clock menu more-arrow no yes mobile

Filed under:

Jump shots, defense, and variance: What has happened to the Texas Longhorns over the last eight games?

Shaka Smart has to be hoping to catch a few breaks in March.

Denny Medley-USA TODAY Sports

By the end of the first week of February, the Texas Longhorns were rolling. Coming off of a ten point win against Texas Tech, Shaka Smart's team had won seven of its previous eight games, and was 7-3 in the Big 12. The following Monday, the Horns would lose a close game at Oklahoma where Texas played well, but didn't do quite enough to beat a top ten opponent on the road. It was not a discouraging loss.

But over the last eight games the Longhorns have failed to finish the season strongly. Texas is 4-4 since then, and while two of those wins came against West Virginia and Oklahoma, several of the losses were rather ugly.

While we as observers can find all sorts of things to attribute Texas' late January and early February success to, and other things to explain their relative slide since then, the numbers point pretty clearly to differences in what has happened on the defensive end of the floor.

Looking at the performance of the Texas defense since mid-January

The last 16 games of the season break up neatly into two halves. During the eight games from January 16 through February 8, Texas ran off a 6-2 record largely based on the apparent success of its defense. During these games, Longhorn opponents averaged 93 points per 100 possessions, and only once did an opponent (Kansas) exceed one point per possession in an individual game.

That one point per possession mark for the Texas defense is an important one, as on the season Shaka Smart's squad is 14-3 when holding opponents below it, and 6-9 when its opponents exceed it.

For all games played since Texas' February 8 loss at OU, Longhorn opponents have averaged 111 points per 100 possessions. Over this period where the Longhorns have a 4-4 record, six of eight Texas opponents have exceeded the one point per possession threshold.

It is the defense that best explains what has happened late in the year for Texas. In contrast to this, the Longhorn offense has been pretty much the same on average during these two time periods. Texas scored 103 points per 100 possessions during its 6-2 stretch, and 104 points per 100 possessions during its final eight games.

What is responsible for the defensive decline?

An 18 points per 100 possessions shift in performance over a period of games is substantial. But this big shift shouldn't come as a surprise to anyone who has followed this team; consider that Texas' four losses during the final eight games prior to today were by margins ranging from 10 to 30 points.

That 18 point difference is largely attributable to opponent shooting percentages. Texas opponents shot 53 percent from two-point range against the Longhorns during the final eight games, compared with 40 percent during the prior period. That difference accounts for 11 points per 100 possessions of the total 18 point change.

Longhorn opponents were also successful behind the arc 39 percent of the time in the final eight games of the season, compared with 32 percent shooting in the prior eight games. This difference is worth about 4.5 points per 100 possessions to the Texas D.

We can further subdivide opponent two-point shooting to get greater clarity around what was going well for the Longhorns during their peak stretch, and what hasn't been going as well lately. Using play-by-play data, I divided opponent two point shots into several groups. Two point shots can either come at the rim or on jump shots. And two point shots can either come in transition, or later in a possession in obvious half court settings. Based on these two characteristics, it is possible to group Texas' opponent shooting into four categories.

We can see how much of the change in opponent shooting is due to a change in each of these four categories. Looking at the data, the change comes from two categories, while the other two categories have essentially no effect. The two changes that matter are opponent two-point jump shots taken in half-court situations (which accounts for about two-thirds of the change in opponent two-point shooting percentage), and shots taken in transition at the rim (which accounts for the remaining one-third of the change).

During the period of games where Texas opponents were scoring 93 points per 100 possessions, they were making about 25 percent of their non-transition two-point jump shots. During the latter period where opponents were able to score 111 points per 100 possessions, their two-point jump shooting percentage in the half-court was 41 percent. This change matters a lot, since 50 percent of opponent two-point attempts taken against the Longhorns over these final 16 games have been these non-transition jumpers.

The change in transition shooting numbers is also significant. During the earlier period where Texas' opponents weren't scoring as much, 10 percent of opponent twos came at the rim in transition, and were converted 57 percent of the time. In the later period 13 percent of opponent twos came at the rim in transition, and the shooting percentage in these cases was 77 percent.

Jump shots are fickle

Between two point jumpers and threes, 11 to 12 points per 100 possessions of Texas' defensive lapse were caused by opponents making more of their jump shots during the final eight games, as compared with the previous eight. So what are we supposed to make of this?

In understanding how a defense can go from watching opponents convert a low percentage of jump shots to a significantly higher percentage of jump shots, we have to contemplate how much of this effect is due to a change in actual defensive performance, and how much is due to shear random chance. This is a difficult question that doesn't produce satisfying answers.

There is a decent amount of evidence that suggests on the whole that the defense doesn't have much influence over what percentages opponents shoot on three-point shot attempts. Similar work has not been done on two-point jump shooting percentages, but it doesn't take a huge leap of faith to suspect something similar would be at play.

But this is not the same as saying that the defense has no effect. Ken Pomeroy's work indicates that the defense does have some limited influence on opponent three-point shooting percentage, although the mechanism for this is hard to discern from the data he has available. And there are examples of teams that year after year hold their opponents to lower than normal three point shooting percentages -- Louisville and Syracuse come to mind specifically. Although, as Ken suggests, it can be hard to tell if this is because these teams are doing that much better of a job contesting threes, or if the overall effects of their defense forces opponents to settle for more low quality threes.

I am not going to be able to settle this question definitively, but I think it is safe to say that one should not have looked at an eight game stretch where Texas' opponents shot 25 percent from the floor on two point jump shots and then concluded that this percentage would hold for the rest of the season.

In hindsight, I am feeling somewhat foolish for not noticing that this was going on at the time, although even if I did I am not sure that I would have pointed it out to any of you, because in the moment I hate to be that guy who tells you that your team isn't all that good, and is just getting lucky. No one wants to invite that guy to parties.

But now after the fact I have no problem being that guy. I leave it up to you to decide if you want to invite me to your party -- although one thing that should sway you is that I make really good cocktails.

Random variation swings both ways. It was swinging in Texas' favor for a while, and now is swinging against the Longhorns. Opponents hitting 41 percent of their two-point jump shots and 39 percent of their threes is a higher than expected percentage, just as during the prior period opponents were shooting at a lower than expected percentage.

I cannot precisely quantify how much of the difference in Texas' defense over these time periods is simply due to random chance or good and bad fortune, and how much is due to an actual change in the performance of the Texas defense. What I can say is that the effect of random variation is real, and it is possibly a bigger portion of this change than anything that is within Texas' control.

Controlling what you can control

The slope from basketball analysis to pop psychology is one that I have no interest in sliding down. But I do think the school of thought of focusing on things within your control and letting other stuff happen is a pretty solid one.

There is a core issue in basketball that everyone has to confront. Once the shot is in the air, the ball either goes through the rim or it doesn't, and there isn't much that can be done about it either way. This is an even bigger deal for the defense, as it has less to say than the offense about the moments leading up to the point where the ball ultimately decides what happens. The side with the ball gets to make more of the decisions than the side without it, and when the other team is hitting shots at a very high rate you are more likely to lose.

Part of the problem for the Texas defense is the Texas offense, which is probably best described as OK, but not particularly strong. The Longhorns don't turn the ball over, but they also don't shoot very well from the perimeter, don't get many offensive rebounds, and are below average when it comes to finishing plays around the basket. In games where the other team is just hitting shots, you have to be able to score some to keep pace, and this is not a thing that Smart's team is able to do on a consistent basis.

Random chance and a single elimination tournament

Of course, it is March, and the NCAA tournament is here. One of the basic facts of a single elimination tournament is that randomness plays a considerable role in the outcome. Randomness does not decide the outcome, but it is has a say in the decision.

The best teams are the best because they are so good that even when the jump shot numbers break somewhat against them, they still do enough else well to stand a decent chance of winning -- although even the very best teams can go down to an opponent who hits an abnormally high number of jumpers.

Texas isn't good enough to be likely to survive this sort of bad break. If an opponent is hitting jump shots, and the Longhorns are not, then the Longhorns are probably going to lose.

But not all breaks are bad, and things are every bit as likely to break in Texas' favor in this game of jump shot roulette.