View previous topic :: View next topic |
Author |
Message |
Eli W
Joined: 01 Feb 2005 Posts: 402
|
Posted: Tue May 17, 2005 11:57 am Post subject: Linear weights calculated from adjusted plus/minus |
|
|
I know that after Dan Rosenbaum developed his adjusted plus/minus statistic, he worked backwards to calculate the linear weights that correlated to adjusted plus/minus. I've been trying to find these weights in order to estimate adjusted plus/minus for this season.
I've tried to use the coefficient estimate numbers from Table 2 on this page - http://www.uncg.edu/bae/people/rosenbaum/NBA/winval2.htm , but the numbers don't seem to be coming out quite right (for instance, should I use eFGA/40 or the square of that?). Can anyone get them to come out correctly? |
|
Back to top |
|
|
Dan Rosenbaum
Joined: 03 Jan 2005 Posts: 541 Location: Greensboro, North Carolina
|
Posted: Tue May 17, 2005 3:29 pm Post subject: Re: Linear weights calculated from adjusted plus/minus |
|
|
John Quincy wrote: | I know that after Dan Rosenbaum developed his adjusted plus/minus statistic, he worked backwards to calculate the linear weights that correlated to adjusted plus/minus. I've been trying to find these weights in order to estimate adjusted plus/minus for this season.
I've tried to use the coefficient estimate numbers from Table 2 on this page - http://www.uncg.edu/bae/people/rosenbaum/NBA/winval2.htm , but the numbers don't seem to be coming out quite right (for instance, should I use eFGA/40 or the square of that?). Can anyone get them to come out correctly? |
There are weights for both eFGA/40 and eFGA/40 squared. To replicate my results, you need use both of them. Also, all of my statistics are pace-adjusted, so if you are very close to what I am getting, then that might be the difference. |
|
Back to top |
|
|
Eli W
Joined: 01 Feb 2005 Posts: 402
|
Posted: Tue May 17, 2005 4:50 pm Post subject: |
|
|
Thanks, that seemed to do the trick. I've managed to use the weights to calculate statistical plus/minus ratings for 2004-05. These are based on the weights in Dan's article, but they are not adjusted for pace (what exactly is the pace adjustment you use, Dan?).
Here are the top ten overall:
1. Andrei Kirilenko
2. Kevin Garnett
3. LeBron James
4. Shaquille O'Neal
5. Dirk Nowitzki
6. Vince Carter (with the Nets)
7. Tim Duncan
8. Brad Miller
9. Larry Hughes
10. Amare Stoudemire
Steve Nash ranks around 70th.
Here are the top rookies, which I found interesting:
1. Andris Biedrins
2. Andre Iguodala
3. Anderson Varejao
4. Devin Harris
5. Delonte West
6. Josh Childress
7. Josh Smith
8. Tony Allen
9. Damien Wilkins
10. Dwight Howard
Ben Gordon and Emeka Okafor rank much lower. |
|
Back to top |
|
|
Ed Küpfer
Joined: 30 Dec 2004 Posts: 785 Location: Toronto
|
Posted: Wed May 18, 2005 12:29 am Post subject: |
|
|
I had my custom pace adjustments at hand when I read this, so I added them to your list. "102" means that during that player's time on the floor, the team played at 102% of the league average possessions/minute (90.3). I'm not sure how the pace adjustment would work, so I'll leave it up to you if you want.
John Quincy wrote: | These are based on the weights in Dan's article, but they are not adjusted for pace (what exactly is the pace adjustment you use, Dan?).
Code: |
Here are the top ten overall:
1. Andrei Kirilenko 96
2. Kevin Garnett 99
3. LeBron James 99
4. Shaquille O'Neal 100
5. Dirk Nowitzki 102
6. Vince Carter 99
7. Tim Duncan 97
8. Brad Miller 103
9. Larry Hughes 104
10. Amare Stoudemire 105
Here are the top rookies, which I found interesting:
1. Andris Biedrins 100
2. Andre Iguodala 105
3. Anderson Varejao 96
4. Devin Harris 104
5. Delonte West 102
6. Josh Childress 101
7. Josh Smith 101
8. Tony Allen 103
9. Damien Wilkins 96
10. Dwight Howard 104 |
|
_________________ ed |
|
Back to top |
|
|
|