Kevin Pelton, SUPERSONICS.COM | January 27, 2005
Oliver is reluctant to rate players individually, but when he does so, he uses his Individual Offensive and Defensive Ratings. We'll let Oliver himself explain the theory, from his book,
Basketball On Paper, where you can also find the actual formulas.
Individual Offensive Ratings: Individual offensive ratings are constructed from the following two statistics:
Individual possessions, which represents the credit an individual gets for the times his/her team ends a possession and gives it back to the opponents.
Individual points produced, which represents the credit an individual gets for the points that his/her team generates on the offensive end.
Individual Defensive Ratings: The premise of individual defense is that players force "stops," preventing the other team from scoring. … An individual can do that by forcing a missed shot that then gets rebounded by his team, by getting a defensive rebound, by forcing a turnover, or by fouling a player who misses both foul shots, the second of which is then rebounded by the defense.
From the individual ratings, a player's winning percentage can be determined using the Pythagorean method for estimating winning percentage from offensive and defensive ratings. This takes the player's offensive rating to the power of 14 and divides by his/her offensive rating to the power of 14 plus his/her defensive rating to the power of 14.
The number of "games" a player is "responsible" for is based on the percentage of the team's possessions the player is responsible for on offense, the percentage of team stops on defense, minutes played and games started. The winning percentage times games is the number of wins the player has created, the remaining games losses.
To see the formulas in work, here are the ratings for the 1995-96 Western Conference Champion Sonics, courtesy Basketball-Reference.com, which has ratings for every team dating back to the 1977-78 season. The league's average rating in 1995-96 was 107.7.
| Player |
ORtg |
DRtg |
W |
L |
Win % |
| Shawn Kemp |
111 |
97 |
11.7 |
1.7 |
.872 |
| Gary Payton |
112 |
102 |
11.4 |
3.0 |
.792 |
| Hersey Hawkins |
116 |
104 |
8.9 |
2.0 |
.819 |
| Detlef Schrempf |
118 |
105 |
7.3 |
1.5 |
.829 |
| Sam Perkins |
113 |
103 |
6.8 |
1.9 |
.784 |
| Vincent Askew |
112 |
106 |
4.2 |
1.8 |
.700 |
| Ervin Johnson |
103 |
98 |
4.1 |
2.1 |
.662 |
| Nate McMillan |
109 |
99 |
3.9 |
0.9 |
.807 |
| Frank Brickowski |
100 |
104 |
1.3 |
2.3 |
.369 |
| Eric Snow |
95 |
101 |
0.5 |
1.2 |
.278 |
| David Wingate |
97 |
107 |
0.5 |
2.0 |
.190 |
| Sherell Ford |
104 |
102 |
0.4 |
0.3 |
.558 |
| Steve Scheffler |
106 |
104 |
0.3 |
0.3 |
.538 |
|
There is a temptation to look solely at the player's winning percentage, but it is important to note that if an efficient role player like Hawkins or Schrempf was forced to play Payton's and Kemp's go-to roles, his winning percentage would likely go down.
