Alternative Post Timeout Ratings

One of the aspects to having a serious, analytical and yes critical audience here at 82games.com is that we are constantly reminded by our readers if they detect anything "scientifically unsound" in an article or data set we may post.

...Was ...Now

Such has been the case with the recent Post Timeout Performance article where we showed the stats and ranks for the various teams in the events immediately following a timeout. Now we had the nerve to throw up some rankings and an overall "score" at the end, which prompted a deluge of emails stating the same thing: raw numbers post timeout should not be used, but instead the difference between a team's full season points per possession efficiencies (offense and defense) and the post timeout ones.

Now it's a more complex issue than this -- what are the league wide expectations following a timeout versus "all game state" moments? What about adjusting for the varying expectations by the actual state (having 0.9 seconds left in the game from the inbounds versus a full 24 seconds, garbage time versus clutch time, etc)? The more frequent "in the bonus" state, intentional fouls, etc...

Still, we aim to please (well occasionally) and so here are the "alternative rankings" for your entertainment.

Post-Timeout "Difference in Performance"

These rankings reflect the difference in team performance between post timeout plays and the full season numbers for each team.

Legend:
Off1 = 1st Offensive possession post timeout when team had ball from inbounds
Off2 = 1st Offensive production when team was on defense coming back from the timeout
Def1 = 1st Defensive possession when team was on defense after the timeout
Def2 = 1st Defensive production when team was on offense coming back from the timeout
Score = weighted ranks formula: [3*Off1)+(3*Def1)+Off2+Def2]/8

Rank	Team	Off1	Off2	Def1	Def2	Score	Wins
1	Utah	2	6	10	2	5.5	41
2	Denver	3	20	5	16	7.5	44
3	Orlando	8	13	4	17	8.3	36
4	New York	11	5	11	4	9.4	23
5	New Orleans/Okla	5	9	8	29	9.6	38
6	Seattle	6	10	18	7	11.1	35
7	Miami	24	4	2	12	11.8	52
8	New Jersey	9	19	15	5	12.0	49
9	Boston	1	12	27	3	12.4	33
10	Charlotte	15	23	7	11	12.5	26
11	Chicago	21	2	9	19	13.9	41
12	Dallas	22	22	3	23	15.0	60
13	Golden State	4	7	25	26	15.0	34
14	Philadelphia	26	16	1	24	15.1	38
15	Memphis	13	1	26	8	15.8	49
16	Indiana	14	21	13	25	15.9	41
17	Milwaukee	20	25	14	6	16.6	40
18	Portland	19	8	16	20	16.6	21
19	Detroit	25	17	6	27	17.1	64
20	Sacramento	7	28	24	18	17.4	44
21	L.A. Lakers	23	15	12	21	17.6	45
22	Washington	16	14	22	15	17.9	42
23	Phoenix	12	30	19	28	18.9	54
24	Atlanta	28	3	17	14	19.0	26
25	Toronto	18	18	21	22	19.6	27
26	L.A. Clippers	10	29	29	13	19.9	47
27	San Antonio	29	24	20	1	21.5	63
28	Houston	17	26	23	30	22.0	34
29	Cleveland	27	11	30	10	24.0	50
30	Minnesota	30	27	28	9	26.3	33

So originally on the "raw look" we had the Dallas Mavericks as the best team in the NBA post timeout, but when you start adjusting by overall performance things change quite a bit. For example, the Mavs 113.3 points per 100 possessions number on offense immediately after a timeout ranked 6th in the league, but when you factor in their overall offensive prowess, they were a mere +1.8 points/100 poss improved, which ranks #22 in the league under the revised scheme.

Similarly the Portland Trailblazers who were the worst offensive team in the league this season for overall efficiency, ranked #29 in the raw post-timout offense ranking, but #19 in this alternative view, largely because there was basically nowhere to go but up!

To our minds this approach really isn't the answer since it penalizes the good teams and rewards the bad teams.

Still for those who disagree, your new overall "best timeout coach" becomes Jerry Sloan of the Utah Jazz who ranked a respectable third in the first version.

The correlation to wins goes way, way down and actually turns negative (again due to the basic reward bad overall teams/punish good overall teams issue):

Off1 = -.21
Off2 = -.27
Def1 = .10
Def2 = -.04
Score = -.15

Only three of the top ten teams made the playoffs, while six of the bottom ten qualified for the post-season...

We prefer version one!