Looking at Lineups: Finding Minutes For What Works

by Ian Levy on Sep 12, 2011 7:55 PM EDT in Indiana Pacers Stats

Without the the certainty of a new season to look forward to, I've found myself spending more and more time mulling over last year. In particular I've been trying to quantify exactly how I feel about Frank Vogel becoming the permanent head coach of the Pacers. Like many Pacers' fans I was both excited and disappointed at times by the Pacers' with Vogel at the helm. I did however find it refreshing to watch how the team changed under their new coach.

The two things I found most frustrating about Jim O'Brien's tenure were a lack of diversity at the offensive end, and the insistence on including ineffective players in his rotations. Specifically, the early season minutes that went to Solomon Jones and James Posey had me pulling my hair out. During his time in Indiana, O'Brien was never blessed with much in the way of depth, but I still found myself constantly frustrated by how minutes were divided.

In a post I just finished at Hickory-High, I tried creating a method to determine how effective different NBA coaches have been at managing their rotations. For each coach I looked at the correlation between the number of minutes they played each five-man unit and how effective that unit was. To measure the effectiveness of each unit I used the Net Rating (Offensive Rating - Defensive Rating) and weighted it by how many minutes that unit had played.

It came as no surprise to me that O'Brien did not perform very well in this analysis. From 2008 to 2010, his full season correlations were -0.224, 0.159 and 0.094. This means at best, the statistics show no connection between how effective a lineup was and how often O'Brien used them. His cumulative correlation for that time span was 0.004, which placed him 31st of the 45 coaches who worked during those seasons. The entire results can be viewed here.

I pulled all the data for this analysis from Basketball Value, a wonderful site on which I rely heavily. One unfortunate limitation to their data is that it is only available for entire seasons. This means I couldn't include in my coaching analysis the numbers from any teams which switched coaches during a season, as the Pacers did last season. Luckily for the readers of Indy Cornrows, I happen to have saved a spreadsheet (downloaded from BV) of the Pacers' five-man unit data from just after the firing of O'Brien, and the promotion of Vogel. I ran the same correlation studies for this data and found a somewhat discouraging set of numbers.

Before his firing, O'Brien had a 0.468 correlation between the effectiveness of a lineup and how many minutes that lineup was used.
After taking over for O'Brien, Vogel had a 0.070 correlation between the effectiveness of a lineup and how many minutes that lineup was used.
During the playoff series against the Bulls, Vogel had a -0.571 correlation between the effectiveness of a lineup and how many minutes that lineup was used.

It certainly doesn't make one feel warm and fuzzy to see numeric evidence that O'Brien managed the effectiveness of his units better than his replacement, or that the Pacers' playoff rotations were a disaster. There are several explanations for this pattern, but the crucial change seems to be found in the starting lineups used by each coach.

The lineup which played the most minutes under Jim O'Brien was the Collison-Dunleavy-Granger-McRoberts-Hibbert configuration. That lineup, which frequently started games for the Pacers at the beginning of the season, played just over 240 minutes together under O'Brien, posting a terrific Net Rating of +15.02. This lineup played just over 203 minutes under Frank Vogel during the rest of the season, mostly maintaining their high level of performance with a Net Rating of +11.73. The starting lineup used by Frank Vogel, Collison-George-Granger-Hansbrough-Hibbert, had played just under 5 minutes together before Vogel took over, posting a Net Rating of -80.56. In the 324 minutes they played together under Vogel, they had a Net Rating of -2.19.

There is no denying that Vogel did make some positive changes to the Pacers' rotations. The two that probably had the biggest impact were shoving James Posey and Solomon Jones to the end of the bench. However, he also took the Pacers' starting lineup, which had been the most effective lineup of the six units which had played at least 40 minutes under O'Brien, and replaced it with a negative. He committed himself to a starting lineup that, although exciting, was much less effective than that of his predecessor. In doing so he was heeding the (mostly) unspoken requirement of being promoted, which was finding more minutes for Paul George, Tyler Hansbrough and A.J. Price. He did as promised, moving George and Hansbrough into the starting lineup. While this provided oodles of opportunities for player development it also took minutes from the Pacers' previous starting lineup, which had been very effective.

That is all understandable to some degree. The Pacers' goals were to develop their youth and make the playoffs at the same time. In the end both were accomplished. Vogel was an interim coach and was working under enormous pressure to deliver success for the team's sake and his own. The thing that worries me is how doggedly he stuck to the trial by fire model for his young players, even when it put the overall performance of the team in jeopardy. In many ways Vogel appeared as rigidly tied to his favored rotations as O'Brien had been.

The playoff series against the Bulls is a perfect example of what I'm talking about. The strength of the Pacers, particularly under Vogel, was their ability to create matchup problems with the second unit. Vogel played the Collison-George-Granger-Hansbrough-Hibbert starting lineup for 17.7% of the total minutes to finish out the regular season. During the playoffs, that lineup played 33.6% of the total minutes. They saw this increase in minutes despite a horrible drop in efficiency, posting a Net Rating of -16.17 against the Bulls. The other 9 units which played at least five minutes in the series had a combined Net Rating of +8.8. From my perspective there were some things that obviously were and weren't working in that series, and the numbers bear that out. Roy Hibbert really hurt the Pacers whenever he was on the floor, -15.94 Net Rating. The Pacers were much better, particularly on offense, when Mike Dunleavy was in the game, +12.22. However, those things didn't appear to make an impact on Vogel's rotations.

Perhaps Frank Vogel did what he needed to do to keep his job. I'm sure there is more to the puzzle than I'm privy to. In the end I don't want to find myself complaining too much about a playoff berth and an extremely competitive series against a much more talented opponent. Maybe Vogel will be able to loosen his rotations this year. Maybe roster changes made by the front office will reduce the Pacers' depth and make some of those rotation decisions more black and white. I just know it will be a load off my mind if Vogel starts the season by demonstrating a willingness to tweak his lineups based on what's working and not expecting the team to get better by just playing through.

I'm sure many of you have noticed it's been quite awhile since my last post here at IndyConrows. A cross-country move, starting a new job and a summer full of travel have kept my schedule overflowing. I'm happy to be able to say, nearly three months since I last wrote, that the schedule is starting to thin out and that this post should mark the return of my regular contributions here. It's good to be back!

Tweet Comment 4 comments | 0 recs |

Do you like this story?

Around SB Nation

Pacers 105, Magic 87: Darren Collison Fuels Fourth-Quarter Indiana Rally, Ends Orlando's Season

May 2012 from Orlando Pinstriped Post - 35 comments

Phoenix Suns Player Evaluation: Robin Lopez

May 2012 from Bright Side Of The Sun - 97 comments

USA Basketball: Olympic Team Adds Harden, Davis To Roster Selection

May 2012 from Barking Carnival - 23 comments

Kevin Durant's Top Ten Plays of the 2011-2012 Season

May 2012 from Welcome to Loud City - 0 comments

Glen Davis' First Orlando Magic Season Doesn't Go His Way

May 2012 from Orlando Pinstriped Post - 64 comments

Comments

Good to see you back, Ian

My take is this: I keep hearing about how the Collison-Dunleavy-Granger-McBob-Hibbert combination was the most effective on the team (I even remember hearing that it was supposedly one of the most effective in the league), but I’m always miffed as to why that starting lineup was part of a team that was no better than basically 10 games below .500. How do we explain that? Is it that the other lineup combinations behind them were so bad that they’re mostly responsible for the lack of Ws? Is it indicative that for all the seemingly wonderful things these advanced lineup/efficiency numbers provide, that they’re just simply not reliable when trying to equate them to Ws and Ls?

Someone help me out because I’m on the verge of dismissing them entirely, and I don’t want to do that in fear of disrespecting the work of some really talented analysts.

by Glenn A. on Sep 12, 2011 8:38 PM EDT reply actions

I missed these.

I have always liked McBob more than Hansbrough, maintaining the thought that Josh is a better starter and Tyler is a better bench player. So, I didn’t like the fact that Hans became the starter all that much.

We all saw how ineffective A.J. Price was on both ends of the floor and, frankly, he shouldn’t be on the team next year. Or, at least, not in the rotation.

Very excited for George this coming season, I believe he will be very much improved on the offensive side.

I was particularly excited about Vogel being hired. I like his energy, positive attitude, and the fact he doesn’t make an attempt to destroy his player’s confidence at any opportunity he gets. It’s still far too hard to determine how good of a coach he is, but for the time being I’ll support him and hope for his success.

"A champion is someone who gets up when he can't."
- Jack Dempsey

by infinityzero.systemerror on Sep 12, 2011 10:10 PM EDT reply actions

The Pacers

Were 19-13 when that lineup started. Under O’Brien, it was 11-10. Under Vogel, it was 8-3.

The Pacers were 4-4 when this unit started in December, 1-6 when it didn’t.

In the 13 games under O’Brien in January, the Pacer went 3-10. This unit only played 29 minutes during that stretch.

Ultimately, O’Brien went away from this line up when Hibbert collapsed in December. Unfortunately, he was stuck with Hibbert. In December, most of the losses came with Rush starting in place of Dunleavy. Come January, they were desperate for offense, so Hansbrough got the start over McRoberts.

The important thing to remember is that under Vogel, the Pacers’ biggest strength was the ability to play a lot of bad teams.

by count55 on Sep 12, 2011 10:12 PM EDT reply actions

Thanks for putting that together!

I didn’t have the time to put the numbers together in response to Glenn’s comment last night, but that was going to be my answer. When the Pacers were playing really well at the beginning of the season this lineup was starting, and playing lots of minutes. O’Brien went away from it when the team started to flounder.

There are really only two plays: Romeo and Juliet, and put the darn ball in the basket. ~Abe Lemons
Find me at IndyCornrows
and Hickory-High

by Ian Levy on Sep 13, 2011 5:26 AM EDT up reply actions