APBRmetrics Forum Index APBRmetrics
The statistical revolution will not be televised.
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

Questions related to adjusted +/- ratings
Goto page Previous  1, 2
 
Post new topic   Reply to topic    APBRmetrics Forum Index -> General discussion
View previous topic :: View next topic  
Author Message
Ben



Joined: 13 Jan 2005
Posts: 202
Location: Iowa City

PostPosted: Thu Sep 20, 2007 12:35 pm    Post subject: Reply with quote

John Quincy wrote:


Beyond that, there's the problem with the magnitudes that I mentioned previously. I'm not sure why those are off by so much, but that's surely part of the reason the team totals don't look right. (To see some actual magnitudes for statistical plus/minus that Dan found, see here - http://www.uncg.edu/bae/people/rosenbaum/NBA/wv2t3.txt )


Judging from the variable names and the magnitude, I would venture to guess that those are suppose to be wins contributed above an average player assuming they play 82 games and 40 minutes.
Back to top
View user's profile Send private message
Mountain



Joined: 13 Mar 2007
Posts: 216

PostPosted: Fri Sep 21, 2007 5:00 pm    Post subject: Reply with quote

I said earlier if overall +/- is considered better, lets go for making it available in public. But it may not be to everyone. Pure +/- is still a useful metric to see as a companion to any individual performance based metric.

Blending the two in overall +/- is something I am comfortable with and a supporter of in general. There are those who haven't been public supporters of other blending metric efforts, but overall +/- with statistical +/- and pure +/- is a hybrid. I hope simply weighted products of several metrics and conceivable meta-metrics with new elements get a full and fair review in performance studies- soon or some day. Win Produced can be called a hybrid too for the way it incorporates team based defensive data into the rating. Will any other hybrid metric be rated in the upcoming conference report?

What is the best metric at explaining the past or more importantly the future, pure or blended? The only way to know is to develop and comparably score the explanatory and predictive power of everything rather than effectively settling on one metric because it is the best currently available among a select list of methods with wide publicity. It shouldn't be the endpoint, unless it truly can beat all comers. If anyone else wants to fairly score new contenders against existing that would be great.
Back to top
View user's profile Send private message
Mountain



Joined: 13 Mar 2007
Posts: 216

PostPosted: Sun Sep 23, 2007 12:38 pm    Post subject: Reply with quote

There is some relevant year to year correlation data here if any are interested in that point:

http://sonicscentral.com/apbrmetrics/viewtopic.php?t=877&postdays=0&postorder=asc&start=195
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    APBRmetrics Forum Index -> General discussion All times are GMT - 5 Hours
Goto page Previous  1, 2
Page 2 of 2

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group