FanPost

Observations on the Rotations

Tuesday, TrueHoop interviewed statistician Wayne Winston, in "Ten Weeks In, Ten Questions with Wayne Winston." I posted the follow quote that grabbed my attention as a FanShot:

10. You have told me that many coaches do things you simply cannot understand, for instance by ignoring some really good lineups. What are some of the best lineups from this season, and are some of them not being played enough?
As I described above, the Lakers have the best lineups. After that, some of the best in the League are:

....
Magic: Ryan Anderson, Matt Barnes, Dwight Howard, Jameer Nelson and J.J. Redick Rating +56
Magic: Matt Barnes, Vince Carter, Rashard Lewis, Dwight Howard and Jason Williams Rating +36

So….what? Sit Carter and Lewis?  Start Jason Williams? What?!

After back-to-back-to-back losses (Chicago, Indiana, Toronto), I’ve spent some time looking at the lineup data available at 82games.com (thru Jan 4, 2010) and basketballvalue.com (thru Jan 6, 2010).  The Top Five-Man Floor Units from 82games are listed below.  This list is "Top-10" based on minutes played together, not team performance.

Top Five-Man Floor Units

#

Unit

Min

Off

Def

+/-

W

L

Win%

1

 Williams-Carter-Pietrus-Lewis-Howard

273.7 

1.10 

1.06 

+21  

11  

7  

61.1

2

 Nelson-Carter-Pietrus-Lewis-Howard

65.8 

1.05 

0.89 

+24  

4  

1  

80.0

3

 Nelson-Redick-Barnes-Anderson-Howard

58.7 

1.56 

0.95 

+70  

6  

0  

100

4

 Nelson-Carter-Barnes-Lewis-Howard

56.4 

1.03 

1.07 

-4  

2  

3  

40.0

5

 Nelson-Redick-Barnes-Bass-Howard

41.6 

0.94 

1.24 

-25  

0  

4  

0.0

6

 Williams-Carter-Barnes-Lewis-Howard

41.2 

1.39 

1.15 

+22  

8  

4  

66.6

7

 Williams-Redick-Barnes-Anderson-Howard

35.2 

1.08 

1.09 

+10  

7  

1  

87.5

8

 A.Johnson-Redick-Barnes-Anderson-Gortat

33.8 

0.98 

1.35 

-25  

4  

5  

44.4

9

 A.Johnson-Redick-Barnes-Anderson-Howard

31.8 

1.29 

1.02 

+11  

4  

3  

57.1

10

 Williams-Redick-Pietrus-Lewis-Howard

29.8 

1.25 

1.11 

+5  

4  

3  

57.1

Top Five-Man Floor Units, Details

#

Unit

eFG

eFGA

FTA

Close

dClose

Reb

T/O

1

 Williams-Carter-Pietrus-Lewis-Howard

.470 

.441 

+26 

31% 

29%  

48.0% 

-4% 

2

 Nelson-Carter-Pietrus-Lewis-Howard

.418 

.457 

+25 

29% 

28%  

52.5% 

+6% 

3

 Nelson-Redick-Barnes-Anderson-Howard

.561 

.422 

+4 

47% 

27%  

64.5% 

+2% 

4

 Nelson-Carter-Barnes-Lewis-Howard

.478 

.422 

-9 

43% 

21%  

48.0% 

-8% 

5

 Nelson-Redick-Barnes-Bass-Howard

.397 

.542 

+3 

43% 

32%  

48.5% 

-3% 

6

 Williams-Carter-Barnes-Lewis-Howard

.556 

.434 

+15 

24% 

26%  

47.0% 

-3% 

7

 Williams-Redick-Barnes-Anderson-Howard

.468 

.474 

+6 

39% 

28%  

51.5% 

-3% 

8

 A.Johnson-Redick-Barnes-Anderson-Gortat

.382 

.556 

+1 

18% 

24%  

46.5% 

-8% 

9

 A.Johnson-Redick-Barnes-Anderson-Howard

.477 

.415 

+14 

27% 

23%  

55.5% 

-3% 

10

 Williams-Redick-Pietrus-Lewis-Howard

.595 

.429 

+14 

24% 

27%  

38.0% 

-1%

In looking at the lineups here are some of my thoughts about how the Magic might improve their rotation.

Assumptions:

1.  As has been pretty well defended at this site by Ben and Eddy, playing Bass with Howard is worse for the team than Lewis or Anderson (see line #5 and compare to #3).  Unit #5 leaves the discussion for now.

2. Anthony Johnson is clearly our 3rd PG.  Units #8 & 9 leave the discussion for now.

3. Howard is better than Gortat.  Other depth chart positions are open for discussion.

4. Getting below the top 10 units in minutes-per-unit makes drawing conclusions more difficult, ‘cause the sample size shrinks.  For the most part, I will keep to the remaining lineups in this list (#1-4, 6, 7, & 10). Though to figure out who plays best with Gortat, I will use data from basketballvalue.com to assess lineups that have played together less than the above-listed 29.8 minutes.

5. I know that stats are descriptive rather than prescriptive.  So, it could be that particular lineups have just coincided with the players’ lack of energy and hustle and/or injuries,  and therefore are really reflective of motivation problems rather than on-floor chemistry issues.

6. Basketball is a 48-minute game, so it is more than just putting the best five players on the floor together to have the single best lineup.  Plenty of teams have stellar starters, but lose games when their bench hits the floor. It is more important to find the best rotation that gives you the best combination(s) versus your opponent over four quarters.  The Magic are plenty deep—maybe the deepest—but are the combinations out of sync?  I know the team’s had to adjust for injuries, illnesses, and suspension, but when everyone is available, what works best?

Observations:

A. The lineup we trotted out the last five games (& three losses) Nelson-Carter-Barnes-Lewis-Howard (#4) is not clicking and hasn’t since it was put in place following the Christmas Day loss to Boston.

B. It seems like units that have Pietrus over Barnes (#1 vs. 6 or #2 vs. 4) with Howard/Lewis/Carter and a PG, have lower (ie. better) Defensive Points Per Possession numbers.  So, Pietrus is a better defender?  Well, seems so.  Four of the six units that have played 18 or more minutes together this season and hold opponents to a sub-0.90 Def rating have Mickael in the lineup (and one of those does not even include Howard!). [Factoid: Barnes is not a defensive slouch…some really killer (i.e. 0.80ish rating) defensive units have Barnes and Pietrus together…again, sample size is small.]

C. Seems Nelson’s units are always better than Williams’ on Defensive Points Per Possession (#2 vs. 1 or #4 vs. 6).  This seems to parallel the conventional view that Nelson is a better defender than Williams.

D. Nelson-Redick-Pietrus-Anderson-Howard does not appear not only in the top 10, it just does not seem to get used.  Not sure why when it is pretty similar to…..

E. The #3 lineup looks like the best unit.  Nelson-Redick-Barnes-Anderson-Howard.  In 58.7 minutes of floor time (nearly five quarters), they have the best eFG to eFGA differential, the best Off. to Def. differential, the greatest percentage of close-in shots, don’t really give up any more close in shots to opponents than any other combo, rebound nearly 12% higher, and have the 2nd best TO ratio.  At first I wondered if this was because this unit plays versus the opponent’s 2nd unit.  It is not.  Often this unit plays when the opponent has one or two bench players in.  Just to show the Redick/Anderson flow, swap Nelson out and put Williams in and you have lineup #7, which with over two quarters of time has the 2nd best W:L ratio at 7:1, but not as good stats as the units below, in F.

F. Our 2nd best combo is a pretty close "tie" between #1, 2, and 6 depending on what stat is most important to you.  In all three you have Howard/Lewis/Carter.  Comparing the units, Williams/Barnes in #6 is the best offensively (and best differentials) and Nelson/Pietrus in #2 is the best defensively.

G. Missing from this data set is who plays well with Gortat (unit #8 is awful).  Basketballvalue.com’s data shows that Gortat does better is certain, select lineups.

i. Do not ever play Gortat and Lewis together.  Ever.  Horrible.  For example, Williams-Carter-Pietrus-Lewis-Gortat in nearly 20 minutes is 0.90 Off rating and Def. rockets to 1.30.  Most of the Adj +/- for Gortat/Lewis units are similarly in the -40 range.

ii. Gortat gets some of his best defensive numbers with Brandon Bass at the 4. That surprised me a little bit. And, the offense is not too bad either. [Factoid: But Howard and Bass do NOT work well together.  Neither do Gortat and Howard.]

iii. Gortat does well defensively with Anderson, but that combo works better with Pietrus than Barnes.  If Barnes and Gortat play together, it worked against Charlotte’s subs to go small play Barnes at the 4.  Best realistic Gortat lineups based on this year’s small sample size?

Gortat-Bass-Pietrus-Redick-Williams (26 minutes, Adj+/- is +16.85)

Gortat-Anderson-Pietrus-Redick-Williams (12 minutes, Adj +/- is +59.09)

Gortat-Anderson-Pietrus-Redick-Nelson (10 minutes, Adj +/- is +24.31)

iv. Despite that third lineup in iii., Gortat and Nelson are not nearly as good as Gortat and Williams.

So…..what does all this mean for the rotation?

Conclusions

Play Howard all game.  Just kidding.

Pietrus to start at the SF. In my opinion, Pietrus works better as a starter with Howard-Lewis-Carter-Nelson.  I know this is the lineup that got killed by Boston, but….  Van Gundy tinkered with the lineup by putting Barnes in at SF just after this Christmas loss and the team has played poorly ever since…even in the wins.  Maybe they were psyched out by the loss to the Celtics, but maybe Barnes doesn’t click with the starters.  Barnes is played lights out for four games (laid an egg in Toronto), but everyone else has been playing lights off.

Or, Williams to start at PG. Williams-Carter-Barnes-Lewis-Howard is the offensively potent lineup #6.  Jameer would be a great 6th man candidate and would finish out games with Carter-Pietrus-Lewis-Howard (lineup #2).

No Gortat and Lewis together.  If Howard gets in foul trouble and Gortat comes in, based on the numbers, Lewis needs to sit, too.  Can keep trying Nelson with Gortat, but keep an eye on whether or not this works.

Gortat plays better with Bass.  Here’s where Bass could get his minutes.

Gortat plays better with Pietrus.  This is a dilemma because if Nelson starts, Nelson/Carter/Lewis/Howard also play better with Pietrus than with Barnes.  However, if Williams starts, putting Barnes in at SF is better offensively than having Pietrus.  This frees Pietrus to platoon with Gortat off the bench and then finish the game with the Nelson/Carter/Lewis/Howard unit.

Play the +56 lineup!  Nelson-Redick-Barnes-Anderson-Howard.  It worked against Toronto, Phoenix, and Detroit.  Try it some more!

I have neither the smarts nor the time to put together a full rotation for the team based on this information.  Here are lineups that I think work, maybe somebody can sort out the minutes in the comments.

Start either:

Williams-Carter-Barnes-Lewis-Howard (#6) – 1st choice or,

Nelson-Carter-Pietrus-Lewis-Howard (#2)

Biggest offensive punch so far this year:

Nelson-Redick-Barnes-Anderson-Howard # 7 (rest Williams, Carter, Lewis)

When Howard is out, either:

Williams-Carter-Pietrus-Bass-Gortat or, (rests Nelson, Barnes, Lewis, Howard) or,

Williams-Redick-Pietrus-Anderson-Gortat (rests Nelson, Carter, Barnes, Lewis, Howard)

Finish with:

Nelson-Carter-Pietrus-Lewis-Howard (#2)



Feel free to comment, correct, commend, debate, disagree, or digest.  Thanks for reading.

This FanPost was made by a member of the Orlando Pinstriped Post community, and is to be treated as the opinions and views of its author, not that of the blogger or blog community as a whole.

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Orlando Pinstriped Post

You must be a member of Orlando Pinstriped Post to participate.

We have our own Community Guidelines at Orlando Pinstriped Post. You should read them.

Join Orlando Pinstriped Post

You must be a member of Orlando Pinstriped Post to participate.

We have our own Community Guidelines at Orlando Pinstriped Post. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9347_tracker