APBRmetrics

Anybody put a set out yet (I know gotbuckets.com was hosting talkingpracticeblog's data last year, but the last updates were the season previews, and their Twitter is dead as well)? Would love to see how some of the top guys stack up.

Anthony Davis in particular is 3rd in the league in J.E.'s RPM:

http://espn.go.com/nba/statistics/rpm/_/sort/RPM

so I wonder how much if it is the box score component, and how much is him doing so well this year that he's distanced himself from his prior even in vanilla RAPM?

Do you want single year or multiyear RAPM?

J.E. wrote:Do you want single year or multiyear RAPM?

Multiyear I suppose, thanks.

J.E. wrote:http://stats-for-the-nba.appspot.com/ra ... _2015.html

Awesome, thanks. Appears that Davis looks pretty good here as well.

J.E. wrote:http://stats-for-the-nba.appspot.com/ra ... _2015.html

wait are you updating these again? that's awesome. Thought I read in another thread that these weren't going to get updated anymore.

I just realized that the above link was not actually multi-year RAPM, but multi-year RPM

I apologize for the confusion.

You can now find pure multi-year RAPM at that same link http://stats-for-the-nba.appspot.com/ra ... _2015.html
What I posted originally - in case you want to still look at it - can be found at http://stats-for-the-nba.appspot.com/ra ... _2015.html

J.E., when you say multi-year, how many years are we talking about? How much does that the prior years have influence in this RAPM in comparison to the current year? Do prior years have more influence than this season?

J.E. wrote:I just realized that the above link was not actually multi-year RAPM, but multi-year RPM

I apologize for the confusion.

You can now find pure multi-year RAPM at that same link http://stats-for-the-nba.appspot.com/ra ... _2015.html
What I posted originally - in case you want to still look at it - can be found at http://stats-for-the-nba.appspot.com/ra ... _2015.html

Ah, no problem. Thanks for sharing. Also if it's not too much trouble, do you have NPI vanilla RAPM for this season as well?

Any way you can do a 1.5 year RAPM? 2014 and 2015 up to the break, of course.

Just wrote up a "tutorial" on how to do RAPM using scikit-learn (using 2014-15 data only):

http://nbviewer.ipython.org/gist/EvanZ/ ... eb14f28d58

EvanZ wrote:Just wrote up a "tutorial" on how to do RAPM using scikit-learn (using 2014-15 data only):

http://nbviewer.ipython.org/gist/EvanZ/ ... eb14f28d58

It's weird how Manu ranks 2nd with just a +5.9 Net plus/minus. It seems like every version of RAPM loves Manu for some reason.

AcrossTheCourt wrote:Any way you can do a 1.5 year RAPM? 2014 and 2015 up to the break, of course.

It was easy enough for me to run this. This is just based on the past 1.5 years:

https://docs.google.com/spreadsheets/d/ ... 1655629432

EvanZ wrote:Just wrote up a "tutorial" on how to do RAPM using scikit-learn (using 2014-15 data only):

http://nbviewer.ipython.org/gist/EvanZ/ ... eb14f28d58

could you (or would you mind if I) post this in a new thread so it can get the attention it deserves?

Thanks for your contribution.

ampersand5 wrote:

EvanZ wrote:Just wrote up a "tutorial" on how to do RAPM using scikit-learn (using 2014-15 data only):

http://nbviewer.ipython.org/gist/EvanZ/ ... eb14f28d58

could you (or would you mind if I) post this in a new thread so it can get the attention it deserves?

Thanks for your contribution.

Sure, go ahead. Could start a nice discussion I suppose.