Been some ups and downs in the early season since I started this thread. Anyone changed their mind either way?
Anyway, I really bumped the thread because I never got back here and broke down the Hollinger preview like I had wanted to, but I figure better late than never:
Hollinger made his season preview predictions (54 wins for the Cs?!?) for a lot of sketchy reasons, including his PER stat. But his in-season power rankings are tied almost entirely to the team scoring margin. And that, right there, is what made Hollinger's preseason prediction idiotic.
Because PER was essentially something that Hollinger just came up with. There was no science involved, he just played with a lot of box score stats until he found a formula that fit his notions of what a good player is. If, historically, there is any predictive value in PER it is purely a side effect.
On the other hand, the team scoring margin as a predictor of team quality has been proven and re-proven by multiple different sources. Hollinger has his version, but so do others like Berri and Sagarin.
And by looking at their history of team scoring margin, it was idiotic to expect the Celtics to suddenly fall off the board barring significantly more injury. To publish a prediction based purely upon a random stat and some very qualitative "analysis" when you have a lot of data in your possession from a better stat that gives more realistic results...that was just dumb.
I wanted to type this message out before Hollinger's new power rankings started coming out because it was a lead-pipe lock that the Cs would spend most of the year on top of that ranking, but alas I am late and the Cs have already taken a perch at the top. Principle still applies, though...your stat might have the Cs on top, Hollinger, but as for you being in agreement with it all along...no you (freakin) wasn't!