jade_hippo wrote:actually, thats the fallacy of trying to group things together that a inherently different. you can group eggs together and grade them because they are all the same, no 2 basketball players are exactlly the same. welcome to statistics.
What an odd thing to say. This system groups good eggs together with rotten eggs. If it consistently picked good eggs, but some eggs were better than others, then fine. But it doesn't. What's the use of this system in the real world? 70% of the time is selects a player that will be at least decent, but it doesn't select the players in order of quality very well, nor does always select the correct sleepers, nor does it always weed out the busts. A system that can't do anything consistently is a system that does nothing.
he had a decent regular season, something like 14p/4r/5a? (technically he never played college ball now, but oh well) where CDR had like 18p/4r/2a on better percentages, so yes, he does deserve a better ranking than Rose, but relax, no one said he was the better player, just that he had a better ranking. unless you can develop a system that calculates game impact without taking numbers into consideration, this is pretty much all you're going to be stuck with.
I don't really know why we're talking about Rose versus CDR. It's fine for Hollinger's system to miss on Rose. But it's not just Rose at all. I'm talking about the system as a whole. Hollinger is trying to use this system to very specifically pick out players.
As longtime readers will know, I've cooked up my Draft Rater the past few years to present a methodical, objective look at how players' collegiate achievements tend to translate to the pro game. It's based on regression analysis that compares college performance using 27 variables, from the obvious (age, height, likely pro position) to the obscure (say, 3-point attempts per field goal attempt). By looking at what talents have led to success at the pro level, we can figure out some things about what current collegiate draft prospects will be able to do in the NBA.
So he's not ranking them on mere productivity in college, just to indicate who was more productive, and let the rest be evaluated subjectively. No. He attempts to actually predict how good they'll be in the NBA. And the system doesn't work! He pretends that it DOES, yet it doesn't.
Because as you've explicitly said, you can't actually infer from statistics, player quality. Hollinger is saying the exact opposite. And he's full of bull****.
