Ambiguous Architecture

Posted by Ian Levy on January 17, 2013 under Commentary | 6 Comments to Read

Solid Foundation?

For the first time in awhile, things are looking up in Dallas. Dirk Nowitzki is healthy, and the Mavericks are on a four game win streak. In their wins over Sacramento, Memphis, Minnesota and Houston, Dallas put up points at the scorching rate of 112.4 points per 100 possessions. This is a tremendous bump for what has been the 18th most efficient offense in the league this year and, at just 100.9 points per 100 possessions, the least efficient Mavericks’ offense of the past 13 seasons.

Offensive firepower of great variety has been the defining characteristic of Mavericks’ basketball for more than a decade, so watching the team struggle so mightily this season has been somewhat disconcerting. The absence of Dirk Nowitzki has certainly made things difficult, but the problems have been so systemic it’s hard to lay them all at the feet of one giant German. Across the entire season the Mavericks have wilted in each of the offensive Four Factors. They rank 8th in the league TO%, but 13th in eFG%, 16th in FTA Rate and 27th in ORB%.

The eFG% is especially troubling. Making shots is what Mavericks do, and under Rick Carlisle in particular, the team has shown a razor-sharp focus on the craft of creating quality open looks. This season however, their miraculous ability to manipulate and manufacture open space has largely fizzled. As dark as things have been, some fragrant Four-Factor-blossoms bloomed in their three most recent wins. They posted an eFG% of just 45.3% against Sacramento but pushed the bounds of offensive efficiency with just nine turnovers and 35 free throw attempts. Against Memphis and Minnesota, Dallas scorched the nets with eFG%s of 55.6% and 66.3% respectively. Against Houston, shooting was again a problem but 10 turnovers and 43 free throw attempts did the job. Those eFG% numbers are exciting to type; they feel like a thick, down sleeping bag with the potential to fend off the long winter weeks still to come. But I’m not sure they are truly a reflection of problems solved.

One of the reasons they Mavericks have struggled to score efficiently this season is that they’ve taken really difficult shots. At Hickory-High I’ve been working on a metric called Expected Points Per Shot (XPPS), which uses the average value of shots from different locations to measure the quality of a player or team‘s shot selection. The chart below shows where the Mavericks’ shot selection ranks this season. XPPS is graphed on the x-axis across the bottom, Actual Points Per Shot is graphed vertically on the y-axis.

*I use trips to the line for two free throws in my XPPS calculations. From here on out when I refer to shots attempts or shot selection, those shooting fouls are included.


As a team the Mavericks have an XPPS of 1.024 points per shot, well below the league average of 1.047. Only four teams, the Warriors, Magic, Sixers and Wizards, have had a less efficient shot-distribution this season. In terms of expected value, the shot locations rank in this order: Two Free Throws (1.511 PPS), Restricted Area (1.183), Corner Three-Pointer (1.157), Above The Break Three-Pointer (1.048), In The Paint Non-Restricted Area (0.793) and Mid-Range (0.788). The Mavericks’ XPPS works out so low because 42.2% of their shot attempts this season have come from those two least efficient areas. They also have a very inefficient ratio between Above The Break Three-Pointers and Corner Three-Pointers — more than 3 to 1 and one of the highest ratios in the league.

The difference between the Mavericks XPPS of 1.024 and the league average of 1.047 may not sound huge, but across an entire season it’s monumental. If we assume the Mavericks will attempt around 7539 shots this season (the league average from 2010-2011) that difference works out to 173 total points, or 2.1 per game. Again, I’m not talking about the Mavericks improving their accuracy to pick-up those 2.1 points per game. They could continue to makes shots at the exact same rates and pocket those extra points just by taking better shots. Not to put too fine a point on it, but I’m also not implying they need to do something extraordinary, merely sidle up alongside the league average.

This shot selection problem has certainly been exacerbated by players having to stretch both their usage and roles in Nowitzki’s absence. But the issues run deeper. Having missed out on Deron Williams and Dwight Howard, the Mavericks’ front office did an admirable job of maintaining flexibility while acquiring solid, professional basketball players. Unfortunately, the Mavericks have ended up with a roster with an incredible amount of duplication in both offensive abilities and tendencies. Simply look at the ten Mavericks’ who have attempted the most shots this season and what percentage of their attempts came from those two least efficient zones.

  • Elton Brand – 73%
  • Chris Kaman – 70%
  • Dirk Nowitzki – 61%
  • Brandan Wright – 57%
  • Shawn Marion – 48%
  • Dahntay Jones – 43%
  • O.J. Mayo – 42%
  • Darren Collison – 42%
  • Vince Carter – 35%
  • Jae Crowder – 25%
I’ve never looked at numbers for a team this way, but I still find them stunning. Nine of the Mavericks’ ten most frequent offensive contributors take more than a third of the shot attempts from the two least efficient areas of the floor. The thing that’s confusing is that none of those players are strikingly inefficient from those areas. Everyone except Jones, Crowder and a still recovering Nowitzki shoots better than 40% from those two areas combined. The thing is, those shots are inefficient on their face. Take the example of Chris Kaman, who is shooting a healthy 47% from those two areas combined. Not counting free throws, he has averaged 0.980 points per shot in those zones. That’s less than Corey Brewer has averaged on three-pointers this season. It’s also less than Sebastian Telfair, Earl Watson and Jodie Meeks have averaged on shots at the rim this season. Even when you’re very good in the mid-range game, it’s still not a frame to build an offense around.
The Mavericks have always had a close relationship with the mid-range jumper. Their two greatest offensive weapons of the past decade, Nowitzki and Jason Terry, relied on them heavily and were among the most accurate in the game. But in the past the Mavericks were adept at using it as a scalpel — in precise situations and often to open up other opportunities. The graph below is the same as the one we looked at above, except it only shows the Mavericks, going back to 2000-2001.


Both of these graphs are screenshots of my Team XPPS visualization. It’s important to know that Tableau automatically adjusts the average lines when it’s filtered. The average lines you see here on this graph are just for the 13 Mavericks’ seasons and are much lower than the league-wide numbers for the same span. Every one of these Mavericks’ seasons is actually below the league-wide average of 1.047, again the effects of long-term reliance on Terry and Nowitzki. It seems painful obvious here that Mavericks offense was at their championship-clinching best when their shot-selection was as well. For the XPPS range the Mavericks find themselves in now, the best-case scenario (in terms of efficiency) looks like the 2008-2010 offenses. Unfortunately the talent pool in Dallas is not quite as deep as it was and the worst-case scenario of last season may be more plausible. This is especially true for a team that is rarely found on the offensive glass or at the free throw line.

Circling back to the current win streak, in the two wins against Memphis and Minnesota, the Mavericks took 73 of their 157 (46.4%) of their shot attempts from those two least-efficient areas. They made an absolutely ludicrous 56.2% of those shots. In that small sample, the Mavericks have been scoring 1.121 points per shot on field goal attempts from zones where the league average is 0.791. They scored 82 points on those shots, where they should have expected to score about 58. That burst of offense was not because of a beautiful metamorphosis, but rather a likely unsustainable run of fortunate accuracy. Against Houston we saw reversion, with the Mavericks again taking 45.3% of those zones, but making just 35.9%. Expecting mid-range accuracy of the kind they displayed against Memphis and Minnesota to be the centerpiece of offensive improvement going forward is probably going to lead to frustration.

The solution really has two prongs. In Rick Carlisle’s hands is the potential to tweak the offensive playbook and the rotations to emphasize more efficient shot-distribution. On the player side is the opportunity to make their team better be simply passing up a few long jumpers for the sake of offensive re-initiation and the possibility of something greater. Shot-selection is a powerful tool because it doesn’t require talent or speed, strength or length, to be implemented. A better offense can be created with simple knowledge, awareness and better decision-making.

In addition to his work for The Two Man Game, Ian Levy is the author of Hickory High, and a contributor to Indy CornrowsHardwood Paroxysm, HoopChalk and ProBasketballDraft. You can follow Ian on Twitter at @HickoryHigh.

  • Breyzh

    It’s an interesting article, but those statistics cannot be properly analyzed without context. Your XPPS measurement doesn’t account for a team’s (or a player’s) accuracy at those types of shots, while one would expect teams who can shoot to have more chances of scoring on those opportunities. The Mavs aren’t the league-average team, and they lack the personnel who can create (or capitalize on) opportunities from 3 point land and/or the restricted area, while they are better than average at making mid-range jump-shots…
    It also fails to take into account specific situations, which have an even higher correlation with shot-making than shot-location does (a contested 3 pointer is much less efficient than an open long 2):
    So yes, the Mavs decision-making needs improvement, but statistics are empty if you fail to contextualise them.

    • Jay

      Agree with Breyzh, interesting read but this Mavs team has the personnel to shoot and make those shots efficiently. I would agree with Ian that it is not a great way to structure a team because there is little floor spacing and certainly overlaps in offensive abilities and preferences. But like noted above, the stat needs more context. League wide it might not be an efficient spot but how do the Mavs players you listed rank versus the rest of the league in those spots and how many of those shots are open versus guarded? Also while RC could somewhat alter the offense to get players in different spots, that doesn’t mean that they would in actuality be more successful from them. These players are shooting from the spots that they are comfortable with, can get to with regularity, and have had the most success with. Having Kaman out at the corner three could spread the floor but that doesn’t mean having him shoot that shot versus a mid range two is a good idea even if it’s wide open.
      I would look at these stats more as a basis for building a well-rounded team than a way of evaluating shot selection. The championship team like you said was based on 2 long jumpshooters; that is not the typical championship team structure but because Dirk and Terry were such good shooters and we had so many players that could hit threes on the floor at any given time, it was extremely effective. A more interesting analysis in my opinion would be breaking down successful teams by player using the statistics you outlined and looking for correlations in those terms.

  • Pingback: Court Vision: The art of setting a screen | The Point Forward -

  • FromWayDowntown

    Cool read but I did feel a constant unease about the XPPS. I lack the statistical prowess to really look through this but Breyzh might be onto something. Anyways I appreciate articles like this which try to dig deeper.

  • Pingback: Coming Up for Air | The Two Man Game

  • Pingback: Dallas Mavericks Player Combinations not working out | The Two Man Game