1 of 3
Author Message
Mike G
Joined: 14 Jan 2005
Posts: 3554
Location: Hendersonville, NC
PostPosted: Sun Sep 21, 2008 7:01 am Post subject: Home Assist Adjustments Reply with quote
Another spin-off thread, inspired by this one:
http://sonicscentral.com/apbrmetrics/vi ... 76&start=0
BenF had his definitions and terms, then several others had theirs. In order to keep my focus separate, I'll define what I'm looking at here. First, some terms best left behind:
- "Scoring" is putting the ball in the basket. It may be a short form of "scorekeeping", unrelated to the kind of scoring that wins games; but it makes for some very confusing discussion when used to mean jotting notes on the sideline.
- "Bias" has been used several ways, but for my purposes it just means home assist generosity towards the home team. It doesn't affect the outcome of games; nor does it indicate whether a team played better. It just affects players' assist totals/rates/%.
Home/away splits, like this one:
http://www.basketball-reference.com/fc/ ... &year=1987
... show that the 1987 Lakers averaged 48 FG at home, and 43.2 on the road. They also averaged 34.1 assist at home, and only 25.2 on the road.
Their home/away Ast/FG are therefore (341/480=) .710 at home, to just .583 on the road. The rest of the league only thought they assisted on (583/710) 82% as many of their FG.
The 'away' Ast/FG is what I'm calling 'true' Ast%. If we adjust the '87 Lakers' home Ast% by the factor .82, then we can estimate their 'real' assists -- per FG, per game, and total.
A player who is assumed to play half his games (minutes) at home can be assumed to have benefitted (or suffered) equally from whatever the home scorekeepers have seen fit to award, for half their totals. To leave the 'away' half unaffected, we halve the adjustment factor. For this team,
fac = (1 + .82)/2 = .91
So we can correct Magic Johnson's 1987 assists by the factor .91 . Does he still win the assist title (his last)?
Code:
year player team fac G Ast Ast2 As/G As/G2 Ast% Ast%2
1987 Magic Johnson LA .911 80 977 890 12.2 11.1 .472 .430
1987 Sleepy Floyd GS .982 82 848 833 10.3 10.2 .380 .373
1987 Isiah Thomas Det .980 81 813 797 10.0 9.8 .393 .385
1987 Glenn Rivers Atl .932 82 823 767 10.0 9.4 .430 .401
1987 John Stockton Uta .959 82 670 643 8.2 7.8 .476 .457
1987
In fact, Magic's lead over Sleepy (12.2 to 10.3 Ast/G) is cut in half; his total-assist margin from 129 down to just 57. Floyd's Warriors were only jacking their players' assists up by a couple %.
This marks the advent of the Stockton era. Listing just the top 3 Ast/G men; plus anyone who, after adjustment, would crack the top 3 (in Ast or Ast/G). Ast% (from b-r.com) and an adjusted version are listed.
Ast%2 = Ast% x fac
Code:
year player team fac G Ast Ast2 As/G As/G2 Ast% Ast%2
1988 John Stockton Uta .965 82 1128 1089 13.8 13.3 .548 .529
1988 Magic Johnson LA .932 72 858 799 11.9 11.1 .454 .423
1988 Mark Jackson NY .919 82 868 798 10.6 9.7 .374 .344
1988 Terry Porter Por .985 82 831 819 10.1 10.0 .359 .354
1988 - Porter should have been 3rd in Ast/G and 2nd in assists.
1989 John Stockton Uta .992 82 1118 1109 13.6 13.5 .543 .539
1989 Magic Johnson LA .919 77 988 908 12.8 11.8 .486 .447
1989 Kevin Johnson Phe .938 81 991 929 12.2 11.5 .404 .379
1989
1990 John Stockton Uta .997 78 1134 1131 14.5 14.5 .574 .573
1990 Magic Johnson LA .948 79 907 859 11.5 10.9 .455 .431
1990 Kevin Johnson Phe .897 74 846 759 11.4 10.3 .444 .398
1990 Muggsy Bogues Cha .982 81 867 851 10.7 10.5 .446 .438
1990 - Muggsy over KJ in all categories.
year player team fac G Ast Ast2 As/G As/G2 Ast% Ast%2
1991 John Stockton Uta .966 82 1164 1125 14.2 13.7 .575 .556
1991 Magic Johnson LA .936 79 989 926 12.5 11.7 .493 .462
1991 Kevin Johnson Phe .973 77 781 760 10.1 9.9 .407 .396
1991
1992 John Stockton Uta .981 82 1126 1105 13.7 13.5 .537 .527
1992 Kevin Johnson Phe .937 78 836 783 10.7 10.0 .406 .380
1992 Tim Hardaway GS .965 81 807 779 10.0 9.6 .333 .321
1992
1993 John Stockton Uta .997 82 987 984 12.0 12.0 .498 .496
1993 Tim Hardaway GS .913 66 699 638 10.6 9.7 .398 .363
1993 Scott Skiles Orl .975 78 735 717 9.4 9.2 .348 .339
1993
year player team fac G Ast Ast2 As/G As/G2 Ast% Ast%2
1994 John Stockton Uta 1.006 82 1031 1037 12.6 12.7 .531 .534
1994 Muggsy Bogues Cha .970 77 780 756 10.1 9.8 .392 .380
1994 Mookie Blaylock Atl 1.000 81 789 789 9.7 9.7 .404 .404
1994
1995 John Stockton Uta .966 82 1011 976 12.3 11.9 .526 .508
1995 Kenny Anderson NJ .926 72 680 630 9.4 8.7 .431 .399
1995 Muggsy Bogues Cha .914 78 675 617 8.7 7.9 .403 .368
1995 Avery Johnson SA 1.035 82 670 693 8.2 8.5 .334 .346
1995 - Avery easily over Muggsy, and almost takes 2nd
1996 John Stockton Uta .958 82 916 877 11.2 10.7 .491 .470
1996 Jason Kidd Dal 1.005 81 783 787 9.7 9.71 .415 .417
1996 Avery Johnson SA .966 82 789 762 9.6 9.3 .391 .378
1996 Rod Strickland Por 1.015 67 640 650 9.6 9.70 .434 .441
1996 Damon Stoudamire Tor .951 70 653 621 9.3 8.9 .374 .356
1996 - A 3-way race for 2nd becomes a 2-way near-tie; Damon wasn't that close.
year player team fac G Ast Ast2 As/G As/G2 Ast% Ast%2
1997 Mark Jackson Den-Ind 1.016 82 935 950 11.4 11.6 .482 .490
1997 John Stockton Uta .991 82 860 852 10.5 10.4 .460 .456
1997 Rod Strickland Was 1.038 82 727 754 8.9 9.2 .381 .395
1997
1998 Rod Strickland Was .999 76 801 800 10.5 10.5 .431 .430
1998 Jason Kidd Phx .942 82 745 702 9.1 8.6 .356 .335
1998 Mark Jackson Ind 1.021 82 713 728 8.7 8.9 .467 .477
1998
1999 Jason Kidd Phx .950 50 539 512 10.8 10.2 .440 .418
1999 Rod Strickland Was .953 44 434 414 9.9 9.4 .459 .437
1999 Gary Payton Sea 1.002 50 436 437 8.7 8.7 .413 .414
1999
2000 Jason Kidd Phx .962 67 678 652 10.1 9.7 .401 .386
2000 Nick Van Exel Den .932 79 714 665 9.0 8.4 .397 .370
2000 Sam Cassell Mil .963 81 729 702 9.0 8.7 .411 .396
2000 Gary Payton Sea .949 82 732 695 8.9 8.5 .377 .358
2000 John Stockton Uta .968 82 703 680 8.6 8.3 .481 .466
2000
2001 Jason Kidd Phx .956 77 753 720 9.8 9.3 .412 .394
2001 John Stockton Uta 1.001 82 713 714 8.7 8.7 .487 .488
2001 Nick Van Exel Den .914 71 600 548 8.5 7.7 .384 .351
2001 Mike Bibby Van .983 82 685 673 8.4 8.2 .383 .376
2001
2002 Andre Miller Cle .924 81 882 815 10.9 10.1 .498 .460
2002 Jason Kidd NJ .998 82 808 806 9.9 9.8 .425 .424
2002 Gary Payton Sea .967 82 737 712 9.0 8.7 .394 .381
2002 - 'Dre almost didn't 'really' win that assist title.
year player team fac G Ast Ast2 As/G As/G2 Ast% Ast%2
2003 Jason Kidd NJ .986 80 711 701 8.9 8.8 .422 .416
2003 Jason Williams Mem .953 76 631 601 8.3 7.9 .419 .399
2003 Stephon Marbury Phx .954 81 654 624 8.1 7.70 .366 .349
2003 John Stockton Uta 1.010 82 629 635 7.7 7.75 .464 .469
2003
2004 Jason Kidd NJ .972 67 618 601 9.2 9.0 .448 .436
2004 Steve Nash Dal 1.008 78 687 692 8.8 8.9 .383 .386
2004 Sam Cassell Min .910 81 592 539 7.3 6.7 .380 .346
2004 Eric Snow Phl 1.001 82 563 564 6.9 6.9 .330 .330
2004 - Sam's fabulous '04 was vastly inflated. Nash was nearly on top already.
2005 Steve Nash Phx 1.003 75 861 863 11.5 11.5 .492 .493
2005 Brevin Knight Cha .969 66 591 572 9.0 8.7 .493 .478
2005 Jason Kidd NJ .961 66 545 524 8.3 7.9 .406 .390
2005 Stephon Marbury NY 1.003 82 668 670 8.1 8.2 .360 .361
2005 - Marbs gets short shrift. He really was better!
year player team fac G Ast Ast2 As/G As/G2 Ast% Ast%2
2006 Steve Nash Phx 1.035 79 826 855 10.5 10.8 .444 .460
2006 Brevin Knight Cha 1.013 69 610 618 8.8 9.0 .423 .429
2006 Chauncey Billups Det .980 81 699 685 8.6 8.5 .393 .385
2006
2007 Steve Nash Phx 1.037 76 884 916 11.6 12.1 .501 .519
2007 Deron Williams Uta .989 80 745 737 9.3 9.2 .416 .411
2007 Jason Kidd NJ .945 80 736 695 9.2 8.7 .414 .391
2007 Chris Paul NO .986 64 569 561 8.9 8.8 .413 .407
2007
2008 Chris Paul NO .932 80 925 862 11.6 10.8 .522 .487
2008 Steve Nash Phx 1.063 81 897 954 11.1 11.8 .472 .502
2008 Deron Williams Uta .984 82 862 848 10.5 10.3 .436 .429
Finally, a clear upset. Nash should definitely have won his 4th straight assist title.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
DLew
Joined: 13 Nov 2006
Posts: 224
PostPosted: Sun Sep 21, 2008 3:50 pm Post subject: Reply with quote
Mike,
If I understand what you're doing correctly, then I think you've made a slight error in your adjustment. Let's say we have a player with 820 assists in 82 games (41 at home 41 on the road). Right now you are assuming that he got 410 assists at home and adjusting those down by a factor, lets say .8 (.8 * 410 = 328). So, you would say this hypothetical player should have 738 assists.
However, isn't it more likely that more likely that the player got a lot more of his assists at home? Isn't more likely that he had 500 assists at home and 320 on the road, which would yield an estimate 720 true assists?
So, basically the quick and easy way to improve this is use the home/road assist ratio to make sure that you apply the adjustment to the appropriate number of assists.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3554
Location: Hendersonville, NC
PostPosted: Sun Sep 21, 2008 4:34 pm Post subject: Reply with quote
David, thanks for the input.
Yes, I do see your point. Most teams score more FG at home, as well as having higher Ast/FG at home. And some players work more minutes on the road (because they aren't in as many blowout wins?) This was a rather more streamlined first approach, pointing out possibilities.
In my 1987 Magic Johnson example, using your suggestion, his 977 assists were 520 at home, 457 on the road. He played more road minutes (in 40 G each) away from LA, 1484-1420.
457 + 520*.82 = 457 + 426 = 883 'adjusted' assists.
My 1st adjustment yielded 890. Given that LA's .911 factor is to date the very worst for a team with the league's assist champ, I doubt I'd get much more insight by looking up each player's splits (from the list above).
The list mostly serves as a taking-off point for closer study. Did Chris Paul get a higher % of his team's assists on the road?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Harold Almonte
Joined: 04 Aug 2006
Posts: 616
PostPosted: Sun Sep 21, 2008 5:22 pm Post subject: Reply with quote
This is what I was trying to say in the other thread. There are bookeepers who inflate locals's assists (especially PGs's or favorites's), while others (very few) are very stingy, or excesivelly fair. But, there are as well a lot of teams which play a lot better at home, with a healthier first factor, meaning some more (really or possible) "fair and deserved assists". How do we separate home inflation from better play?
Back to top
View user's profile Send private message
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 708
Location: Raleigh, NC
PostPosted: Sun Sep 21, 2008 5:32 pm Post subject: Reply with quote
Quote:
How do we separate home inflation from better play?
This is why we examine the percentage of made field goals that ended up with an assist awarded. If each team theoretically plays the same style, then playing better should not affect this metric.
That doesn't we have a clear picture, but it's the best we can get without trying to examine the rates a team should get for specific types of shots, lineup usage, or any other factor that would affect this metric.
I do think it's more informative to simply look at differences between arenas, instead of jumping straight to the conclusion if bias. This is clearly going to happen, but it might be just as informative (if not more) if we understand the difference between the way different people (in this case arenas) view an assist.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Harold Almonte
Joined: 04 Aug 2006
Posts: 616
PostPosted: Sun Sep 21, 2008 6:15 pm Post subject: Reply with quote
Quote:
This is why we examine the percentage of made field goals that ended up with an assist awarded. If each team theoretically plays the same style, then playing better should not affect this metric.
Why not? They can have some percentage of variation even remaining at the same "style" (same pace, same PGs, same scorers, same coaching, even same FT/FG). Don't home advantage affect other stats? Are they supposed to have the same SOS at home than road?
About the arenas, what we could see from Mike's chart was that bookeepers who were "suposedly" stingy with locals, invariably did the same with visitors (both teams suposedly changed styles - by what force?); and some teams which were known as very good defenders, the local bookeeper were stingy with visitors only, but the opposite with local. Then we have of both effects.
Back to top
View user's profile Send private message
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 708
Location: Raleigh, NC
PostPosted: Sun Sep 21, 2008 7:10 pm Post subject: Reply with quote
Well that is the assumption we're using, at least. Of course it is possible that a team playing "better" could in turn shoot more from assist-able positions, but I'm not sure we can even measure this.
The other things I mentioned before are certainly measurable in some way, and in reality can be a direct result of "playing better". I understand what you're saying, and I think some of that can come from measuring other variables that directly affect the ratio.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3554
Location: Hendersonville, NC
PostPosted: Mon Sep 22, 2008 6:43 am Post subject: Reply with quote
Of 614 team-seasons since 1987, just 90 (14.7%) have had more assists on the road. Only 78 (13%) have scored more FG on the road. 151 (24.6%) have a higher 'away' Ast/FG ratio. The average road team has been awarded assists on .582 of their FG; at home, .620 .
Virtually no team has gotten more FT on the road, or actually played better (won more) overall.
Sometimes scorekeepers inflate their home players' assist rates in response to a downward trend in the assist department. Perhaps they honestly believe their guys 'should be' getting a certain number; or they deliberately cook the books to make them 'look better' in this one category. Since it doesn't affect the game, they can operate without much scrutiny.
I doubt than many home assists are actually unrecorded, anywhere or anytime. Even a 'stingy' team like the Suns is probably giving out a certain number of iffy assists. And everyone else is giving more.
The stingiest team of all was the Clippers of 1996-1999. For these 4 years, their away Ast/FG was just about average -- .582 -- while at home they got only .466 . If we were to use this as a baseline, then 20% of all NBA assists are inflated, extra, in addition to what is strictly defined. On top of this 20% extra, most scorekeepers pad even more onto their home teams: on average, +6%.
That 6% brings the average home inflation to some 126% of the Clipper Line, and includes the 15% of teams that actually resist the leaguewide assist generosity (fewer Ast at home). The 1997 thru '99 Raptors gave their boys another 32% at home. In the same years, a player in Toronto would get 50-55% more assists (at home) than he'd get for the Clipps.
So, Damon Stoudamire may have looked like an assistinger player than Pooh Richardson, in those years. But in fact, maybe he wasn't.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Harold Almonte
Joined: 04 Aug 2006
Posts: 616
PostPosted: Mon Sep 22, 2008 5:41 pm Post subject: Reply with quote
I think a 20% is too high, I prefer to think is maybe half this, and that this bookeeper just was doing a negligent job. If it was even a league historical overall generosity of 10%, then probably are the rules which need to be adapted (custom makes rules, and assists just count for ratings-and rates, and not for win-lose or possession game decissions as scored points, or like TOs, or fouls). Probably they got to be impressed and are rewarding good playmaking, by crediting some "one, and two moves by scorer"- assists, as a result of an excellent or very decisive pass and playmaking (the Suns's bookeeper would be an stoic for not to say a game un-lover, if this is the case).
If this is the case, it should be the league or metricians (teams's and independents) who need to be worry about?
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3554
Location: Hendersonville, NC
PostPosted: Mon Sep 22, 2008 6:21 pm Post subject: Reply with quote
Harold,
I'm not suggesting that everyone since 1987 (or until whatever time we may come up with home/away splits) should have their assists reduced by 0-35%, leaving intact rates before 1987. As you suggest, probably there's always been home bias (and 'arena bias') happening.
But for comparing players in a given year, since 1987, these ratios are enlightening. Maybe if we can bring these discrepancies to light, homerist assist-counters won't be quite so blatant.
In any case, whatever adjustment one makes should be fair to all players in a given season. There does seem to be less home bias in the 21st century; so maybe, before '87, there was even more?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Mike G
Joined: 14 Jan 2005
Posts: 3554
Location: Hendersonville, NC
PostPosted: Mon Nov 24, 2008 8:58 am Post subject: Reply with quote
Last year's most biased home-assist scorekeepers, those in Denver, seem to be up to their old tricks again. On the road, just 58% of their FG have been assisted, vs 73% at home. One of every six is bogus!
The Suns, meanwhile, are getting just 51% of their FG counted as assisted at home, while their opponents' scorekeepers have adjudged that 66% of Suns' FG are assisted.
Factoring in these (small-sample) discrepancies, some guys whose assist rates appear pretty similar, perhaps shouldn't :
Code:
team AFac player Ast% Adj%
Den .90 Billups 32.8 29.5
Phx 1.14 S Nash 36.9 42.2
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Mike G
Joined: 14 Jan 2005
Posts: 3554
Location: Hendersonville, NC
PostPosted: Sat Nov 29, 2008 6:45 am Post subject: Reply with quote
Great news, I think. This year, home assist bias has been significantly less than it was last year. The average team is only boosting their homeboys' assists by 3.3% (over what the same players get on the road), compared to 5.3% last year.
Code:
As/FG Home Away diff
0708 .598 .568 .030
0809 .573 .555 .018
The Suns were 2nd-stingiest (after Sac) last year in home assist bias. This year, they're off the charts, though. A few teams were relatively stingy last year and are even stingier this season:
Code:
tm 2008 2009 diff
Phx 1.064 1.164 .100
Mia 1.041 1.077 .036
Ind 1.011 1.026 .015
These are the 'home assist factors' described upthread (with the DLew suggestion). At this early stage in the season, assuming 1/2 of games at home and away is generally wrong. Phx, though has played 8 at home and 9 on the road. Knowing that, their factor could be changed to 1.150; but the league as a whole wouldn't change.
Last year, these teams were anti-biased, but they've relaxed their standards this season:
Code:
tm 2008 2009 diff
Orl 1.010 .973 -.037
Cha 1.023 .982 -.041
Chi 1.007 .956 -.052
NYK 1.061 1.009 -.052
Mem 1.033 .937 -.095
A few teams that had bias last year and have been even worse this year:
Code:
tm 2008 2009 diff
Atl .921 .885 -.036
Was .974 .936 -.038
Phl .957 .911 -.046
The Hawks have played just 6 of 15 at home.
15 teams had bias last year but have reduced (or reversed) that so far this season. Significantly:
Code:
tm 2008 2009 diff
SAS .959 1.046 .087
GSW .949 1.029 .080
Det .943 1.003 .060
LAL .952 1.011 .059
Bos .919 .974 .055
Min .977 1.026 .049
Por .921 .968 .048
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
gabefarkas
Joined: 31 Dec 2004
Posts: 1313
Location: Durham, NC
PostPosted: Mon Dec 01, 2008 12:01 pm Post subject: Reply with quote
I think it might have been mentioned already either in this thread or the one it spun off from, but I'm curious what people think of this article examining home team bias from refs. It makes some similar points, citing a few numbers to back them up, too.
Back to top
View user's profile Send private message Send e-mail AIM Address
Mike G
Joined: 14 Jan 2005
Posts: 3554
Location: Hendersonville, NC
PostPosted: Fri Mar 06, 2009 7:44 am Post subject: Reply with quote
Big shout-out to the Atlanta Hawks, or to their scorekeeper. Giving assists on .659 of their FG at home, while the rest of the league gives them .511 (in other cities). The difference (jack factor) of .148 is highest of any team since Den'04 (.161) and 8th worst since 1987.
The league overall has reduced from an average of .030 jack last year, to .025 this year. The Hawks regressed most, .049 worse than last season; followed by Dal (.047) and Mem (.044).
Most-improved are SAS (-.064), Por (-.061), Det (-.049), Ind (-.046).
Phx is still the model of moderation; giving their boys but .523 Ast/FG, while the rest of the league gives them .619 . If .100 were subtracted league-wide, we'd guess about 10% of all assists are 'bogus'.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Mike G
Joined: 14 Jan 2005
Posts: 3554
Location: Hendersonville, NC
PostPosted: Mon Mar 09, 2009 8:33 am Post subject: Reply with quote
Who is the Real assist champ this year?
Chris Paul appears to be walking away with it. But again, his assists have been greatly inflated by homer scorekeeping, relative to others.
The prefix 'e' can be read 'equivalent'.
Code:
Ast Ast/G Ast% Assists tm AsFac G eAst eAs/G eAs%
619 11.05 .551 Paul,Chris NOH .936 56 579 10.35 .516
521 8.40 .401 Rondo,Rajon Bos .943 62 491 7.92 .378
521 8.40 .342 Kidd,Jason Dal .936 62 487 7.86 .320
513 9.68 .429 Nash,Steve Phx 1.092 53 560 10.57 .468
497 10.57 .479 Williams,Deron Uta .983 47 488 10.39 .471
463 7.85 .290 Duhon,Chris NYK 1.027 59 475 8.06 .298
452 7.53 .409 Wade,Dwyane Mia 1.057 60 477 7.96 .432
417 6.95 .373 James,Lebron Cle .959 60 399 6.66 .358
411 6.74 .314 Felton,Raymond Cha 1.006 61 413 6.78 .316
408 8.50 .396 Calderon,Jose Tor 1.003 48 409 8.53 .397
Ast% is the fraction of teammate FG the player assists (from b-r.com). eAst, eAs/G, and eAs% are just the official numbers multiplied by team Assist Factor (AsFac).
AsFac = ((AAst/AFG)/(HAst/HFG)*HFG)+AFG)/(HFG+AFG)
H and A are Home and Away season totals for the team.
AAst/AFG represents a team's 'true' % of FG assisted: what the rest of the league records.
page 2
Author Message
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Fri Apr 03, 2009 6:55 am Post subject: Reply with quote
Mike G wrote:
Big shout-out to the Atlanta Hawks, or to their scorekeeper. Giving assists on .659 of their FG at home, while the rest of the league gives them .511 ...
Holy cow. Since this appeared (Mar.6), Atl has played 11 home games, and awarded assists on only .514 of their FG.
They also won 9 of the 11. Instant karma!
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Odysseus
Joined: 08 Apr 2009
Posts: 1
Location: Charleston, SC
PostPosted: Thu Apr 09, 2009 9:36 pm Post subject: assists discrepancy Reply with quote
Have you considered the road teams maybe be negatively affected?
What I mean is that the home scorers grade the home team appropriately but are negatively biased against visiting teams?
Perhaps seeing the home team run so many of the same plays they can piece together longer strings of action and assign assists while for the visiting team (because the faces, skills, and team plays change) the scorers have trouble keeping track of the action and are less likely to attribute causality to the point guards of the other team?
_________________
I am interested in studying basketball and predicting the outcome of future games.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Mon Apr 13, 2009 6:57 am Post subject: Reply with quote
Home team bias may be manifested as being overly-generous to the home team, or overly-stingy to the visitors. It's hard to say which it is.
A couple of years ago, some here were trying to explain why Phx opponents got so few assists. Was the Suns' D such that opponents got lots of 1-on-1 opportunities and hardly had to pass the ball? A lot of unlikely possibilities.
All along, right under our noses, was the home-away split of Suns and their opponents. For this year --
http://www.basketball-reference.com/fc/ ... &year=2009
(per game averages)
Code:
@Phx FG Ast A/FG
Suns 42.6 22.7 .533
Opp 39.6 18.8 .475
@opp FG Ast A/FG
Suns 39.7 23.7 .597
Opp 40.3 23.7 .588
At Phx, the Suns get 12% more assists (per FG) than their opponents. On the road, they get just 1.5% more. Whether this bias is deliberate or unconscious, it's about par for the league.
Suns' opponents in those opponents' houses get assists on 59% of their FG, while playing the Suns --which is right at the league average. So it's nothing to do with Phx style of play; just that Phx scorekeepers don't grant assists very freely.
Meanwhile, Suns' opponents' scorekeepers have no apparent trouble noticing Suns' assists, and actually give them 12% more (than their own scorekeepers give). I don't call this bias (as in 'for the home team') but generosity.
What I've been tracking this year is the bias aspect that can be found in these splits. The question of whether the league as a whole is too generous (or the degree of the generosity) is more a matter of aesthetics, to me.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
CareyScurry
Joined: 29 Jul 2005
Posts: 8
PostPosted: Sun May 17, 2009 11:13 pm Post subject: A junk response to the issue... Reply with quote
Granted, it's implied in the above posts - but - to be clear...
Isn't this two issues:
* A given scorer might be overly (dis-)inclined to grant assists and,
* A given scorer might show (unusually little/excessive) favoritism to home players
Accordingly, my quick-n-dirty response back when I first thought about this was to create separate factors, multiply them together, and use the square root (didn't revert to mean, only used single year of data at a time). This is fraught with problems (why presume these issues are of equal importance, % of pt given to types of lineups/strategies/indiv. players is the same in both settings, this functional form, yadda yadda), but - again - this was a junk response.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Thu May 21, 2009 8:41 am Post subject: Reply with quote
Through 2 rounds of playoffs, Orlando Ast/FG ratio has been .61 on the road and just .46 at home.
Boston scorekeepers, meanwhile, have given their Celtics assists on .65 of their FG at home, while they've gotten .567 on the road.
Versus Chicago, there was no apparent bias: Celts Ast/FG was .56 in Bos and .57 in Chi. It all blew up in the Orlando series. Celts' A/F was .56 in Orl, .76 in Bos.
Meanwhile, Orl got .48 at home and .64 in Boston.
Relative to Orlando's scorekeeper, Boston's gave Orl 34% more assists and the Celts 36% more.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
mtamada
Joined: 28 Jan 2005
Posts: 376
PostPosted: Thu May 21, 2009 4:02 pm Post subject: Re: A junk response to the issue... Reply with quote
CareyScurry wrote:
Isn't this two issues:
* A given scorer might be overly (dis-)inclined to grant assists and,
* A given scorer might show (unusually little/excessive) favoritism to home players
Yeah, it's a classic example of what econometricians call the "identification problem", when one observed statistical pattern can be explained by two or more competing hypotheses.
There's an additional complication: forget about the scorers, some teams have higher or lower true (meaning unbiased by scorers) assist rates. Which of course is obvious, that's basically what we're trying to measure in the first place. But in addition, what if some teams have higher or lower assist rates at home than on the road? Maybe they fastbreak more at home, or rely more on the 3-pointer on the road, or whatever. Now we have to disentangle that from the home-road scorer's bias, adding further to the identification problem.
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Fri May 22, 2009 7:59 am Post subject: Reply with quote
It was precisely to distinguish between these 2 tendencies that this thread was created. Home team 'favoritism', which I called 'bias' is what matters in determining a 'standardized' measure of players' assist rates.
Here are the 2009 Home and Away ratios of Ast/FG (Ast%) . A team's Away Ast% should be pretty close to their 'real' rates -- after all, these represent the average rates given them by the other 29 NBA teams; and the ratio of HAst%/AAst% reveals Home Assist Bias, or home jack factor.
The JacFac column represents Home + Away FG and Ast. So a player for that team could have his assists divided by this factor to get his season 'equivalent/standardized' assists (presuming he played equal home/away minutes).
Code:
2009 HAs% AAs% Jack JFac
Den .659 .529 1.245 1.111
Atl .608 .511 1.191 1.089
Dal .613 .522 1.173 1.082
LAL .619 .534 1.158 1.074
NOH .594 .513 1.157 1.072
Mil .641 .554 1.156 1.075
Was .583 .514 1.134 1.063
Cle .579 .520 1.114 1.055
Phl .577 .522 1.107 1.051
Por .578 .524 1.104 1.051
Bos .633 .576 1.099 1.049
NJN .584 .535 1.092 1.044
LAC .607 .559 1.086 1.041
GS .548 .509 1.077 1.038
Det .584 .549 1.063 1.031
tm HAs% AAs% Jack JFac
Hou .577 .546 1.056 1.028
Uta .658 .630 1.044 1.022
Chi .564 .544 1.035 1.017
Cha .616 .600 1.026 1.013
SAS .577 .564 1.022 1.011
Tor .602 .599 1.005 1.003
Orl .542 .546 .991 .996
Sea .550 .559 .984 .992
Mem .491 .503 .976 .988
Min .554 .568 .975 .987
Sac .526 .558 .942 .970
Ind .531 .572 .929 .963
NYK .527 .575 .916 .956
Phx .536 .596 .898 .945
Mia .513 .589 .871 .930
At the bottom are 9 teams who actually gave fewer assists at home.
Leaguewide, Away Ast/FG was .551 (home .579). So we see here several teams who (by official assist totals) seem to pass the ball less than the average team; but in fact, they pass/assist more than average.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
mtamada
Joined: 28 Jan 2005
Posts: 376
PostPosted: Fri May 22, 2009 1:37 pm Post subject: Reply with quote
Mike G wrote:
IA team's Away Ast% should be pretty close to their 'real' rates
It'd be very close to their real ROAD Asst%, but that's only going to be an approximation of their real overall (and home) Asst%. If a team's home Asst% is identical to its road Asst%, no problemo we're done. But that's an unverified (and difficult to verify) assumption: do teams have the same (real, unjacked) Asst% at home and on the road?
Back to top
View user's profile Send private message
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Fri May 22, 2009 3:59 pm Post subject: Reply with quote
mtamada wrote:
.. do teams have the same (real, unjacked) Asst% at home and on the road?
You mean, as in :
Quote:
Maybe they fastbreak more at home, or rely more on the 3-pointer on the road, or whatever.
?
Do fastbreaks get more Ast/FG, or do 3-pt attempts? I sure don't know. Nor would I suppose teams have a 'road game plan' and a 'home game plan'.
Teams' home Ast/FG tend to be distinctly higher than away Ast/FG. Actual assists (and FG) are not so dramatically different. All teams shoot more FT at home, and no assists are granted for these.
Maybe some scorekeepers are trying (perhaps unconsciously) to make up for passes which lead to a foul: 2 FT but no assist. If you were inclined to favor the counting of 'assisted fouls drawn', you might toss in an extra assist here and there.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 708
Location: Raleigh, NC
PostPosted: Fri May 22, 2009 5:00 pm Post subject: Reply with quote
Quote:
Do fastbreaks get more Ast/FG, or do 3-pt attempts?
I think this is an important question to answer before we start talking about any inherent bias. Hopefully I will have an answer on this soon.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Sat May 23, 2009 7:23 am Post subject: Reply with quote
Ryan, to get useful estimates wouldn't you just about have to visually review every single assist? How much additional adjustment would you optimistically expect to create?
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 708
Location: Raleigh, NC
PostPosted: Sat May 23, 2009 9:28 am Post subject: Reply with quote
I'm simply talking about looking at the AST/FGM ratios given how the play started.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Mon May 25, 2009 9:34 am Post subject: Reply with quote
Ryan J. Parker wrote:
... before we start talking about any inherent bias.
But the cat is long gone from the bag. I'll boldly predict that the extreme difference (1245/871) of some 43% easier-to-get assists in Denver (vs in Miami) will hardly be dented by any inference about home/away playing styles.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 708
Location: Raleigh, NC
PostPosted: Mon May 25, 2009 10:52 am Post subject: Reply with quote
Could you reference where the "1245/871 of some 43%" comes from? Not sure I understand you exactly.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3574
Location: Hendersonville, NC
PostPosted: Mon May 25, 2009 12:49 pm Post subject: Reply with quote
Denver tops the list above, at 1.245 as many Ast per FG at home (as away).
Miami gave their own team only .871 as many at home.
Drop the decimal, the ratio remains the same.
DWade got 7.2 Ast/G in Miami. Acquire the Nuggets' scorekeeper, and he gets 7.2 * 1245/871 = 10.3 Ast/G in Miami. Pretty dramatic, no?
page 3
Author Message
mtamada
Joined: 28 Jan 2005
Posts: 377
PostPosted: Wed May 27, 2009 12:56 am Post subject: Reply with quote
Mike G wrote:
Do fastbreaks get more Ast/FG, or do 3-pt attempts?
I sure don't know.
Precisely. So should we just assume neutrality of teams' performances, and statistics, when we compare their home Asst% to their road Asst%? Or is it possible that they differ?
Quote:
Nor would I suppose teams have a 'road game plan' and a 'home game plan'.
It doesn't have to be a difference in plans, simply a difference in the RESULTS, a difference in the statistics that they generate (meaning, the "true" statistics, not stats biased by scorekeepers). You're not denying that teams have different stats on the road vs at home are you? Including, quite possibly, a different Asst% on the road vs at home -- a true one, not one caused by scorekeeper bias.
IOW, we have an identification problem. A team's road Asst% gives us a good estimate of their true road Asst%, but we don't know what their true home Asst% should be. If it's higher, is it because of a friendly home scorekeeper? Or because they truly generate a higher Asst% at home?
Back to top
View user's profile Send private message
wmchad
Joined: 16 Apr 2009
Posts: 13
Location: Portland, OR
PostPosted: Wed May 27, 2009 9:25 am Post subject: Reply with quote
Quote:
It doesn't have to be a difference in plans, simply a difference in the RESULTS, a difference in the statistics that they generate (meaning, the "true" statistics, not stats biased by scorekeepers).
I think that, if you have the play by play data, you should be able to get better answers. Particularly, you can see what the Asst% is for various shot types and see how those change at home vs. on the road. That wouldn't eliminate the uncertainty entirely, but it may shed some additional light on the issue.
Back to top
View user's profile Send private message
mtamada
Joined: 28 Jan 2005
Posts: 377
PostPosted: Thu May 28, 2009 8:17 pm Post subject: Reply with quote
wmchad wrote:
Quote:
It doesn't have to be a difference in plans, simply a difference in the RESULTS, a difference in the statistics that they generate (meaning, the "true" statistics, not stats biased by scorekeepers).
I think that, if you have the play by play data, you should be able to get better answers. Particularly, you can see what the Asst% is for various shot types and see how those change at home vs. on the road. That wouldn't eliminate the uncertainty entirely, but it may shed some additional light on the issue.
That's an excellent idea, with a lot of potential. I don't know how well the play-by-play data distinguishes between various types of mid-range shots (open, covered, double-teamed), but 3-pointers on the one hand and dunks and layins on the other are clearly distinguishable and probably have different assist rates (especially if one leaves out the dunks and layins generated by offensive rebounds).
Back to top
View user's profile Send private message
NewCollegeHokage
Joined: 14 Apr 2009
Posts: 9
PostPosted: Thu May 28, 2009 11:06 pm Post subject: Reply with quote
Play by plays on basketball value. I've got a feeling there's a program to figure this out, Project Super Awesome Stats the Rockets Should be Giving us Anyways. If not, anyone want to split up some games and find Ast% for 3s, dunks and layups?
Back to top
View user's profile Send private message AIM Address
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Thu May 28, 2009 11:16 pm Post subject: Reply with quote
Quote:
anyone want to split up some games and find Ast% for 3s, dunks and layups?
Hey this is something I can do. Cool For the 08-09 season:
Code:
Type FGM AST AST%
3pt 16,024 13,751 85.8%
Dunks 9,135 6,649 72.8%
Layups 26,400 12,557 47.6%
Any data collection errors are my own.
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
wmchad
Joined: 16 Apr 2009
Posts: 13
Location: Portland, OR
PostPosted: Fri May 29, 2009 11:41 am Post subject: Reply with quote
Ryan J. Parker wrote:
Hey this is something I can do. Cool For the 08-09 season:
Code:
Type FGM AST AST%
3pt 16,024 13,751 85.8%
Dunks 9,135 6,649 72.8%
Layups 26,400 12,557 47.6%
Any data collection errors are my own.
If it isn't too much of a problem, can you break it down by home and away, as well as including 2pt jumpers (or other categories you have)? I assume that there are some missing shots here since the total assisted percentage is 63.9%, whereas from basketball reference the number is 56.5%.
Back to top
View user's profile Send private message
Ryan J. Parker
Joined: 23 Mar 2007
Posts: 711
Location: Raleigh, NC
PostPosted: Fri May 29, 2009 11:47 am Post subject: Reply with quote
I'll do you one better than that, but it'll take some time. The work I'm doing now will allow me to flow right into that sort of analysis, but I gotta get it done first. Very Happy
_________________
I am a basketball geek.
Back to top
View user's profile Send private message Visit poster's website
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Sun Feb 20, 2011 8:57 am Post subject: Reply with quote
This year's Home and Away Assist disparities, as of the All-Star break.
Code:
Assists per FG home Blocks per 100 Opp FGA home
tm away home A/H xAst tm away home A/H xBlk
LAC .545 .643 .848 108 Cha 5.51 8.72 .631 72
Atl .579 .668 .868 86 Chi 5.81 8.84 .657 71
Cle .573 .657 .872 78 LAC 4.37 7.08 .617 65
GSW .540 .603 .897 76 SAS 4.75 6.68 .712 44
Min .495 .561 .881 72 NOH 4.70 6.56 .716 40
Okl .509 .575 .884 68 Min 5.81 7.44 .782 38
Was .518 .584 .887 66 Was 6.62 8.36 .791 38
Chi .555 .611 .908 61 Sac 5.27 6.71 .785 34
SAS .574 .630 .911 59 Den 4.07 5.35 .761 32
NOH .556 .604 .921 49 GSW 4.86 6.07 .800 30
Den .538 .581 .927 48 Okl 6.09 7.30 .834 27
Phl .568 .616 .923 46 Ind 6.34 7.48 .847 26
Orl .535 .567 .943 35 Phx 4.42 5.39 .819 24
Cha .585 .618 .946 33 Phl 4.59 5.73 .801 24
Det .542 .571 .951 31 Uta 7.00 7.84 .893 20
Assists per FG home Blocks per 100 Opp FGA home
tm away home A/H xAst tm away home A/H xBlk
Tor .555 .581 .956 28 LAL 5.75 6.59 .872 19
LAL .562 .587 .956 27 Mil 6.17 7.12 .865 19
Bos .632 .656 .964 27 Det 4.95 5.64 .878 16
Dal .621 .638 .973 18 Mia 6.54 7.13 .917 12
NJN .569 .571 .997 2 Orl 5.10 5.57 .917 11
Mil .545 .540 1.009 -4 Hou 5.44 5.86 .928 10
Por .596 .585 1.020 -11 Cle 4.67 5.10 .917 9
Mia .548 .527 1.040 -19 Atl 4.92 5.34 .920 9
Ind .566 .547 1.035 -19 NJN 5.64 6.03 .935 9
Phx .607 .588 1.032 -21 Mem 6.42 6.73 .954 7
Sac .533 .510 1.045 -25 Tor 4.98 5.23 .953 6
Hou .621 .594 1.045 -29 Bos 5.64 5.86 .963 5
NYK .572 .545 1.050 -29 Dal 5.22 5.42 .963 5
Mem .529 .500 1.056 -30 NYK 7.51 7.18 1.046 -7
Uta .671 .620 1.082 -58 Por 6.63 5.85 1.133 -15
avg .565 .589 .961 26 avg 5.53 6.54 .854 23
xAst, or extra assists at home, could be subtracted from teams' assist totals, and their home Ast/FG ratios would then equal their away Ast/FG.
Same deal for xBlk.
_________________
`
36% of all statistics are wrong
Last edited by Mike G on Sun Feb 20, 2011 6:25 pm; edited 1 time in total
Back to top
View user's profile Send private message Send e-mail
greyberger
Joined: 27 Sep 2010
Posts: 50
PostPosted: Sun Feb 20, 2011 4:55 pm Post subject: Reply with quote
Thanks for posting this, Mike.
Back to top
View user's profile Send private message Send e-mail
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Mon Feb 21, 2011 10:16 am Post subject: Reply with quote
Finding 116 players who have at least 70 Assists at home, or at least 25 home blocked shots, or both, I've reviewed their home/away splits to find who seems to have big disparities.
The Clippers have the biggest Ast H/A disparity -- their Away Ast/FG ratio is just .545 (NBA avg = .565) and at home they're .643 . They're a below-average Ast team posing (at home) as an elite Ast team: Only the Jazz have a higher Ast% (.671 on the road).
Not surprisingly, then, an LAC player leads the league in 'extra' assists at home.
Code:
xHAs at Home tm HAst AAst HAs% AAs% eHAs
48 griffin,blake LAC 126 70 .214 .132 78
46 paul,chris NOH 298 259 .497 .420 252
38 parker,tony SAS 197 178 .398 .322 159
36 hinrich,kirk Was 124 86 .267 .191 88
35 rose,derrick Chi 259 176 .432 .374 224
34 jackson,stephen Cha 125 78 .216 .158 91
33 nash,steve Phx 349 240 .572 .518 316
32 billups,chauncey Den 150 121 .287 .226 118
31 wright,dorell GSW 115 62 .140 .102 84
31 gibson,daniel Cle 86 55 .226 .144 55
29 stuckey,rodney Det 132 83 .284 .222 103
27 horford,al Atl 105 78 .208 .155 78
26 harris,devin NJN 209 202 .456 .399 183
26 terry,jason Dal 138 105 .255 .208 112
24 ridnour,luke Min 155 96 .313 .265 131
24 ellis,monta GSW 176 123 .240 .208 152
Assist% = Ast/((TmFG*MPG/48)-FG)
which is % of teammate FG a player has assisted, while on the court.
Equivalent Home Assists: eHAst = (AAst%/HAst%)*HAst
That is, a players HAst if his Home Ast% were = his Away Ast%.
Extra Home Assists: xHAst = HAst - eHAst
Chris Paul has been the perennial leader in this dubious category. This year, it won't help him win the Assist title.
Rajon Rondo has just 1 questionable assist among his 316 in Boston. Either that, or he gets just as many away from home.
It's odd to see Nash up there, since most years he's been in the negative -- Phx' scorekeepers have not and still do not write down as many assists for the Suns as others' do.
His Away Ast% (.518) is just about the same as last year's official Ast% (.509).
Just a handful of players would have more than 14 additional assists if they could get them at home as readily as on the road:
Code:
xHAst tm HAst AAst HAs% AAs% eHAst
-31 holiday,jrue Phl 142 205 .255 .312 173
-30 gasol,marc Mem 51 99 .093 .146 81
-28 bibby,mike Atl 79 123 .159 .216 107
-27 frye,channing Phx 20 39 .031 .074 47
-27 mcgrady,tracy Det 85 104 .206 .270 112
There's Frye representing the Suns who don't get more assists at home. Just 42% as well, in fact.
Player and team minutes and FG also factor into the Ast%.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Mon Feb 21, 2011 8:54 pm Post subject: Reply with quote
Since I have them, here are dubious Block splits.
Code:
xHBk team HBlk ABlk HBk% ABk% eHBk
35 mcgee,javale Was 78 43 3.55 1.96 43
28 jordan,deandre LAC 63 31 2.64 1.47 35
25 bogut,andrew Mil 75 56 3.79 2.50 50
22 thomas,tyrus Cha 39 17 1.74 .76 17
18 dalembert,samuel Sac 51 28 2.18 1.39 33
18 milicic,darko Min 68 49 2.87 2.09 50
18 gibson,taj Chi 53 30 2.28 1.50 35
18 biedrins,andris GSW 31 11 1.26 .54 13
17 west,david NOH 33 17 1.52 .73 16
15 hibbert,roy Ind 54 39 2.41 1.74 39
xHBk team HBlk ABlk HBk% ABk% eHBk
-14 chandler,wilson NYK 27 42 1.19 1.81 41
-12 smith,josh Atl 39 55 1.84 2.39 51
- 9 williams,lou Phl 1 12 .05 .49 10
- 9 aldridge,lamarcu Por 28 42 1.41 1.86 37
- 8 wall,john Was 5 13 .23 .59 13
- 8 camby,marcus Por 30 43 1.51 1.90 38
Not much to report on the 'too few blocks at home' side. Probably random.
The top 40 shotblockers (>40 Blk) average 23% more at home (per Opp FGA), or 8 blocks in ~24 G.
_________________
`
36% of all statistics are wrong
Back to top
View user's profile Send private message Send e-mail
kjb
Joined: 03 Jan 2005
Posts: 865
Location: Washington, DC
PostPosted: Tue Feb 22, 2011 5:53 pm Post subject: Reply with quote
Is there a similar split with steals?
With assists, I wonder how much is home scorekeeper bias and how much is the old "teams play worse on the road" thing. The Wizards, for example have been awful on the road this year. I have no idea how to tease out the difference. Are they getting fewer assists because they're shooting worse? Are they shooting worse because their passing is worse?
Back to top
View user's profile Send private message AIM Address Yahoo Messenger
Mike G
Joined: 14 Jan 2005
Posts: 3597
Location: Hendersonville, NC
PostPosted: Tue Feb 22, 2011 10:12 pm Post subject: Reply with quote
What's distinguishing teams here is not their home/away FG% or FG/G, but their ratio of Ast to FG, which is higher at home for 2/3 of teams.
For the Wizards:
Code:
per game FG FGA FG% Ast Ast/FG
home 37.1 83.0 .447 21.7 .585
away 36.7 82.7 .444 18.9 .515
Very little difference in FG, FGA, or FG% home and away.
In the table further up the page, they're shown with 66 'extra' home assists this year.
Over 28 home games, that's about 2.4 per game.
Subtract that from 21.7 Ast/G at home, and their home and away Ast/FG are both .515 .
It's not really better or worse to have 51 vs 58% of your FG assisted. But the higher figure does pad your players' assist totals.
Some FG are assisted, some missed FG are blocked, and some TO are steals. There are more occasions to insert another Ast or Blk among all the made and missed shots.
Sure, someone could or should look at steals, home and away.