and finished first their relative performance would be underestimated by pure average rank.
I'll take whatever you want to give as I am sure compiling the numbers is a PITA and I do appreciate the effort.