View previous topic :: View next topic |
Author |
Message |
Jon Nichols
Joined: 18 Aug 2005 Posts: 370
|
|
Back to top |
|
|
back2newbelf
Joined: 21 Jun 2005 Posts: 276
|
Posted: Mon Dec 07, 2009 5:18 am Post subject: |
|
|
link doesn't work for me |
|
Back to top |
|
|
Jon Nichols
Joined: 18 Aug 2005 Posts: 370
|
Posted: Mon Dec 07, 2009 10:36 am Post subject: |
|
|
Which link? They both seem to be working for me... |
|
Back to top |
|
|
schtevie
Joined: 18 Apr 2005 Posts: 414
|
Posted: Mon Dec 07, 2009 11:29 am Post subject: |
|
|
Jon, just thinking out loud here. I can't say I have much intuition on the point, but to the extent that coaches try to match up to avoid being out rebounded at both ends of the court, wouldn't the marginal effect of projected ORR be diminished by not including projected opponent DRR in the regression? Curious to see those results (both the extent to which projected ORR is correlated with projected opponent DRR and the relationship to actual ORR and DRR) if it wouldn't be too much trouble. |
|
Back to top |
|
|
Jon Nichols
Joined: 18 Aug 2005 Posts: 370
|
Posted: Mon Dec 07, 2009 11:46 am Post subject: |
|
|
schtevie wrote: | Jon, just thinking out loud here. I can't say I have much intuition on the point, but to the extent that coaches try to match up to avoid being out rebounded at both ends of the court, wouldn't the marginal effect of projected ORR be diminished by not including projected opponent DRR in the regression? Curious to see those results (both the extent to which projected ORR is correlated with projected opponent DRR and the relationship to actual ORR and DRR) if it wouldn't be too much trouble. |
That definitely makes sense. I'll see what I can do about that. |
|
Back to top |
|
|
cherokee_ACB
Joined: 22 Mar 2006 Posts: 157
|
Posted: Mon Dec 07, 2009 6:05 pm Post subject: |
|
|
schtevie wrote: | Jon, just thinking out loud here. I can't say I have much intuition on the point, but to the extent that coaches try to match up to avoid being out rebounded at both ends of the court, wouldn't the marginal effect of projected ORR be diminished by not including projected opponent DRR in the regression? Curious to see those results (both the extent to which projected ORR is correlated with projected opponent DRR and the relationship to actual ORR and DRR) if it wouldn't be too much trouble. |
Your intuition is correct, although the coefficients do not change much
viewtopic.php?p=19835#19835 |
|
Back to top |
|
|
RChung
Joined: 28 Nov 2009 Posts: 10
|
Posted: Mon Dec 07, 2009 6:56 pm Post subject: |
|
|
Have you considered doing a Tukey MD plot and overlaying a local regression? |
|
Back to top |
|
|
Jon Nichols
Joined: 18 Aug 2005 Posts: 370
|
Posted: Mon Dec 07, 2009 8:55 pm Post subject: |
|
|
Ryan, here is the modified Assists chart using the 0.8 multiplier for each individual assist rate:
With the change, the slope is larger. |
|
Back to top |
|
|
Ryan J. Parker
Joined: 23 Mar 2007 Posts: 711 Location: Raleigh, NC
|
Posted: Mon Dec 07, 2009 9:05 pm Post subject: |
|
|
Cool, thanks. Not a huge difference, but I was expecting to see a larger slope. I think that because of the way assist rate is constructed, there is definitely more data massaging that must be done.
To try and get around this, perhaps look just at assist rate on 3pt shots. It still matters which players are taking the shots, but maybe that gives more insight into what is going on. _________________ I am a basketball geek. |
|
Back to top |
|
|
|