After a hectic offseason in the summer of 2013 that saw the introduction of three new starters, the new look Dallas Mavericks proved to be a pleasant surprise for MFFLs.1 Now stop me if you’ve heard this one before: they were an exciting team built around the shooting prowess of Dirk Nowitzki, with a top-three offense and a bottom-ten defense. In other words, they were pretty much the same Mavs we’ve seen, with a few exceptions, for more than a decade now.
The Mavs have already made it clear that they intend to upgrade defensively by trading Jose Calderon for Tyson Chandler. While there were some things the Mavs did well defensively last season,2 they were among the worst teams in the league at keeping opponents off the free throw line and holding down opponents’ FG%. In addition, they allowed their opponents to shoot from favorable locations on the court. Consider three areas a shot could come from — near the basket (<8 feet), three-point range, and midrange. League-wide, close shots yield 1.11 points per shot,3 three-pointers yield 1.08 points per shot, and midrange shots yield just 0.79 points per shot. This is why smart coaches are trying to find ways to force opposing teams to take more midrange shots.
The graph below shows the defensive rating of all 30 teams plotted against the percent of opponents’ shots that come from the midrange. The Mavericks were atrocious at forcing these defensively favorable midrange shots: only 29% of their opponents’ shots came from the midrange, the fifth worst mark in the league. By contrast, the Pacers’ opponents shot 38% of their shots from midrange! As you can see, forcing your opponent to shoot lots of midrange jumpers is closely associated with a strong defensive rating.
Even if a team is only average at actually defending the midrange jumper, forcing opponents to take it can significantly improve a team’s defense. The table below shows opponent shot location for the 2013-2014 Mavericks and Pacers. Assume that both the Pacers’ opponents and the Mavericks’ opponents hit each type of shot at the league average rate. If we take the expected points per shot for each shot type, multiply each by the percent of opponent shots that originate from that region, and add them all together,4 we get an expected point value per shot, shown in the last column of the table below.
|Team||% Opp. shots midrange||% Opp. shots close||% Opp. shots 3||Expected points per shot|
Purely by virtue of the locations they allowed opponents to shoot from, the Mavericks gave up, on average, an additional 0.027 points per shot (1.011 – 0.984). The average team will take about 76 shots in 100 possessions. That means that over the course of 100 possessions, the Mavericks will give up 76 * 0.027 = 2.052 more points than the Pacers simply because of opponent shot location. The Mavericks had a defensive rating that was 9.2 points higher than the Pacers’ defensive rating. Shot location alone therefore accounts for 2.052 / 9.2 = 22.3% of the difference in defense between the Mavericks and the Pacers! A back of the envelope estimation using pythagorean wins suggests that if the Maverick could have forced the shot locations that the Pacers did, the Mavs would have won 6 more games last season.5
How will acquiring Tyson Chandler help? I ran three simple regressions using all shots taken in the NBA last season to determine how Chandler’s defensive presence alters the location of opposing shots. The dependent variable for each regression is a dummy indicating if the shot was close or not, midrange or not, and a 3-pointer or not. I control for other defensive players, other offensive players, and both teams. The results show that Chandler increases the chance that an opponent’s shot will come from the midrange by 5.7% while reducing close shots by 2.5% and threes by 3.2%.
Chandler was on the floor for about 63% of all shots taken against the Knicks last season.6 To find out how his presence would alter the distribution of shots taken by Mavericks’ opponents, I take a weighted average of their current distribution (37% weight) and their current distribution + the likely contribution of Chandler (63% weight).7 The result is shown in the table below.
|% Opp. shots midrange||% Opp. shots close||% Opp. shots 3||Expected points per shot|
|Actual Mavericks Distribution||28.9%||42.8%||28.3%||1.011|
|Hypothetical Distribution w/Chandler||36.9%||37.8%||25.3%||0.986|
Chandler’s presence projects to lower an opponent’s expected points per shot by 0.011, or 0.836 per 100 possessions. This would give the Mavericks the 19th best defensive rating in the league, compared to their current ranking of 22nd.8
Last year’s Mavericks were an elite offensive team, and proved to be cagey defenders in a few playoff games against the Spurs where they were able to force turnovers and play strong defense in pivotal moments. Getting a better seed in the playoffs and getting out of the first round in 2014-15 will depend in large part on whether Chandler can anchor their otherwise anemic defense.
- That’s a “Mavs fan for life” for those of you who aren’t drinking the kool-aid. ↩
- They were third in the league at forcing opposing turnovers, for example. ↩
- 56% of shots made * 2 points per shot ↩
- So for the Mavericks, 0.289*0.79+0.428*1.11+0.283*1.08=1.009. The number shown in the chart, 1.011 is more accurate, from a spreadsheet I am using that does not round the location percentages at one decimal. ↩
- Good for the fourth seed in the vicious West. ↩
- When he was in the game. He missed 27 games. ↩
- Conveniently, my regressions show that Dalembert affects opposing shot location very little, so there is no real need to consider his lost contributions. ↩
- Remember I did all this holding the FG% of opposing teams constant. This is an improvement that comes only from how Chandler affects opposing shot locations. He will also improve the defense by reducing opponent FG%, which makes my estimate here quite conservative. ↩