|
APBRmetrics The statistical revolution will not be televised.
|
View previous topic :: View next topic |
Author |
Message |
Albert Lyu
Joined: 20 Oct 2009 Posts: 2 Location: Evanston, IL
|
Posted: Sat Aug 14, 2010 5:47 am Post subject: The Value Of A Blocked Shot |
|
|
Hey all, I've been a lurker in this forum for a few years now, but I'm glad to be posting for the first time here. Hopefully I can contribute to some basketball research.
I wanted to share some of the work I've done on the value of blocks, trying to emulate what John Huizinga and Sandy Weil had presented at last year's Sloan Sports Analytics Conference (bchaikin did revive an old thread in order to share the links). I did not get a chance to attend the conference or see Huizinga and Weil's paper, but I was intrigued about the idea of factoring the preblock situation and the type of shot in order to ascertain the value contributed by shot-blockers.
Using the PbP data provided by Ryan J. Parker, I took my own look at blocks between 2007-2010 from two perspectives: 1) Points saved per block based on shot location, and 2) Points saved per block based on shot type. The premise is the same as that of Huizinga's in that not all blocks are created equal. I attempt to calculate value of the block based on the value of the shot which was blocked.
If you would like to take a look at some of my findings, go ahead to the following links:
http://thinkbluecrew.blogspot.com/2010/08/anatomy-of-block-introduction-part-1.html
http://thinkbluecrew.blogspot.com/2010/08/anatomy-of-block-by-shot-location-part.html
http://thinkbluecrew.blogspot.com/2010/08/anatomy-of-block-by-shot-type-part-3.html
Just a sample of some of my findings, I found that the shots which Andris Biedrins and Brendan Haywood blocked were consistently of lower value based on both location of shot and type of shot, from season to season. I also found that Joel Przybilla, Chris Anderson, and Gerald Wallace were consistently among the leaders in high value blocks.
This is my first time doing a research study like this and presenting it to a high-caliber audience like that of this forum, and I wanted to address some concerns about my methodology and findings and see what you guys think (perhaps Sandy Weil could help me out, if you still hang out here). I am aware that the shot types recorded may include bias in the sense that layups that are made tend to include additional descriptions such as "driving layup" or "reverse layup." Concerning the generic shot types vs. the specific shot types, while it leads to interesting studies, it indicates a reason to be careful/skeptical about generalizing certain shot types, especially the generic shot types.
Secondly, I have tried to think of ways to consolidate points per shot based on shot location with points per shot based on shot type, but realize that it is more difficult than it seems. The sample size either becomes too small as we get more specific (say, a turnaround jump shot in a different X-Y coordinate) or there is a presence of multicollinearity since most shot types are defined precisely by where they are located on the court. I'm not sure if I am understanding this correctly from a statistical standpoint, and feel free to correct any of my statements.
Finally, I do understand there is more to the value of a block than just the average value of the shot that was blocked. I was wondering what other research has also been done out there on blocks concerning the trade-off with shooting fouls, goaltending, "Points Created" and transition baskets, etc.
Anyway, definitely let me know if you have any interesting or critical comments about all this.
Thanks all! _________________ I am the author of Think Blue Crew. Follow me on Twitter. |
|
Back to top |
|
|
Albert Lyu
Joined: 20 Oct 2009 Posts: 2 Location: Evanston, IL
|
Posted: Mon Aug 16, 2010 1:46 am Post subject: An update |
|
|
I concluded my series on blocked shots on my blog. Here in Part 5, I started looking at "Points Created" by first calculating "Possessions Gained Per Block" (similar if not the same as the way Sandy Weil did it), then added "Points Saved" with "Points Created" to figure out an overall block value per 36 minutes played. I'm a little bit concerned about the validity of using league PPP with possessions gained to calculate points created, as well as adding points saved with points created.
http://thinkbluecrew.blogspot.com/2010/08/anatomy-of-block-points-created-part-5.html _________________ I am the author of Think Blue Crew. Follow me on Twitter. |
|
Back to top |
|
|
Mike G
Joined: 14 Jan 2005 Posts: 3618 Location: Hendersonville, NC
|
Posted: Mon Aug 16, 2010 7:36 am Post subject: |
|
|
Albert, have you looked at breaking your samples into home and away blocks? Some players (some years) have gotten twice as many blocks at home as on the road. If half of their home 'blocks' are just due to imaginative scorekeeping, might there be a lower correlation with points prevented?
Players mentioned in your articles have a broad range of home/away block rate differentials. Do Biedrins' blocks not translate to many prevented points? Or does he really not block so many as are reported?: Code: | Andres Biedrins
yr Tm HMin AMin Hlk ABk HBk36 ABk36 H/ABk xBlk
2006 GSW 464 527 32 14 2.48 .96 2.60 20
2007 GSW 1240 1138 85 51 2.47 1.61 1.53 29
2008 GSW 1064 1012 57 37 1.93 1.32 1.47 18
2009 GSW 866 988 55 41 2.29 1.49 1.53 19
Marcus Camby
yr Tm HMin AMin HBk ABk HBk36 ABk36 H/ABk xBlk
2006 Den 973 881 112 72 4.14 2.94 1.41 32
2007 Den 1128 1243 121 111 3.86 3.21 1.20 20
2008 Den 1396 1365 162 124 4.18 3.27 1.28 35
2009 LAC 868 1054 63 69 2.61 2.36 1.11 6
Erick Dampier
yr Tm HMin AMin HBk ABk HBk36 ABk36 H/ABk xBlk
2006 Dal 946 982 49 57 1.86 2.09 .89 -6
2007 Dal 956 957 47 35 1.77 1.32 1.34 12
2008 Dal 940 816 72 34 2.76 1.50 1.84 33
2009 Dal 978 859 51 44 1.88 1.84 1.02 1
Tim Duncan
yr Tm HMin AMin HBk ABk HBk36 ABk36 H/ABk xBlk
2006 SAS 1436 1353 89 73 2.23 1.94 1.15 12
2007 SAS 1363 1365 103 87 2.72 2.29 1.19 16
2008 SAS 1272 1382 83 68 2.35 1.77 1.33 20
2009 SAS 1194 1331 52 74 1.57 2.00 .78 -14
|
xBlk are estimated 'extra' blocks at home, relative to away.
H/ABk is just (HBk/36)/(ABk/36)
-- which in turn are per-36-minute blocks, home and away.
Gotten from pages such as these:
http://www.dougstats.com/06-07.html
See this thread for fuller career tables and more:
http://sonicscentral.com/apbrmetrics/viewtopic.php?t=2430
Earlier, see also:
http://sonicscentral.com/apbrmetrics/viewtopic.php?t=2292
http://sonicscentral.com/apbrmetrics/viewtopic.php?t=2310 _________________ `
36% of all statistics are wrong |
|
Back to top |
|
|
|
|
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
|