Layups: More Disturbing Stories on the Book-Cooking Front
Posted by Neil Paine on August 28, 2009
Deadspin first brought us the story of Nick Van Exel's 23, um, creatively assigned assists in a game, and now they have more sordid tales of NBA statstical manipulation, including the use of assists, blocks, and steals as PR devices for popular players. Meanwhile, to read the statistical communty's reaction, go here, here, and here.
August 29th, 2009 at 12:45 am
I can see why this would offend the sensibilities of fantasy basketball enthusiasts. It would suck to lose a roto league because a homer scorekeeper gave Ramon Sessions 57 assists on the final game of the season.
August 29th, 2009 at 10:18 pm
If it's a league wide phenomenon don't the effects get washed out?
August 30th, 2009 at 3:25 am
I'd always heard Utah was notorious for doing this, so I decided to take a look at John Stockton's splits throughout his career for home and road games:
Home
Games - 669
Assists - 7644
APG - 11.4
Road
Games - 671
Assists - 7137
APG - 10.6
So, over Stockton's entire career, he played in almost the exact same number of home and road games (actually two more road than home) and had 507 more assists at home than he did on the road, or 0.8 more APG. I have no idea whether that's a high amount or not, but it seems like it. It would be interesting to see what the highest differential is between home and road assists for an elite point guard.
The actual year by year breakdown is: (Total/Per Game)
86-87: 64/1.6
87-88: 60/1.5
88-89: 30/0.7
89-90: 70/1.8
90-91: 88/2.1
91-92: 36/0.9
92-93: 15/0.4
93-94: 19/0.5
94-95: -29/-0.7
95-96: 12/0.3
96-97: -10/-0.2
97-98: 41/1.8
98-99: 32/1.3
99-00: 51/1.2
00-01: 13/0.3
01-02: 26/0.6
02-03: -11/-0.3
August 30th, 2009 at 8:30 pm
I forgot to mention this before but I didn't include Stockton's first two years in his assist totals since I couldn't find home/road splits for them. And the year by year breakdown should be read as 1986-87: 64 more assists in home games than road games, or 1.6 more APG.
August 30th, 2009 at 9:19 pm
SO Luke, we conclude that Utah started handing out extra assists, until the rest of the league caught on / Stockton gained a "reputation" and then it balanced up with both home and Away figures having the extras :)
August 31st, 2009 at 5:25 am
Wasn't Utah also notoriously better at home then on the road? Were the statkeepers cooking the actual score as well?
August 31st, 2009 at 8:04 am
Luke, that's very interesting. Stockton in 1989-90 did have 70 more assists at home, or 1.8 per home game. But also note that the Jazz shot better at home, making 2.5 more FG per game:
http://www.basketball-reference.com/fc/tsplit.cgi?team=UTA&year=1990
The team got 1.8 more Ast/G at home, and Stock got all of them. Their ratio of Ast/FG was .666 at home and .662 on the road. Thus, they were actually one of the 5 _least_ home-friendly assist-giving crews that year.
As Mrparker says, one should consider actual (point differential) changes in home/away performance, before attributing all H/A difference to scorekeeper bias.
No one (we think) had an assist title taken away between 1987 and 2007. Read more here:
http://sonicscentral.com/apbrmetrics/viewtopic.php?t=1897
August 31st, 2009 at 12:06 pm
I could be way off here, but it might be good to check Stockton's foul's per game and minutes per game home and away as well. It is my recollection that the Jazz tended to get away with more physicality in Salt Lake than away (which I'm sure is true of most teams), so playing time / ability to remain aggressive might have had something to do with his assist distribution being more prolific at home.
August 31st, 2009 at 4:02 pm
For comparison - here's Kidd's year by year home - road assists:
'95 8.1 - 7.2 = 0.9
'96 10.2 - 9.2 = 1
'97 9.5 - 8.5 = 1
'98 9.6 - 8.6 = 1
'99 10.6 - 11 = -0.4
'00 10.6 - 9.6 = 1
'01 10.8 - 8.7 = 2.1
'02 11 - 8.7 = 2.3
'03 9.2 - 8.3 = 0.9
'04 9.1 - 9.4 = -0.3
'05 9.3 - 7.3 = 1
'06 9.2 - 7.6 = 1.6
'07 9.5 - 8.9 = 0.6
'08 10.2 - 9.9 = 0.3
'09 9.9 - 7.5 = 2.4
I don't want to take the time to total his home / away numbers right now, but this actually looks a little more home-biased than Stockton's list. We might want to check this against a great scorer too, just to see if a relatively objective stat like points swings like this.