 |
APBRmetrics
The statistical revolution will not be televised.
|
| View previous topic :: View next topic |
| Author |
Message |
Mike G
Joined: 14 Jan 2005
Posts: 3605
Location: Hendersonville, NC
|
 Posted: Mon Sep 08, 2008 9:28 am Post subject: |
 |
|
| dsparks wrote: | ...
Utah is the easiest place for visiting teams to get credit for an assist. One possible implication is that Utah may have inflated all assist numbers, including those of John Stockton.
... |
David,
Unversed as I am in modern statistical terminology, I can't make heads or tails of your tables. My bad. But it seems that the relevant statistic for whether or not Stockton got too many assists is related not to how easily visitors (to Utah) got assists; but to how Stockton's assists in Utah compare to his assists elsewhere.
From 1987 to 2003, the Jazz home/away ratio of Ast/FG ranges from 1.097 to .980; averaging 1.039. That is to say, for the bulk of Stockton's career, the Utah scorekeepers gave their homeboys 1.039 as many assists as the rest of the NBA did on other courts. The 2008 NBA average was 1.058 .
The most inflated Utah/Assist season, 1998, suggests that Stockton's league-best 47.8% Ast% (b-r.com) should be only 45.7%. And his career 50.2 could be downgraded to merely 49.3% .
Of players with >1000 career minutes, Chris Paul is 2nd at 44.3 Ast%; Magic 3rd at 40.9. Before adjustment, that is.
Thanks for the availability of these home away splits goes to pages like this one:
http://www.basketball-reference.com/fc/tsplit.cgi?team=UTA&year=1994
_________________
`
36% of all statistics are wrong
Last edited by Mike G on Mon Sep 08, 2008 9:41 am; edited 1 time in total |
|
| Back to top |
|
 |
Harold Almonte
Joined: 04 Aug 2006
Posts: 616
|
| Posted: Mon Sep 08, 2008 9:31 am Post subject: |
|
|
Ben F:
| Quote: | | if Luis Scola enters the ball to Tracy McGrady posted up at the elbow, and Mcgrady turns and jabs and fakes and then shoots a fadeaway jumper, some scorers will say that it was an assist because the player shot after a pass without dribbling. |
It shouldn't be an assist, since a fake is already an after pass - individual shot creation move (no matter the "ball on the floor" rule). It should be downgraded to "potential".
Mike G:
| Quote: | | In any case, I'm curious about the availability of an 'adjustment factor' we can apply to players' assist rates. As mentioned above, the rest of the NBA (scorekeepers) seem to think the Nuggets only assist 55% of their FG, so I'd hit every Nugget's assist rate with a 548/676 = .81 factor. But that would apply to just their home games; since that's half their schedule, split the difference: Each Denver assist is worth .905 NBA assist. |
I agree the generous hASTs should be adjusted to away's (wich appears to be already some kind of Lg. averaged team's assist look), but what about teams like NYK and SAC, that seems to be overpunished by their own bookeepers, almost like if they hate them? It's fair to apply the anti-anti-generous adjust here?
It's fair to set a Lg ave. normal hAST biasing, available to adjust every home bookeeping, before any other adjust? |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3605
Location: Hendersonville, NC
|
| Posted: Mon Sep 08, 2008 10:03 am Post subject: |
|
|
Harold,
It took me most of a week to decide the league average home/away Ast/FG ratios are not necessary. Consider that Phx opponents only register 53% (NBA averages 60%) on their own courts when playing the Suns (Phx on the road). Indicating that the specific opponent dictates what kind of offense a team uses.
The Suns 'cheat' themselves at home by awarding merely 89% of the assists they get around the league (per FG). But they award only 80% (of norms) to their opponents.
The Kings and Knicks give themselves the least (88-89%), but give opponents even less (85-86). My suspicion is that everyone else is overly generous: Mostly to themselves; and to cover their tracks a bit, to their visitors.
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
Harold Almonte
Joined: 04 Aug 2006
Posts: 616
|
| Posted: Mon Sep 08, 2008 10:58 am Post subject: |
|
|
Giving a deeper look to your last chart, it's obvious that the most of teams treat opps's assists like they treat themselves with respect to league (keeping a certain patented level of home bias, of course).
Accepting that the diversity in this "treatment to opponents" is minimal; it doesn't link the home team/opp team relationship, to each team home bkeepin /away bkeeping relationship. There is diversity in the way the teams use passing assists, but there is over and under-use, out of rules standards (for not to say abuse), by some teams's official bookeeper. But, I agree that trying, or the way, to adjust for this, would be as subjective as the assist treatment they give, since we, or the NBA, would need to set a standard or a kind of correction after tv replay. |
|
| Back to top |
|
|
dsparks
Joined: 22 Feb 2008
Posts: 61
|
| Posted: Mon Sep 08, 2008 11:56 am Post subject: |
|
|
Mike G: You are right that we cannot deduce anything about Stockton's Home/Away assist splits from the regression. However, I deliberately chose to compare visiting teams coming to arena A versus their performance as visitors in all other arenas. This is because I assumed that differences in individual home/away assist credit was likely associated more with differences in actual home/away performance than differences in home/away scorekeeping. Teams do generally perform better (final scoring margins are greater) at home than away, and I assumed this performance variable would cloud any inference I could derive about scorekeeper bias.
_________________
David
http://arbitrarian.wordpress.com |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3605
Location: Hendersonville, NC
|
| Posted: Tue Sep 09, 2008 8:55 am Post subject: |
|
|
| Mike G wrote: | ...
From 1987 to 2003, the Jazz home/away ratio of Ast/FG ranges from 1.097 to .980; averaging 1.039. That is to say, for the bulk of Stockton's career, the Utah scorekeepers gave their homeboys 1.039 as many assists as the rest of the NBA did on other courts. |
By comparison, the Lakers' scorekeepers have been consistently overzealous in giving their players extra assists.
In Kobe's time, the Lakers at home have granted themselves 1.107 times as many assists as they've gotten on the road (per FG).
In Magic's last 5 (real) seasons, he got an extra 16.7% (1.167) at home. One of every 7 (home) assists was bogus!
Interestingly, the 3 dominant Shaq/Kobe years (2000-02) netted almost no home bias in granting assists. In '07 they were back to +20%.
All the ups and downs in Magic's annual Ast% coincide with greater/lesser home assist bias.
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
|
| Posted: Tue Sep 09, 2008 9:10 am Post subject: |
|
|
I think the main thing we are really talking about here is "the park factor", as they call it in baseball. That basic factor is from this:
PF_Arena = [(OStatH + DStatH)/GH]/[(OStatR + DStatR)/GR]
It's easy to calculate and reasonably well accepted in baseball (there are more sophisticated things and some have suggested alternatives).
So, you get this list for assists in basketball in 2008:
| Code: |
Tm PF_ast
den 1.139
atl 1.114
por 1.106
cle 1.100
mil 1.099
lac 1.089
njn 1.076
nor 1.067
bos 1.066
lal 1.064
phi 1.055
gsw 1.054
was 1.029
det 1.018
hou 1.014
min 1.004
uta 0.980
san 0.979
tor 0.978
sea 0.972
chi 0.963
dal 0.939
ind 0.929
orl 0.905
mem 0.904
cha 0.901
nyk 0.876
mia 0.867
pho 0.864
sac 0.861
|
So you could (potentially) adjust assists up in Sac and Pho, down in Den and Atl.
For 2007, you get this list:
| Code: |
Tm PF_ast
lal 1.141
njn 1.121
gsw 1.100
lac 1.077
den 1.075
mil 1.070
bos 1.059
phi 1.058
det 1.047
hou 1.034
chi 1.031
min 1.028
was 1.026
cle 1.022
por 1.021
sea 1.021
san 0.984
tor 0.981
dal 0.974
uta 0.967
atl 0.963
cha 0.961
ind 0.953
nyk 0.952
nor 0.934
mem 0.928
orl 0.904
sac 0.888
pho 0.848
mia 0.848
|
About a 76% correlation.
For kicks, here is 2002:
| Code: |
Tm PF_ast
chi 1.120
cle 1.117
den 1.112
min 1.097
por 1.086
lac 1.067
mil 1.051
phi 1.044
ind 1.039
bos 1.036
cha 1.034
pho 1.025
mem 1.015
gsw 1.002
sea 0.998
lal 0.997
tor 0.987
sac 0.976
was 0.975
hou 0.966
nyk 0.955
orl 0.952
uta 0.947
det 0.934
dal 0.930
njn 0.918
san 0.905
atl 0.877
mia 0.869
|
Less of a correlation. You'd probably want to know whether the scorers have been consistent in the different locations. And, for the NBA, we probably should call it Arena Scorer Factor (ASF) since we don't believe that there is nothing about the physical arenas that matter; rather, it's the scorers.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers. |
|
| Back to top |
|
|
Ben F.
Joined: 07 Mar 2005
Posts: 391
|
| Posted: Tue Sep 09, 2008 9:15 am Post subject: |
|
|
| HoopStudies wrote: | For kicks, here is 2002:
...
Less of a correlation. |
Any idea how often teams change who is scoring the games? I would think it would be possible that there is a fair amount of turnover in 6 years, but I don't have anything to base that off of. |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3605
Location: Hendersonville, NC
|
| Posted: Tue Sep 09, 2008 9:31 am Post subject: |
|
|
| HoopStudies wrote: | ... "the park factor", as they call it in baseball.
... we don't believe that there is nothing about the physical arenas that matter; rather, it's the scorers. |
Well, a lefty hitter may hit more homeruns in Yankee Stadium, but that's not a subjective call. Whether he plays for the home team or not, if he's a lefty he benefits.
The total assists (or Ast/FG) granted in an arena doesn't address the home-bias of the granting. In Denver, it seems 3/4 of the 'extra' assists go to the Nuggets.
And Ben, why does it matter who is scoring the games? We know the annual totals and the biases for each year, in each arena.
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
|
| Posted: Tue Sep 09, 2008 9:33 am Post subject: |
|
|
EDIT: I figured it out.
_________________
I am a basketball geek.
Last edited by Ryan J. Parker on Tue Sep 09, 2008 10:05 am; edited 1 time in total |
|
| Back to top |
|
|
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
|
| Posted: Tue Sep 09, 2008 9:43 am Post subject: |
|
|
| Ben F. wrote: | | HoopStudies wrote: | For kicks, here is 2002:
...
Less of a correlation. |
Any idea how often teams change who is scoring the games? I would think it would be possible that there is a fair amount of turnover in 6 years, but I don't have anything to base that off of. |
It varies. I think there is a good amount of stability, though. I don't know exact figures, but I do think it's a key thing with stats like this.
If you calculate this for efg, tovp, etc., you find that there is greater variation in ASFs for assists than for the others. That's consistent with it being more subjectively scored.
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers. |
|
| Back to top |
|
|
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
|
| Posted: Tue Sep 09, 2008 10:00 am Post subject: |
|
|
| Mike G wrote: | | HoopStudies wrote: | ... "the park factor", as they call it in baseball.
... we don't believe that there is nothing about the physical arenas that matter; rather, it's the scorers. |
Well, a lefty hitter may hit more homeruns in Yankee Stadium, but that's not a subjective call. Whether he plays for the home team or not, if he's a lefty he benefits.
The total assists (or Ast/FG) granted in an arena doesn't address the home-bias of the granting. In Denver, it seems 3/4 of the 'extra' assists go to the Nuggets.
And Ben, why does it matter who is scoring the games? We know the annual totals and the biases for each year, in each arena. |
This formula doesn't address extra bias that a scorer may have for their home team, just for the scorer overall bias for awarding assists. It looks at the overall rate in one arena (the numerator) vs the overall rate in other arenas for equivalent offense/defensive players (the denominator).
Such factors for things like offensive ratings and efg are pretty close to 1, as you'd expect.
| Code: |
atl ASF_efg = (efgH + defgH)/(efgR + defgR)
=(0.497+0.489)/(0.463+0.512)
=1.011
|
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers. |
|
| Back to top |
|
|
Mike G
Joined: 14 Jan 2005
Posts: 3605
Location: Hendersonville, NC
|
| Posted: Tue Sep 09, 2008 10:35 am Post subject: |
|
|
| HoopStudies wrote: |
This formula doesn't address extra bias that a scorer may have for their home team, just for the scorer overall bias for awarding assists. |
| Code: | ...
chi 0.963
dal 0.939 ... |
These teams rank 21 and 22 by your formula. But no one should expect similar treatment in Chicago and in Dallas, regarding the awarding of assists.
The Mavs got 5% more assists than their opponents, when playing in Dallas. On the road, they were out-assisted by 15%.
Bulls' opponents got 2% more assists (per FG) in Chicago, and 6% more in their own arenas.
The Bulls were one of the least biased (for-home or against-visitor) in the NBA. The Mavs were the most biased.
Some teams get and/or allow more assists. That doesn't mean their arena allows them or their opponents more.
_________________
`
36% of all statistics are wrong |
|
| Back to top |
|
|
HoopStudies
Joined: 30 Dec 2004
Posts: 705
Location: Near Philadelphia, PA
|
| Posted: Tue Sep 09, 2008 11:13 am Post subject: |
|
|
| Mike G wrote: | | HoopStudies wrote: |
This formula doesn't address extra bias that a scorer may have for their home team, just for the scorer overall bias for awarding assists. |
| Code: | ...
chi 0.963
dal 0.939 ... |
These teams rank 21 and 22 by your formula. But no one should expect similar treatment in Chicago and in Dallas, regarding the awarding of assists.
The Mavs got 5% more assists than their opponents, when playing in Dallas. On the road, they were out-assisted by 15%.
Bulls' opponents got 2% more assists (per FG) in Chicago, and 6% more in their own arenas.
The Bulls were one of the least biased (for-home or against-visitor) in the NBA. The Mavs were the most biased.
Some teams get and/or allow more assists. That doesn't mean their arena allows them or their opponents more. |
So are you saying that the arena/scorer has NO IMPACT? And that it's only about how teams play?
_________________
Dean Oliver
Author, Basketball on Paper
The postings are my own & don't necess represent positions, strategies or opinions of employers. |
|
| Back to top |
|
|
Ben F.
Joined: 07 Mar 2005
Posts: 391
|
| Posted: Tue Sep 09, 2008 12:45 pm Post subject: |
|
|
| Mike G wrote: | | And Ben, why does it matter who is scoring the games? We know the annual totals and the biases for each year, in each arena. |
Well, where do you think the systematic bias comes from? If the differences in awarded assists don't track with who is scoring the games, it would indicate that the theory of "certain scorers have different definitions/biases" was incorrect. |
|
| 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
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
