Gestaltwandler

Posted by Ian Levy on February 25, 2011 under Commentary | 7 Comments to Read

Screen shot 2011-02-25 at 10.34.36 AM

Ian Levy is the author of Hickory High, a contributor to Indy Cornrows, and is now a part of The Two Man Game family. He’ll be bringing his intelligent brand of — mostly quantitative — analysis here on a weekly basis. You can follow Ian on Twitter at @HickoryHigh.

Last week, in looking at the possible impacts of Rodrigue Beaubois’ return, the impact of pace became a topic of interest. This season the Mavericks have played at an extremely slow pace. Going into the All-Star break they’ve averaged 90.5 possessions per game, 23rd in the league. Last season their pace factor was 92.5, still in the bottom half of the league but not nearly as slow as they’ve played this season. It hasn’t appeared to have a detrimental effect on their offensive efficiency; the Mavs are scoring 109.3 points per 100 possessions, the ninth best mark in the league.

One of the takeaways from last week’s Beaubois piece was how much the offensive pace changes for the Mavericks when he’s on the floor, particularly in three-guard lineups. His return may be only tangentially responsible, but it seems to be having the expected effect. By my calculations the Mavericks have averaged 98.2 possessions in the three contests since he returned.

This got me curious about the ability of a team to adjust their pace with different lineups. I started by looking at all the lineups the Mavericks have used this season which have played at least 30 minutes together. I then did a rough approximation of the pace factor for each lineup. These calculations are not exact because most lineups haven’t played the same number of offensive and defensive possessions. In order to estimate pace factor from the available data I took the average of each lineup’s possessions on both sides of the ball, and then looked at how many standard deviations the pace of each lineup was away from the Mavericks’ overall average pace for the season. The table below shows the results.

UnitMinutesNet RatingPaceNumber of SD from Mean
Barea - Terry - Marion - Nowitzki - Haywood305.37+11.8788.417-0.48753
Kidd - Stevenson - Butlet - Nowitzki - Chandler257.48+24.1891.440+0.22007
Kidd - Terry - Butler - Nowitzki - Chandler182.30+16.4090.444-0.01303
Kidd - Terry - Stevenson - Marion - Chandler148.12-4.3689.117-0.32376
Kidd - Terry - Marion - Nowitzki - Chandler118.20+11.7293.198+0.63157
Kidd - Terry - Butler - Marion - Chandler90.87+5.1891.912+0.33042
Barea - Terry - Marion - Nowitzki - Mahinmi76.63+21.9888.947-0.36357
Kidd - Stevenson - Pavlovic - Nowitzki - Chandler72.78-9.6687.057-0.80600
Kidd - Terry - Butler - Marion - Haywood53.92-30.2791.246+0.17470
Kidd - Terry - Stevenson - Marion - Haywood52.18-12.2691.989+0.34826
Kidd - Barea - Terry - Nowitzki - Chandler44.33+9.7894.744+0.99347
Kidd - Stevenson - Stojakovic - Nowitzki - Chandler43.82-29.5793.108+0.61055
Barea - Terry - Marion - Nowitzki - Chandler42.48-15.6693.785+0.76906
Kidd - Barea - Stevenson - Nowitzki - Chandler38.48+21.8886.694-0.89086
Kidd - Barea - Butler - Marion - Chandler38.43-1.7788.681-0.42588
Barea - Terry - Butler - Marion - Haywood38.07-4.0593.932+0.80345
Kidd - Terry - Marion - Cardinal - Chandler36.60+5.2891.148+0.15158
Kidd - Terry - Marion - Nowitzki - Haywood35.10+0.1587.521-0.69727
Kidd - Stevenson - Marion - Nowitzki - Chandler33.82+33.5487.995-0.58633
Kidd - Terry - Stevenson - Nowitzki - Chandler30.93+2.9890.010-0.11477

As you can see, the Mavericks have been remarkably consistent in their slow pace this season. It’s by a small margin, but none of these lineups have played at a pace one full standard deviation above or below the team’s average pace. Dallas hasn’t demonstrated much of an ability to use their personnel to change the tempo of a particular game or even a specific quarter. It’s entirely possible that this is by design, but we’ll talk about that in a minute.

The next step was to compare the Mavs to some of the other very efficient offenses in the NBA. I looked at Denver, Miami, Los Angeles, San Antonio, Houston, Oklahoma City, Phoenix, New York and Utah, which along with Dallas have the ten highest Offensive Ratings this season. I then did the same analysis for each team as for the Mavericks. There’s too much information to compile in one table, so if you want to see the raw data you’ll have to follow this link to the google spreadsheet.

I found that each of those other top-10 offensive teams had at least three lineups which played at one or more standard deviations above or below their team’s average pace. Every one of these teams (except the Mavericks) also had at least one lineup which has posted a positive net rating while playing at least one standard deviation above or below their team’s average pace.

The table below summarizes all the lineups for each team which have played a standard deviation above or below their team’s average pace and posted a positive net rating.

TeamUnitMinutesNet RatingPaceNumber of SDs from Team Pace
DENBillups - Afflalo - Smith - Martin - Hilario61.85+19.8786.531-1.98999
DENBillups - Smith - Anthony - Harrington - Hilario33.48+53.5291.039-1.01162
DENLawson - Afflalo - Smith - Anthony - Harrington30.48+24.03108.661+2.81337
LALBlake - Brown - Walton - Odom - Bynum54.75+6.9987.233-1.01407
LALBlake - Bryant - Artest - Odom - Gasol54.65+5.9283.001-2.04391
MIAChalmers - Wade - Jones - Bosh - Anthony85.33+9.0886.910-1.09813
MIAHouse - Jones - James - Haslem - Ilgauskas51.27+14.7595.026+1.02894
MIAHouse - Wade - James - Bosh - Anthony40.97+39.9984.940-1.61424
MIAChalmers - Miller - Jones - James - Ilgauskas36.65+1.5485.130-1.56461
MIAHouse - Miller - Jones - James - Anthony31.48+35.9896.823+1.49987
MIAArroyo - Wade - James - Bosh - Dampier30.78+13.6284.211-1.80546
SASParker - Ginobili - Jefferson - McDyess - Duncan98.98+10.9097.717+1.22517
SASParker - Hill - Jefferson - McDyess - Duncan65.03+2.5587.467-1.13667
SASParker - Hill - Ginobili - Jefferson - Duncan63.07+7.9697.416+1.15577
SASParker - Hill - Ginobili35.68+34.85100.224+1.80299
SASHill - Neal - Ginobili - Jefferson - McDyess33.15+26.8987.602-1.10568
SASHill - Neal - Jefferson - Bonner - McDyess32.78+15.5885.662-1.55269
OKCMaynor - Harden - Green - Ibaka - Collison121.07+5.4886.826-1.47614
OKCWestbrook - Harden - Thefolosha - Durant - Green48.93+10.42101.042+1.92277
HOULowry - Martin - Budinger - Scola - Hayes33.9+14.89101.239+1.35113
PHONash - Richardson - Hill - Warrick - Frye93.65+7.3698.922+1.36305
PHONash - Carter - Dudley - Frye - Gortat34.27+11.7089.641-1.43461
NYKFelton - Fields - Chandler - Gallinari - Mozgov35.43+7.0599.577+1.10583
NYKFelton - Chandler - Gallinari - Williams - Stoudemire30.32+5.4388.654-2.47117
UTHPrice - Watson - Miles - Elson - Fesenko35.62+34.1682.875-1.81042
UTHWatson - Bell - Miles - Millsap - Jefferson33.57+1.6186.506-1.00137
UTHWilliams - Bell - Miles - Millsap - Elson30.68+14.5297.001+1.33718
UTHPrice - Watson - Miles - Kirilenko - Jefferson30.57+16.3686.359-1.03404

Some of those lineups are merely a memory after recent trades.

Having the ability to play at different tempos is not a prerequisite for being a top offense, but it seems to be an attribute most of them share. I also think it’s interesting that San Antonio, Oklahoma City and Miami all have intact lineups with positive net ratings at one standard deviation above AND below their average pace. These teams truly have the offensive versatility to be successful at any tempo.

You can look at the Mavericks’ situation in a couple of ways. Over the years they have acquired a roster of complimentary players whose offensive skills fit very well in a particular system. There is certainly no shame in sticking to that system, especially when it’s been successful enough to make them a top ten offensive team. You could make the argument that Dallas recognizes that they operate best playing the game at a certain speed and they make a point of maintaining that speed on a nightly basis.

The problem is that at some point that system is going to be stretched, forcing them to play the game at a different pace. I certainly don’t advocate Dallas changing their offensive style willy-nilly and wildly switching things up game to game. The pace they play at is a calculated decision that balances and feeds into what the team hopes to accomplish at both ends of the floor. However, having some additional flexibility when the playoffs roll around will allow Rick Carlisle to counter certain matchup advantages of their opponents as well use his personnel to exploit possible advantages in the other direction.

I do think it will be to their benefit to spend the next month and a half developing some lineups that are comfortable, confident and successful in an up-tempo game. Beaubois and some three guard arrangements seem like the most likely candidates. A path to the Western Conference Finals is likely going to involve matchups against at least two if not three of the teams we looked at here. These other teams are obviously comfortable playing at different speeds. Bending your opponent’s style of play to fit your own is an admirable goal but it can’t hurt for the Mavericks to start developing some contingency plans, just in case.

  • http://pulse.yahoo.com/_34QO2NBPONHDTZGJZ45TESCJCM Ian

    My apologies for the typo. The Mavericks do not have someone named “Butlet” playing small forward in their second most used lineup of the season. Normally I would just fix it but that one game me a laugh.

  • Pingback: The Point Forward » Posts Court Vision: The latest around the league «

  • http://www.shattertheglass.com Bgalella

    I hate that Butlet is out for the year.

  • finzent

    Very interesting stuff. Had to look up what a standard deviation is, though.

    • http://pulse.yahoo.com/_34QO2NBPONHDTZGJZ45TESCJCM Ian

      Sorry. A little more explanation next time!

  • Aditya Challa

    Would it be possible to test who is dictating the pace? i.e. when a slow 'Pace' lineup faces a fast 'Pace' lineup?

    That could be interesting. We might find that the mavericks tend to dictate the pace against opposition.

    • http://pulse.yahoo.com/_34QO2NBPONHDTZGJZ45TESCJCM Ian

      The only way to do that in a reasonably feasible manner would be to look at a very small sample size, for example a single game. It's an interesting idea though. It might be really beneficial in analyzing a playoff series.