Basketball Prospectus: Unfiltered Everything Else is Fluff.

January 18, 2012

Referee Ratings

Filed under: Uncategorized — Neil Paine @ 8:58 am

Random question of the day: Which referees are contributing most to offensive or defensive basketball?

To find the answer, I ran a variation on Basketball-Reference‘s Simple Rating System. Like in the SRS, each team was assigned an offensive and defensive rating that minimized the squared error when predicting the outcome of every game… However, I also tacked on variables for all 60 NBA referees, in an attempt to determine their impact on the outcome, independent of the quality of teams playing or home-court advantage effects (assuming referees’ impact on offense/defense is equally allocated to both teams).

Here were the resulting team ratings (all numbers are through Monday’s games):

Team         Poss       Off    Def   Ovr
----------------------------------------
1.  PHI     1200.4      4.5   -7.6  12.1
2.  CHI     1328.8      1.2   -8.4   9.6
3.  MIA     1200.2      0.8   -7.0   7.8
4.  ATL     1286.1      6.2   -0.5   6.7
5.  OKC     1278.8      6.8    0.9   5.9
6.  SAS     1199.4      5.7    0.4   5.3
7.  POR     1222.8      1.6   -3.5   5.1
8.  DEN     1250.6      4.7    0.2   4.5
9.  LAC      907.7      6.5    2.4   4.1
10. LAL     1363.2     -1.1   -4.6   3.5
11. DAL     1285.2      0.2   -3.1   3.3
12. ORL     1073.7      4.6    1.3   3.3
13. IND     1094.5     -2.0   -5.1   3.1
14. MIN     1201.5      1.8    0.7   1.0
15. MEM     1108.7     -3.2   -4.1   0.8
16. CLE     1118.6      1.1    0.4   0.7
17. UTA     1103.7     -2.0   -2.6   0.6
18. HOU     1197.5      3.0    2.9   0.1
19. PHO     1083.5      1.9    3.3  -1.4
20. MIL     1099.2     -3.4    0.3  -3.6
21. GSW     1108.1      0.0    4.4  -4.4
22. NYK     1208.2     -3.9    0.5  -4.4
23. NOH     1142.0     -5.5   -0.9  -4.6
24. BOS     1051.0     -1.1    4.0  -5.1
25. TOR     1232.9     -1.8    3.5  -5.3
26. NJN     1242.3      0.9    7.6  -6.7
27. SAC     1300.9     -5.7    4.0  -9.7
28. DET     1125.6     -2.4    7.5  -9.9
29. CHA     1305.5     -7.3    4.3 -11.6
30. WAS     1196.8    -11.4    0.8 -12.2
----------------------------------------
    HCA                              5.2
    Lg Constant                    103.2
----------------------------------------

And, of course, the effect of each referee:

Referee           Poss  Rating
-------------------------------
KarlLane          798.6   8.7
MarcDavis         729.9   8.3
ScottTwardoski    905.1   7.7
JamesWilliams     932.3   7.1
JTOrr             891.5   7.1
HaywoodeWorkman   839.5   6.0
KaneFitzgerald    992.2   5.9
KevinFehr         713.4   5.8
MichaelSmith      914.9   5.5
EliRoe            839.0   5.3
OlandisPoole      927.5   4.5
KenMauer          819.0   3.9
ZachZarba        1041.7   3.8
MaratKogut        810.4   3.7
LeonWood         1014.8   3.7
VioletPalmer      804.8   3.7
EdMalloy          919.7   3.0
NickBuchert      1027.1   2.7
JoshTiven         935.3   2.2
KevinCutler       934.4   1.9
MarkLindsay       921.9   1.6
DanCrawford       838.4   1.3
DerekRichardson  1021.5   1.2
GregWillard       900.7   1.1
TreMaddox         721.4   0.8
CurtisBlair       886.5   0.8
BrianForte        828.2   0.5
TommyNunezJr      912.8   0.3
EricDalen         913.3   0.3
MikeCallahan      805.8   0.3
BennieAdams      1016.3   0.2
JoeCrawford       914.8   0.0
TonyBrown         890.8  -0.1
PatFraher         856.1  -0.4
TomWashington     916.7  -0.6
DerrickCollins   1023.9  -0.6
ScottWall         842.8  -0.7
JohnGoble         904.6  -1.2
RodneyMott       1005.3  -1.5
GaryZielinski     845.9  -2.2
MarkAyotte        897.6  -2.7
MontyMcCutchen   1000.2  -2.9
RonGarretson     1001.3  -2.9
DickBavetta       827.2  -3.3
BrentBarnaky      730.1  -3.4
DavidGuthrie      661.6  -3.4
ScottFoster       820.6  -3.4
LeroyRichardson   897.8  -3.9
JamesCapers       838.9  -4.1
EricLewis        1000.4  -4.4
EddieF.Rush       731.3  -4.5
TonyBrothers     1006.9  -4.6
DerrickStafford  1025.4  -5.4
MattBoland        821.3  -5.8
SeanWright        821.3  -5.9
BillKennedy       988.4  -6.4
JasonPhillips     801.8  -6.8
BillSpooner       838.3  -7.9
CourtneyKirkland  895.9  -8.8
DavidJones        911.1 -10.7
-------------------------------

How do you interpret these?

Well, for any given matchup, the home team’s offensive rating (aka the road team’s defensive rating) can be predicted by:

ORtg_H = Lg_Constant + .5*HCA + Ref1_Rtg + Ref2_Rtg + Ref3_Rtg + Home_Off + Road_Def

Likewise, the road team’s offensive rating can be predicted by:

ORtg_R = Lg_Constant – .5*HCA + Ref1_Rtg + Ref2_Rtg + Ref3_Rtg + Road_Off + Home_Def

This means, for instance, that games officiated by Karl Lane feature offensive ratings 8.7 pts/100 poss. higher than average, all else being equal, while games officiated by David Jones have ratings 10.7 pts/100 poss. lower than average.

Although the 2012 season is a small sample right now (making the standard errors on these estimates quite large) those are some pretty big referee effects!

Email Neil at np@sports-reference.com. Follow him on Twitter at @Neil_Paine.

No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress