page 1
Author Message
Italian Stallion
Joined: 04 Mar 2009
Posts: 112
PostPosted: Mon Apr 26, 2010 12:03 am Post subject: Question about rebounding Reply with quote
I'm trying to understand the value of a rebound better.
One of the issues I am tossing around is the fact that when a player gets a rebound (especially a defensive rebound), there are often several teammates in the area any of which could have gotten the same rebound if that player was not there. In my mind, that sort of diminishes the value of a rebound relative to a blocked shot that also changed a possession (just as an example).
Has anyone ever looked into what percentage of rebounds could have easily been gathered by teammates?
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Mon Apr 26, 2010 5:21 am Post subject: Reply with quote
It's true that there are often teammates who could have gotten your rebound. And sometimes you could have gotten one, but a teammate did.
So everyone gets only some fraction of the DReb he could have gotten. In the long run, that fraction which is counted represents your true rebounding ability. It's self-correcting.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
schtevie
Joined: 18 Apr 2005
Posts: 409
PostPosted: Mon Apr 26, 2010 6:00 am Post subject: Reply with quote
A wonderful contribution from Eli W, who is mostly gone but not forgotten...
http://www.countthebasket.com/blog/2008 ... g-returns/
Back to top
View user's profile Send private message
Italian Stallion
Joined: 04 Mar 2009
Posts: 112
PostPosted: Mon Apr 26, 2010 6:46 pm Post subject: Reply with quote
Mike G wrote:
It's true that there are often teammates who could have gotten your rebound. And sometimes you could have gotten one, but a teammate did.
So everyone gets only some fraction of the DReb he could have gotten. In the long run, that fraction which is counted represents your true rebounding ability. It's self-correcting.
Excellent point. I'm pissed I didn't think about that. Embarassed
However, wouldn't a leading rebounder (like a C) gain more value from the rebounds he got than some of the smaller players that sometimes defer to the main guy. The C will almost never defer to a PG that's also is in the area.
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Tue Apr 27, 2010 8:08 am Post subject: Reply with quote
Italian Stallion wrote:
Mike G wrote:
It's true that there are often teammates who could have gotten your rebound. And sometimes you could have gotten one, but a teammate did.
So everyone gets only some fraction of the DReb he could have gotten. In the long run, that fraction which is counted represents your true rebounding ability. It's self-correcting.
Excellent point. I'm pissed I didn't think about that. Embarassed
However, wouldn't a leading rebounder (like a C) gain more value from the rebounds he got than some of the smaller players that sometimes defer to the main guy. The C will almost never defer to a PG that's also is in the area.
Also see Ed Kupfer's work here: http://sonicscentral.com/apbrmetrics/vi ... .php?t=173 and the earlier Eli W post here: http://www.countthebasket.com/blog/2008 ... -rebounds/
There are more studies around as well.
Basically, what we want to value is how much the player adds to the team above a bad player at his position. How much does Marcus Camby's defensive rebounding% (32%) add above Nenad Krstic's (12.6%)? This is where that chart that Eli W posted in the first link posted on this thread is key. Only about 27% of the improvement or decrease in player DRB% actually gets transferred over to the team as a whole--whereas 77% of the ORB% of the player get's translated over to the team as a whole.
More, the relations are relatively linear.
Thus, each offensive rebound a player gets should be worth, to the team, about 3 times what each defensive rebound the player gets is worth.
In other words, DeJuan Blair's 16% ORB% causes his rebounding overall to be the most beneficial to his team of any player--even though is TRB% is behind several other players.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Italian Stallion
Joined: 04 Mar 2009
Posts: 112
PostPosted: Tue Apr 27, 2010 1:01 pm Post subject: Reply with quote
I have enough of a mathematical aptitude to follow many of these statistical studies, but I don't have enough of a background to do them myself.
As an experiment, I've been tying to find evidence of the value using an entirely different approach.
I've been watching some of the playoff games and noting every time a player got a rebound with a teammate in position to get it also and without a teammate in position to get it also.
By this standard, it is already clear that offensive rebounds have more value than defensive rebounds. But it appears to me that defensive rebounds have a lot more value than .3. I'm reluctant to quote a value because the sample is small and there's an element of subjective judgment involved, but it is possible that some of these studies that suggest the value to be .3 are missing something.
Back to top
View user's profile Send private message
Italian Stallion
Joined: 04 Mar 2009
Posts: 112
PostPosted: Tue Apr 27, 2010 1:04 pm Post subject: Reply with quote
By the way, thanks a bunch for the links to those studies. Very Happy
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Tue Apr 27, 2010 2:49 pm Post subject: Reply with quote
Italian Stallion wrote:
I have enough of a mathematical aptitude to follow many of these statistical studies, but I don't have enough of a background to do them myself.
As an experiment, I've been tying to find evidence of the value using an entirely different approach.
I've been watching some of the playoff games and noting every time a player got a rebound with a teammate in position to get it also and without a teammate in position to get it also.
By this standard, it is already clear that offensive rebounds have more value than defensive rebounds. But it appears to me that defensive rebounds have a lot more value than .3. I'm reluctant to quote a value because the sample is small and there's an element of subjective judgment involved, but it is possible that some of these studies that suggest the value to be .3 are missing something.
I think to some extent, it's a factor of scheme--if you know you've got a dominant rebounder on defense, most of the rest of the team won't bother to crash the boards. I think.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Tue Apr 27, 2010 3:34 pm Post subject: Reply with quote
In some ways, the DReb is more essential than the OReb. You can compete and win many games with few OReb. You get blown out any time you can't get DReb.
A team almost always has a couple of its best DReb guys on the floor. No coach is willing to experiment with trying to win with OReb, neglecting DReb -- because some say they're less valuable.
The rarity of an event is not equal to its value.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Tue Apr 27, 2010 3:44 pm Post subject: Reply with quote
Mike G wrote:
In some ways, the DReb is more essential than the OReb. You can compete and win many games with few OReb. You get blown out any time you can't get DReb.
A team almost always has a couple of its best DReb guys on the floor. No coach is willing to experiment with trying to win with OReb, neglecting DReb -- because some say they're less valuable.
The rarity of an event is not equal to its value.
Interestingly, the graphs posted by Eli W. dispute this. For lineups with a predicted DRB% of about 61% (based on the 5 players), the total DRB% still stays up at about 71%. In other words, even if you don't put your best defensive rebounders on the court, you won't lose too many rebounds.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DLew
Joined: 13 Nov 2006
Posts: 224
PostPosted: Tue Apr 27, 2010 4:16 pm Post subject: Reply with quote
I think MikeG is confusing the value of defensive rebounds with the value of defensive rebounders. He is absolutely right that you can't win if you don't get defensive rebounds, no question. So at the team level defensive rebounds are very valuable. However, from a player evaluation standpoint Eli has shown that putting additional good defensive rebounders on the court does not yield as many additional defensive rebounds as you would expect based on the raw rebound rates of the players involved. This suggests that though teams should covet getting more defensive rebounds, putting more good defensive rebounders on the court is not a terribly effective strategy for achieving this.
Back to top
View user's profile Send private message
Italian Stallion
Joined: 04 Mar 2009
Posts: 112
PostPosted: Tue Apr 27, 2010 4:25 pm Post subject: Reply with quote
So is it generally accepted that value of a defensive rebound is .3 (give or take)?
That's a huge difference from the value of 1 that D Berri assigns and is even quite a bit different than my own intuitive expectations of .8 or .7 based on preliminary observation.
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Tue Apr 27, 2010 4:43 pm Post subject: Reply with quote
Italian Stallion wrote:
So is it generally accepted that value of a defensive rebound is .3 (give or take)?
That's a huge difference from the value of 1 that D Berri assigns and is even quite a bit different than my own intuitive expectations of .8 or .7 based on preliminary observation.
No, I don't think that 0.3 is the accepted value of a defensive rebound. All we've concluded is that a player's offensive rebounds are worth ~3 times what a player's defensive rebounds are worth. What are rebounds worth, relative to points?
Honestly, I don't think you would be too far off if you used 0.3 for defense and 0.8 for offense. A team rebound is worth basically 1 point (league average is ~1 PPP). The player's contributions are worth less, because his efforts "overlap" with others on the team.
That is a lot different than D Berri's work, and more similar to what statistical +/- assigns as a value. Values vary with the SPM regression, though.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DLew
Joined: 13 Nov 2006
Posts: 224
PostPosted: Tue Apr 27, 2010 5:07 pm Post subject: Reply with quote
Not sure you should assume that an offensive rebound is three times as valuable as a defensive one. Not even sure what that means really. Eli's study did show that an offensive rebounds are three times less sensitive to diminishing returns. But, in terms of value to the team, the values Berri derived from his team efficiency regressions are probably more accurate. And if the question is how should we weight rebounds for individual player evaluation, running a well specified regression relating box score stats to the way the player effects the team (generally referred to as statistical plus-minus) is the best method for coming to an answer.
Basically I reject the concept that there is a universal value we can assign to an offensive and defensive rebound because, as Berri has very clearly illustrated time and time again, the value of certain stats to the team are different than the value that the players getting those stats contribute to team success.
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Tue Apr 27, 2010 5:46 pm Post subject: Reply with quote
DLew wrote:
Not sure you should assume that an offensive rebound is three times as valuable as a defensive one. Not even sure what that means really. Eli's study did show that an offensive rebounds are three times less sensitive to diminishing returns. But, in terms of value to the team, the values Berri derived from his team efficiency regressions are probably more accurate. And if the question is how should we weight rebounds for individual player evaluation, running a well specified regression relating box score stats to the way the player effects the team (generally referred to as statistical plus-minus) is the best method for coming to an answer.
Basically I reject the concept that there is a universal value we can assign to an offensive and defensive rebound because, as Berri has very clearly illustrated time and time again, the value of certain stats to the team are different than the value that the players getting those stats contribute to team success.
Basically, for looking at the individual value, we need to look at the value of marginal rebounds, which is where the Eli W. study comes in. I agree that the values Berri derived for team efficiency contribution are likely close to accurate.
The issue with SPM is that it's somewhat unstable, and the results can be a little weird, depending on the sample. Sometimes, stats are indicators for other contributions--thus the exceptionally high value for steals in the current regression. It's a tricky regression to run right. I'm a big fan of SPM in general, though.
recovered page 2 of 3
Author Message
Mike G
Joined: 14 Jan 2005
Posts: 3509
Location: Hendersonville, NC
PostPosted: Tue Apr 27, 2010 5:49 pm Post subject: Reply with quote
DSMok1 wrote:
... For lineups with a predicted DRB% of about 61% (based on the 5 players), the total DRB% still stays up at about 71%. ...
Well, if a team goes small, the opponent may go small.
For a team to get only 60% DRb chances, the opponent would have to get 40% ORb.
A player who is generally the 3rd or 4th best DRb guy on the floor, finding himself the 2nd best DRb guy will almost have to get more DRb's. There's not some teammate getting them, and there they are.
This follows common sense, I think. It isn't a lot different than putting 5 high-usage guys on the floor and finding they do not total 125% Usage. See the 2004 Mavs.
Quote:
.. A team rebound is worth basically 1 point (league average is ~1 PPP). The player's contributions are worth less, because his efforts "overlap" with others on the team..
In football, several defenders may pressure the QB, but only 1 or 2 will be credited with a 'sack'. They give 'half sack' credits.
If they didn't credit half-sacks, there would be a greater chance that some guys would not get credit for as many sacks as they helped to create. But there is still one sack credited for every instance of a sack.
There aren't half-rebounds in b-ball, it's still a multi-player effort much of the time, and there are still as many rebounds credited to players as are grabbed by someone. Again, in a given time span, some guys don't get as much credit, while others get more.
A player shoots and makes a basket. Quite likely, a teammate (or 4) could have shot, may have hit it; but he didn't. The guy who shot may have thought, "Well, if I pass it to that guy, he may get a shot; but I will get a shot."
If Joe regularly defers to Moe (just boxes out), Moe might not be as good as his rebound totals. Joe might be better than his totals. But if they total 20 rebounds between them, it's worth 20 however you divide them.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Italian Stallion
Joined: 04 Mar 2009
Posts: 90
PostPosted: Tue Apr 27, 2010 7:36 pm Post subject: Reply with quote
"If Joe regularly defers to Moe (just boxes out), Moe might not be as good as his rebound totals. Joe might be better than his totals. But if they total 20 rebounds between them, it's worth 20 however you divide them."
Excellent way of explaining one of the complexities.
I guess what I'm saying is that I have a better idea what a turnover, foul etc... is worth than a rebound because of the greater impact of diminishing returns on rebounds. I'm looking for a "ballpark" estimate of the value of offensive and defensive rebounds for an individual.
I can make a more detailed analysis of the player as a separate exercise and shade my opinion.
Intuitively, I know it has to be less than 1, but I also thought it would be worth more than .3 and .6 or .3 and .8 etc...
Back to top
View user's profile Send private message
ARucker
Joined: 10 Feb 2010
Posts: 34
PostPosted: Thu Apr 29, 2010 6:10 am Post subject: Reply with quote
Mike G wrote:
In some ways, the DReb is more essential than the OReb. You can compete and win many games with few OReb. You get blown out any time you can't get DReb.
Really?
http://www.basketball-reference.com/box ... 30DET.html
http://www.basketball-reference.com/box ... 50CHI.html
If you look at a team's game log and sort by opponent's ORB, I suspect you'll discover that serving up ORs at volume at the single-game level isn't nearly as correlated with losing as you appear to be predicting.
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 565
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Apr 29, 2010 6:59 am Post subject: Reply with quote
ARucker wrote:
Mike G wrote:
In some ways, the DReb is more essential than the OReb. You can compete and win many games with few OReb. You get blown out any time you can't get DReb.
Really?
http://www.basketball-reference.com/box ... 30DET.html
http://www.basketball-reference.com/box ... 50CHI.html
If you look at a team's game log and sort by opponent's ORB, I suspect you'll discover that serving up ORs at volume at the single-game level isn't nearly as correlated with losing as you appear to be predicting.
At a team level, obviously either 1 team will get the ORB or the other will get the DRB--no matter how good or bad the 2 teams are at those facets, it'll still sum to 100%.
At the team level, ORB and DRB are equal, because getting 1 is the inverse of the other team getting the other type.
What we are looking at, though, is the impact of each individual on the overall team rates in each of those categories.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3509
Location: Hendersonville, NC
PostPosted: Thu Apr 29, 2010 8:35 am Post subject: Reply with quote
That first link shows the Pistons getting 41% of their misses back (ORb%) and still losing by 30.
Meanwhile, the Raptors got just 17% ORb and won by 30.
The second link is similar; the ORebounding team loses.
Doesn't that corroborate what I wrote?
The winning team in these examples managed to get almost 60% (DRb%) of opponent misses. Reverse those % and you get opposite results.
Unless, as sometimes happens, one team shoots 60% and the other shoots 28%.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 565
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Apr 29, 2010 8:46 am Post subject: Reply with quote
Mike G wrote:
That first link shows the Pistons getting 41% of their misses back (ORb%) and still losing by 30.
Meanwhile, the Raptors got just 17% ORb and won by 30.
The second link is similar; the ORebounding team loses.
Doesn't that corroborate what I wrote?
The winning team in these examples managed to get almost 60% (DRb%) of opponent misses. Reverse those % and you get opposite results.
Unless, as sometimes happens, one team shoots 60% and the other shoots 28%.
As I said... at a team level, both offensive and defensive rebounding are equal. One is the inverse of the other. And all rebounding combined is not nearly as important as shooting efficiency! Even if you have 15 more "possessions" (not by the rigorous definition), if you're shooting worse and getting to the line less often, it won't be enough.
Back to top
View user's profile Send private message Visit poster's website
Italian Stallion
Joined: 04 Mar 2009
Posts: 90
PostPosted: Thu Apr 29, 2010 8:57 am Post subject: Reply with quote
I was looking at some individual player stats last night and saw a case where a C was a terrible offensive player around the basket one year (low efficiency) but seemed to be a great offensive rebounder that year.
A little light went off in my head.
The guy was so bad inside he was accumulating excessive offensive rebounds on his own misses. So you have to wonder what the values were in that case. Very Happy
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 565
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Apr 29, 2010 9:12 am Post subject: Reply with quote
I ran a quick regression over the last 10 years to verify the value of each of the 4 factors, both for offense and defense.
Team efficiency margin was the dependent variable.
Code:
Regression Statistics
Multiple R 0.993
R Square 0.985
Adjusted R Square 0.985
Standard Error 0.584
Observations 296
Coefficients Standard Error t Stat P-value
Intercept -39.445 2.286 -17.258 2.9E-46
OeFG% 144.714 2.057 70.338 5.4E-183
OTOV% -150.434 3.729 -40.345 2.7E-120
ORB% 42.045 1.536 27.376 5.2E-82
OFT/FGA 31.431 1.556 20.196 4.8E-57
DeFG% -147.144 1.903 -77.329 3.3E-194
DTOV% 133.876 3.626 36.925 4.9E-111
DRB% 42.101 1.885 22.339 9.5E-65
DFT/FGA -27.928 1.479 -18.880 3.0E-52
Obviously, ORB% and DRB% have identical coefficients, so they are worth the same on a team level.
Last edited by DSMok1 on Thu Apr 29, 2010 9:14 am; edited 1 time in total
Back to top
View user's profile Send private message Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 565
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Apr 29, 2010 9:12 am Post subject: Reply with quote
Italian Stallion wrote:
I was looking at some individual player stats last night and saw a case where a C was a terrible offensive player around the basket one year (low efficiency) but seemed to be a great offensive rebounder that year.
A little light went off in my head.
The guy was so bad inside he was accumulating excessive offensive rebounds on his own misses. So you have to wonder what the values were in that case. Very Happy
A prime reason ORB should have a value less than 1! But he also mitigated his own mistakes--if he missed the shot and got the rebound, then kicked it out to his teammates--the missed shot had no net ill effect on the team.
Back to top
View user's profile Send private message Visit poster's website
Jon Nichols
Joined: 18 Aug 2005
Posts: 369
PostPosted: Thu Apr 29, 2010 10:06 am Post subject: Reply with quote
For what it's worth, I took a look at the diminishing returns of rebounds this past December. I also included home/road splits and the diminishing returns of other stats:
http://basketball-statistics.com/blog1/ ... her-stats/
Back to top
View user's profile Send private message AIM Address
DSMok1
Joined: 05 Aug 2009
Posts: 565
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Apr 29, 2010 11:02 am Post subject: Reply with quote
Jon Nichols wrote:
For what it's worth, I took a look at the diminishing returns of rebounds this past December. I also included home/road splits and the diminishing returns of other stats:
http://basketball-statistics.com/blog1/ ... her-stats/
Yow! I forgot about those charts.
Did you ever figure out why the diminishing returns were so strong on assists? Perhaps because there are only a limited number to go around on the team; only 1 player can be the primary ball handler at a time?
This data is very useful in constructing an actual, logical player rating system, rather than simply looking at empirical regressions.
Back to top
View user's profile Send private message Visit poster's website
Jon Nichols
Joined: 18 Aug 2005
Posts: 369
PostPosted: Thu Apr 29, 2010 11:28 am Post subject: Reply with quote
DSMok1 wrote:
Jon Nichols wrote:
For what it's worth, I took a look at the diminishing returns of rebounds this past December. I also included home/road splits and the diminishing returns of other stats:
http://basketball-statistics.com/blog1/ ... her-stats/
Yow! I forgot about those charts.
Did you ever figure out why the diminishing returns were so strong on assists? Perhaps because there are only a limited number to go around on the team; only 1 player can be the primary ball handler at a time?
This data is very useful in constructing an actual, logical player rating system, rather than simply looking at empirical regressions.
I actually never looked into the assists further. Your theory is probably about right, though.
Back to top
View user's profile Send private message AIM Address
Mike G
Joined: 14 Jan 2005
Posts: 3509
Location: Hendersonville, NC
PostPosted: Thu Apr 29, 2010 12:29 pm Post subject: Reply with quote
And the fact that there are diminishing returns on assists -- defined as, "you can't get a lot more assists by putting more passers in the game" -- does not actually affect the value of an assist, does it?
The same analysis would likely hold true for scoring. You might even get fewer points by putting your 5 best "scorers" on the floor.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DSMok1
Joined: 05 Aug 2009
Posts: 565
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Apr 29, 2010 1:54 pm Post subject: Reply with quote
Mike G wrote:
And the fact that there are diminishing returns on assists -- defined as, "you can't get a lot more assists by putting more passers in the game" -- does not actually affect the value of an assist, does it?
The same analysis would likely hold true for scoring. You might even get fewer points by putting your 5 best "scorers" on the floor.
True about the assists--good point. There seems to be a fairly solid benefit for handing out assists--somewhere about 0.6-0.7 points per assist (per 40 minutes).
I looked at the value of assists in terms of points added per possession; I found it to be 0.4 points per assist. Of course, adding marginal points like that is actually worth more than simply the points themselves.
There is no diminishing returns associated with "scorers" per se, however. The FGA are limited, obviously, but FG% is equal to the sum of the parts (weighted, of course, by who actually takes the shots). (Jon did a follow-up post on scoring.)
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3509
Location: Hendersonville, NC
PostPosted: Thu Apr 29, 2010 2:25 pm Post subject: Reply with quote
DSMok1 wrote:
... 0.6-0.7 points per assist (per 40 minutes)...
Don't mean to nit-pick (except as nit-picking may be necessary), but:
Does the per 40 minutes mean anything here?
- 10 assists in 40 min are worth 6-7 pts
- 5 assists in 20 min are worth 3-3.5 pts
- 5 ast in 40 min are worth 3-3.5 pts
- 1 ast is worth .6 - .7 pts, in any minutes
Yes?
Quote:
There is no diminishing returns associated with "scorers" per se, however. The FGA are limited, obviously, but FG% is equal to the sum of the parts...
Well, I was referring to scoring, as in points, not to FG%.
I can hardly imagine that low-FGA guys wouldn't shoot more often if there were 5 of them on the floor at once. This would almost certainly yield more points from most or all of them than they normally produce.
Since your 5 best scorers are generally not also your 5 best ball-handlers, rebounders, defenders, etc, we almost never see 5 scoring specialists take the floor. Other than some final possession situation, where those other skills are unnecessary.
Coaches generally understand this 'diminishing returns' concept, even if they don't know they do. You get the best overall production (and point differential) with a balanced lineup. Contingent upon the opponent lineup, of course.
page 3
Author Message
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Apr 29, 2010 3:17 pm Post subject: Reply with quote
Mike G wrote:
DSMok1 wrote:
... 0.6-0.7 points per assist (per 40 minutes)...
Don't mean to nit-pick (except as nit-picking may be necessary), but:
Does the per 40 minutes mean anything here?
- 10 assists in 40 min are worth 6-7 pts
- 5 assists in 20 min are worth 3-3.5 pts
- 5 ast in 40 min are worth 3-3.5 pts
- 1 ast is worth .6 - .7 pts, in any minutes
Yes?
Yes, you're right.
Quote:
There is no diminishing returns associated with "scorers" per se, however. The FGA are limited, obviously, but FG% is equal to the sum of the parts...
Well, I was referring to scoring, as in points, not to FG%.
I can hardly imagine that low-FGA guys wouldn't shoot more often if there were 5 of them on the floor at once. This would almost certainly yield more points from most or all of them than they normally produce.
Since your 5 best scorers are generally not also your 5 best ball-handlers, rebounders, defenders, etc, we almost never see 5 scoring specialists take the floor. Other than some final possession situation, where those other skills are unnecessary.
Coaches generally understand this 'diminishing returns' concept, even if they don't know they do. You get the best overall production (and point differential) with a balanced lineup. Contingent upon the opponent lineup, of course.[/quote]
Scoring doesn't mean much of anything, in my opinion, without the FG% also.
An interesting note--I recently estimated the threshold TS% to contribute to the overall effort. I found it to be 35.6%. When I include a squared term (Value = (TS%-Threshold)*TSA + ((TS%-Threshold)*TSA )^2)), I get a threshold of 35%.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3578
Location: Hendersonville, NC
PostPosted: Thu Apr 29, 2010 3:50 pm Post subject: Reply with quote
DSMok1 wrote:
...
Scoring doesn't mean much of anything, in my opinion, without the FG% also.
An interesting note--I recently estimated the threshold TS% to contribute to the overall effort. I found it to be 35.6%. ..
I would think the value of a given TS% (or FG%) would depend on opponent TS%, ORb%, and etc.
In general, your average possession has to be worth more than your opponent's. If they're knocking down 58%, then 56% is not good enough.
In the '70s, the Celtics won 2 titles, with terrible shooting and 34-35% ORb%. Jo Jo White's nightly 8-18 FG shooting was fine when 3-4 misses were recovered by Cowens or Silas.
Only reason I argue for flexible 'threshold of contribution' in shooting is that the playoffs are happening. A given playoff series has its own parameters for what's good enough.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Italian Stallion
Joined: 04 Mar 2009
Posts: 112
PostPosted: Thu Apr 29, 2010 7:47 pm Post subject: Reply with quote
There is something I don't understand about the relationship between diminishing returns and the value assigned to something.
If I understood it , there seems to be a large element of diminishing returns associated with blocks. That makes perfect sense to me because there is usually one primary shot blocker in the paint for the defense. So adding another floating around outside the paint may not help much.
However, that may have little to do with the value of a blocked shot.
Assume a guy blocks a shot and takes possession. That should be worth 1 unless someone else that was on the court with him would have blocked it if he didn't.
If you ask me that rarely happens. Usually only one guy has a chance to block any given shot.
So I suspect the value of a block is related to percentage of the time the blocked shot causes a change in possession and is not just swatted away off the court etc... without much consideration for the other players or diminishing returns.
Rebounding is an entirely different animal because there are always several people in the area with a chance to get a defensive rebound if one particular didn't.
Back to top
View user's profile Send private message
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Apr 29, 2010 7:51 pm Post subject: Reply with quote
Italian Stallion wrote:
There is something I don't understand about the relationship between diminishing returns and the value assigned to something.
If I understood it , there seems to be a large element of diminishing returns associated with blocks. That makes perfect sense to me because there is usually one primary shot blocker in the paint for the defense. So adding another floating around outside the paint may not help much.
However, that may have little to do with the value of a blocked shot.
Assume a guy blocks a shot and takes possession. That should be worth 1 unless someone else that was on the court with him would have blocked it if he didn't.
If you ask me that rarely happens. Usually only one guy has a chance to block any given shot.
So I suspect the value of a block is related to percentage of the time the blocked shot causes a change in possession and is not just swatted away off the court etc... without much consideration for the other players or diminishing returns.
Rebounding is an entirely different animal because there are always several people in the area with a chance to get a defensive rebound if one particular didn't.
Yep.... it's a complicated subject (which is an understatement!)
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Thu Apr 29, 2010 7:53 pm Post subject: Reply with quote
Mike G wrote:
In the '70s, the Celtics won 2 titles, with terrible shooting and 34-35% ORb%. Jo Jo White's nightly 8-18 FG shooting was fine when 3-4 misses were recovered by Cowens or Silas.
Only reason I argue for flexible 'threshold of contribution' in shooting is that the playoffs are happening. A given playoff series has its own parameters for what's good enough.
I would argue that Jo Jo White was not contributing much; it was Cowens and Silas that were the ones winning the games. In fact, I'm sure of it. The average efficiency of the league in points per possession has not fluctuated that much.
Back to top
View user's profile Send private message Send e-mail Visit poster's website
DSMok1
Joined: 05 Aug 2009
Posts: 602
Location: Where the wind comes sweeping down the plains
PostPosted: Fri Apr 30, 2010 3:33 pm Post subject: Reply with quote
Tom Haberstroh with a contribution:
Offensive Orientation and Re-Estimating ORR