Recently, we tackled Shaka Smart's offense, trying to identify the sorts of results that his VCU teams have frequently achieved. While Smart's offenses have been up and down, on defense Smart's teams have been more consistent, and generally better. VCU's defenses have been outstanding, ranking in the top 25 in Ken Pomeroy's adjusted defensive rating in three of the last four seasons.
While there were quite a few similarities between the VCU offense and Texas' offense under Rick Barnes (both teams derived significant advantages from protecting the ball and crashing the offensive glass, and both frequently relied on the perimeter players to create scoring opportunities), there are essentially no similarities between the two defenses, outside of the fact that both have been rather effective. Smarts approach to defense, which focuses almost exclusively on forcing turnovers, could not be more different from Rick Barnes' D, which sought to limit opponents to one low percentage shot per possession.
The numerical differences are stunning. This season the Texas Longhorns forced turnovers in 14 percent of opponent possessions, which ranked 350th out of the 351 teams playing D-I basketball. In three of the six seasons that Smart coached VCU, the Rams had the highest opponent turnover rate in the nation. On the other hand, Texas opponents had the second-lowest effective field goal percentage in the nation, at 42 percent, while Smart's teams have at best barely cracked the top 100 in this measure.
The trade-offs of a pressure defense
One of the basic facts of basketball defense is that there are inherent trade-offs that come with different styles. Even within the frame work of man-to-man defense, there are many different approaches that a coach can take, with each approach creating different strengths and weaknesses.
Like Rick Barnes, Shaka Smart is a man-to-man defense coach. The first decision that a man-to-man coach must make is what to do with his off-ball perimeter defenders. There are a variety of options here, but at their core the coach must decide if he wants some or all of these defenders denying passes, or playing in gaps to defend dribble penetration.
Shaka Smart's teams at VCU have generally played in the passing lanes, as shown in the photo below. Notice where the wing defenders are in this photo, playing up on their men to attempt to deny the first pass. This approach to half-court defense, when combined with Smart's aggressive full-court press, leads to a lot of chances to deflect passes and create turnovers. In fact, creating deflections is one of the primary goals of Smart's defense; Texas fans are going to hear the word "deflection" a lot when Smart talks.
An interesting contrast to that is shown in the photo below, which is from Texas' defense last season. In this photo, we see that Texas' off-ball wing defender (Javan Felix) is not playing up close to his man, and is not contesting the pass. Instead, he is standing in the gap, trying to prevent dribble penetration.
In the photo's above, we clearly see that the VCU approach and the Texas approach to off-ball defenders is different. This difference has real consequences. VCU's more aggressive approach created more turnovers, but also allowed more dribble penetration and surrendered more layups.
Over the past few seasons, VCU has given up quite a few layups, and not all of these were layups that came in transition against the full-court press. Even in obvious half-court situations, Smart's teams have typically allowed something like 35 to 40 percent of opponent shot attempts to occur at the rim. Contrast that with Texas' more conservative defense, that allowed only 26 percent of opponent shots to occur near the basket against a set half-court defense.
So here we have a basic trade-off. Smart's defense is likely to both create more turnovers, and to allow more shot attempts near the basket than the defense used by his predecessor. This leaves us with one obvious question -- a question for which mathematics potentially provides an answer.
Is Texas' defense likely to be better or worse adopting Smart's pressure philosophies?
To set the table for answering this question, let's first list the characteristics of some of Rick Barnes' best defenses at Texas:
- Exceptionally low opponent eFG%. Season averages ranging between 42 and 47 percent were common.
- Below average opponent turnover rates, which occasionally were exceptionally low. This season, Texas' opponent turnover percentage was 14 percent, the second-lowest rate in the country. Last season, it was only somewhat better, at 16 percent. The upper typical range for a Rick Barnes defense was about 20 percent.
- Good to outstanding defensive rebounding, with opponent offensive rebounding percentages as low as 28 percent.
- Average or slightly better than average opponent eFG%. Something ranging from 47 to 50 percent is typical.
- One of the highest opponent turnover rates in the nation, frequently ranging between 24 and 29 percent.
- Average or below average defensive rebounding. Teams that allow a lot of drives to the basket also tend to give up some offensive rebounds. Generally, something between 31 to 35 percent is what the result ended up being most seasons.
We can do a simple estimate to determine which approach is likely to be better. It turns out that in general a one percentage point change in eFG% is worth about 1.3 points per 100 possessions, which is the same as a one percentage point change in turnover rate. This is also about what a two percentage point change in offensive rebounding rate is worth.
First, let's assume(*) just for the sake of argument that we swap out typical VCU values under Smart for the typical Texas values under Barnes. In a few paragraphs we will dig into this assumption further, and can explore some of the cases where it is not correct.
(*When you make assumptions, you potentially make an ass out of you and me. When you don't make assumptions, you do nothing to move forward our understanding of anything, and might as well sit in the dark and wait for death.)
Effective field goal percentage. With Rick Barnes' defense, we expected this to range between 42 and 47 percent in the best seasons. With Smart, under our initial assumptions we might expect it to range from 47 to 50 percent. This difference will hurt the Texas defense by between 0 and 10 points per 100 possessions. It is highly unlikely that there will be improvement in effective field goal percentage defense in changing from Barnes' defense to Smart's.
Turnover percentage. With Rick Barnes' defense, turnover percentages ranged from 14 to 20 percent. With Smart, we might expect this value to vary from 22 to 29 percent. This difference is likely to help the Texas defense by between 3 and 20 points per 100 possessions. Texas' opponent turnover percentage will almost certainly improve right away, and the improvement from the ridiculously low 14 percent of last season to something like 22 or 24 percent next year will be substantial, and will be worth around 10 or 15 points per 100 possessions.
Rebounding percentage. Pressure defenses typically give up more offensive rebounds than defenses designed to limit dribble penetration. With Rick Barnes' defenses, opponent offensive rebounding rates typically varied from 28 to 34 percent, but were more likely to be at the lower end of this range. VCU's rebounding percentages varied from 31 to 37 percent. This difference is most likely to hurt Texas' defense, but should not hurt by any more than about 5 points per 100 possessions. Texas' defensive rebounding percentage is likely to be worse next season than it was this season, but this difference may be so small that it won't matter all that much.
So if we add all of this up, we may reasonably expect that the range of potential outcomes for the Texas defense will be anything from getting worse by about 12 points per 100 possessions to improving by 20 points per 100 possessions. That is an exceptionally broad range of possible outcomes, but honestly I don't think the near-term future is quite as uncertain as this.
Keeping with our basic assumptions, if I were to just guess based on this analysis, I would expect that the Texas effective field goal percentage defense is likely to get worse by about 5-7 points per 100 possessions, Texas' rebounding will cost the team 2-4 points per 100 possessions, and the extra turnovers will help the defense by about 10 or 15 points per 100 possessions, relative to what we saw this last season when both effective field goal percentage and turnover rates were exceptionally low. Adding it all up, I would expect that the Texas defense might be between 1 point per 100 possessions worse and 8 points per 100 possessions better in its first season under Smart. In other words, I expect that the defense most likely will improve.
I want to spend a few minutes considering the assumption that I made in simply applying VCU statistics to Texas without any sort of adjustment. This is a completely uncontrolled assumption, and in fact is not much better than a guess. So let's consider it a little further.
Just to make things more concrete, let's take the example of effective field goal percentage. Effective field goal percentage depends on a number of factors, some of which are largely due to defensive structure dictated by the coach, some of which are based on opponent strength, some of which are dependent on the abilities of the defenders, and some of which are almost totally random.
Smart's pressure defense is just going to allow opponents more shots near the basket than Barnes' tightly compressed defense did. But beyond that, there are other factors that have nothing to do with defensive structure that are going to have a big effect on eFG%.
Opponent strength is a big one. VCU has played in the Atlantic 10 for the last three seasons, which is a very good league. The Big 12 is a better league. With better opponents, we might expect that Smart's opponents are going to have the ability to convert a few more shots against him at Texas than they did against VCU.
A factor that competes with this one is that Smart will now enjoy a roster with some bigger interior defenders. Smart did have a decent rim protector in Mo Alie-Cox at VCU, but Prince Ibeh and Cameron Ridley are better. Texas under Smart is likely to block more shots than VCU under Smart did.
The effects of better (and bigger) opponents and better shot blockers are likely to at least partly cancel each other out. This gives me some reason to believe that my somewhat optimistic projections for the defense can hold.
If there is a worry, it is in the turnover rate. It will likely take a couple of seasons for Smart to get the turnover rate up to full speed with Texas.
The hope for Texas fans is that Smart, with access to the types of recruits that he can pursue at Texas, will over the long term be able to build a defense that can both force a high rate of turnovers, and can defend the rim at a high level. This is the winning combination that Rick Pitino struck with his national championship team at Louisville.
The ability to achieve this combination will ultimately set the ceiling for Smart's program at Texas. If Smart can strike this balance, and can pair this sort of defense with the perimeter shooters that he needs to thrive, then the sky will be the limit for Texas.
If he can't, Texas will have good teams, but probably won't do much better than that.