This Ken Pom Thing

[kyleslamb]

Has anyone ever checked whether KenPom's statistics/rankings are actually correlated with higher winning percentages? It's nice that he goes into all this detail, but I think it's bizarre that he's never bothered to show that they work. I know for a fact that plenty of the really exciting new baseball stats are not predictive at all (UZR being a notable offender).

I believe I heard that over a several-year sample, it was correct on 77% of SU picks. Against the spread, it's about 48-49% (which obviously isn't good enough to use religiously for making money ATS).

 

[BuffOrange]

Has anyone ever checked whether KenPom's statistics/rankings are actually correlated with higher winning percentages? It's nice that he goes into all this detail, but I think it's bizarre that he's never bothered to show that they work. I know for a fact that plenty of the really exciting new baseball stats are not predictive at all (UZR being a notable offender).

It's a better predictor than seeds are and the projected scores are usually pretty close to the Vegas lines.