LOL @ Hollinger

[polandprzem]

I'm sure timvp can dunk on that basket

[bdubya]

Another mistake that probably explains these results is that Hollinger used an equation with coefficients fitted to stats over a complete 82-game season.

Ah. That rings true.

I can't help wondering - if he inputs the stats from each of the past several 82-game years, does it reproduce the corresponding playoff results? Or does it give you a fourth Lakers le followed by a Suns three-peat, or what? I want to see this done.

[Purple & Gold]

Hollinger and his PER is

[RonMexico]

I wish he would provide these stats and just say "here they are" instead of trying to justify them and provide in-depth analysis.

Pretty much like his rookie PER - Big Baby has probably played like 12 mins in garbage time and gotten a good chunk of reboundsa and layups, so of course he'll have a higher PER than Horford or Durant... but would you say he's rookie of the year? The only PER's that are indicative of success right now are LeBron's and Manu's...

[mathbzh]

The only PER's that are indicative of success right now are LeBron's and Manu's...

Not true his PER ranking is as good as any number can be to judge players... at least for starters playing enough minutes.

[ShoogarBear]

Not true his PER ranking is as good as any number can be to judge players... at least for starters playing enough minutes.

Here we go again. "As good as any number can be".

Do you have a shred of data (not opinion) to back that statement up?

When I fit a model to data, I can produce goodness-of-fit statistics to ascertain whether it is better or worse than another model. What's the goodness-of-fit for PER?