APBRmetrics

There is some discussion of RMSE and R^2 (in terms of predicting game margins) in my Sloan paper, which is available on the conference website:

http://www.sloansportsconference.com/wp ... thLogo.pdf

The paper is now available on the Sloan website:

http://www.sloansportsconference.com/?p=2798

Yep, anyone who wants the paper (which was presented in 2010, actually) can send me an email at joe underscore sill at yahoo and I'll send the pdf to you. I have asked various people associated with the Sloan conference to put a link to the pdf on the conference page for the paper, without success: ...

Joe Sill is the one who originally developed RAPM: http://www.sloansportsconference.com/?p=2798 . (BTW, Joe -- is the Sloan paper no longer available online, or could I just not find it?) Yeah, at one point the paper was available for download on the site and now it's not- why, I couldn't tell you....

I get Kobe, Dirk, Paul Pierce, Manu, KG as my top 5 for 1-year 2007-2008 RAPM, and that top 5 doesn't change much for different choices of the regularization parameter as long as it's reasonable. If Ronnie Price is ranking number 1, there is probably an error either in the implementation or the data.

Congratulations, John!

I wonder if the advanced stats pages on ESPN will always be known as "Hollinger Player Statistics" and "Hollinger Team Statistics". John has left ESPN, but maybe his brand will live on!

I'm going to try to swing by for a little while early on Thursday night, and I'll probably make it to the Sunday pickup game, too. Let me offer a little pre-game scouting: Mike Z may say he's rusty, but all I know is that last year he was drilling numerous 3-pointers against some pretty decent compe...

Indeed, congrats to Roland (a pioneering force in advanced basketball stats) and the rest of the Mavericks organization. It's great to see a stats guy go from starting a website on his own to traveling with a team and winning a title!

If you're going to predict things on a per-possession basis (or per-game-segment-without-player-substitution basis), then another possibility would be simply to remove those possessions/game segments where rookies played from the evaluation, and only test the models on data where all 10 players on t...