Compared with a year ago, the Texas Longhorns basketball team is significantly improved. At this point last season, Texas was 4-3 after being run off the floor by Northwestern, beaten down by Colorado, and suffering a double-digit home loss to UT-Arlington. This season Texas is 6-2, with its only two losses coming in overtime against ranked opponents — No. 1 Duke and No. 17 Gonzaga.

That the Longhorns are playing differently from a year ago is not surprising given that Shaka Smart's team features a rather different lineup, with four of the top seven players making their first appearances for Texas this season. But it isn't just the new players who are helping; the three returning players — Kerwin Roach II, Andrew Jones, and Eric Davis Jr. — have all started off this season better than the last one. Their improvement shows up in a number of different ways, but today we will focus on only one. We will focus on the offensive contributions of Kerwin Roach.

Texas' two-headed point guard

A season ago, Kerwin Roach and Andrew Jones shared primary ball handling responsibilities. It did not go well; the Texas offense was a mess of turnovers and general dysfunction. This year Texas added a true freshman point guard in Matt Coleman, who has done well to start the season. He is valuing the basketball and helping to run the team.

With Coleman's arrival, there has been a lot of talk about how much better Roach is doing after moving "off the ball." I am going to contest that point of view to a degree, because Roach hasn't entirely moved off the ball. He is handling the ball quite a lot.

Texas is still sharing ball-handling responsibility, with Roach lightening Coleman's load significantly — Coleman is the designated point guard while Roach remains Texas' secret point guard. He isn't being called a point guard, but he is doing a lot of point guard stuff.

The difference is this season he is doing it a lot better than last season.

In his role as secret point guard, Roach is doing much of the actual play-making for Texas in half court situations. Using data from hoop-math.com, we can see that 14 of Matt Coleman's 32 assists have occurred in likely transition situations, meaning that they have occurred within the first 10 seconds after a live ball change of possession. 18 of Coleman's assists came outside of these early offense cases. Meanwhile, 18 of Kerwin Roach's 25 assists also did not come in early offense.

This (along with a similar split in ball screen rates that I will describe below) suggests what careful viewers have probably noticed — Roach and Coleman are engaged in an approximately 50/50 split of half court play making. You could almost call this arrangement a "point guard by committee" if you were specifically interested in rankling those who rigidly cling to arbitrary positional definitions. Let's be clear — I am specifically interested in rankling these people — the truth is on the damn floor. There once was a time when guards were simply guards; the boundary between the guard positions was established during the late 70s and early 80s. Meanwhile, the basketball does not care what you call the person who dribbles it.

Much like last season, Texas remains a "point guard by committee" team; the simple difference is that this year the committee is better.

To date, the Coleman/Roach partnership has worked pretty well. The Longhorns have an exceptionally low turnover rate, giving the ball up in fewer than 16 percent of possessions, which ranks around 25th nationally. Roach is helping Coleman, a freshman who has played well, but is still adapting to the college game. And Coleman helps Roach, by helping Texas more quickly advance the ball up the court and get into its offense.

Jones gives the group a third ball handler, which is helpful in transition and when dealing with defensive pressure, but the reorientation of things has allowed Jones to spend more time doing the things that he does best — finishing in transition and playing as a spot up perimeter shooter.

Texas' ball screen game

There was a specific maddening element to Roach's play last season. It was how frequently Roach almost made a play on offense that would lead to a score for Texas, only to have things fall apart for him right at the critical moment. This season that is happening far less often, and as a result, Roach has emerged along with Jones as one of Texas' two most effective offensive players.

Within the context of the Texas offense, one of the key responsibilities of the guards is to use ball screens to break down the defense and create chances to score. According to the Synergy Sports video database, Roach and Coleman have a roughly equal split of logged ball screens, with 55 logged for Roach and 50 logged for Coleman. (Note that the Synergy database only logs plays that end in a shot or a turnover. As a result many ball screens end up not being logged). Both players have performed fairly well in the ball screen game, with Texas scoring 58 points on Roach's 55 logged possessions and 47 points on Coleman's 50 logged possessions. This is a substantial improvement over the performance Roach delivered last year, when his 277 logged ball screens resulted in only 225 points.

So what is making the ball screen game more effective for Roach this season? The answer turns out to be pretty simple. Just like last year, half of Roach's logged ball screened possessions end with him passing, and half end with him keeping the ball for himself. When he passes, Texas averages roughly one point per play — the same result as last season. The difference really comes when he keeps the ball. Last year when Roach attempted to score for himself off the ball screen the Longhorns averaged about 0.6 points per play. This season when Roach keeps it, he has scored 29 points in 28 chances. That is a lot better, although to be clear, this is still a relatively small sample of data (only about 1/5 of the total ball screens he tried to score on last season).

All of this improvement in Roach's ball screen activity this season has come from the fact that he is shooting for a much higher percentage on these sorts of plays. A year ago Roach shot 28-83 coming off of ball screens; this year he has converted 14-19 of his shots.

This is a pretty substantial difference.

It’s a little harder to understand just why Roach is converting on a much higher rate of his ball screen scoring opportunities this season, because it forces us to synthesize a bunch of different pieces of information, and because we have a relatively limited sample of data so far this year. But at least we can make some tentative observations.

How Roach is making the ball screen work

One thing that surfaces from the more finely-sliced cuts of Roach's Synergy ball screen data is a difference in the frequency with which he pulls up for a jump shot after using a ball screen. Roach is far from the best shooter in the game, but even for much better shooters than Roach, a pull up jumper off of a ball screen is frequently a low value shot. A season ago, about a quarter of Roach's ball screen drives ended with a pull up jumper, while so far this season he has only taken two such shots.

Instead, Roach is far more focused on attacking the basket (where he is converting at an incredibly high rate this year) or finding one of his teammates. I thought it might help to go back and look at two examples from the VCU game where Roach made the ball screen work for the Texas offense. This was a game where Roach was solid, going 3-3 from the floor, 1-4 from the line, and registering seven assists to only three turnovers. Roach made a number of plays for the Longhorns in that game, but these two were selected because they seemed particularly interesting, and reveal the things that make Roach so difficult to guard.

The first example is selected from early in the second half, and shows a case where Roach decisively attacks what is probably a defensive mistake. To set up the action, the photo below shows Roach preparing to take a high ball screen by Mohamed Bamba.

Bamba prepares to set an angled screen to allow Roach to dribble to his right. The defense responds by using a specific tactic known by a number of names — I will use the term "ice" because it is my favorite one — where the defender guarding the ball forces the ball handler away from the screen. We see this in the photo below.

Against such a defensive tactic, the best move for the ball handler is usually to refuse the screen and attack off the dribble. The man guarding the screener usually is responsible for attempting to contain this dribbler. However, in the photo below we see that after advancing a few frames that is not happening. Rather than helping guard the penetration, Bamba's defender instead decides to hang out with Bamba 20 feet from the basket, and out of the play. While Bamba is personable and would probably be fun to hang out with, this isn't the time for chit-chat.

Thankfully for the Rams, there are secondary defenders who can potentially help out. Dylan Osetkowski's defender helps off to try to slow Roach down.

Slowing Roach down, who by now has a full head of steam, is easier said than done. One of the harder things to do is to stop an attacking defender in an open area, and this surely qualifies. Roach refuses to be stopped, crossover dribbles to his right hand, and blows by the defender with an eye on the rim.

VCU still has one final defender, but he is forced to make an awful choice. He can either try to stop Roach, leaving a 40 percent three-point shooter and NBA Jam-hot Jones wide open in the corner, or he can stay with Jones and give up the layup. He perhaps should have helped, but he did not. Roach scored on an uncontested layup that was created by a combination of his read of the defense, a blown assignment, his speed with the ball, and the spacing of his teammates on the floor.

Isn't it nice how interconnected everything is?

Let's take a look at a second example, drawn from late in the first half, where Roach again used his speed to blow away the defense. And this time a missed defensive responsibility wasn't involved in the play.

We pick up the action with the photo below, where Bamba is again setting a high ball screen for Roach. The angle of this screen is of particular importance. Because of Roach's limitations as a shooter, opposing teams may prefer to try to send the defender under ball screens so as to try to cut off penetration. But a proper screening angle where the screener's shoulders are at an angle to the sideline, as opposed to parallel to the sideline, makes it difficult to go under the screen without losing the ball handler entirely.

A few moments later and we find Roach using the screen, as shown in the photo below. There are several specific things developing in the image that we need to discuss. First, the defense is attempting to use Bamba's defender to slow Roach down in a maneuver that is sometimes referred to as either "showing" or "hedging." "Showing" and "hedging" typically refer to things that are subtly different, but for our purposes here that difference doesn't matter. In both cases this second defender's job is to slow the ball handler down long enough to allow the primary defender to get over the screen and recover. At least for the moment, the defense appears to be in decent position.

A second thing we need to account for is the particular offensive concept that Texas is using. Bamba is going to screen and roll, while the other big man on the floor (in this case Royce Hamm) will pop out and replace as the play proceeds. For reasons that are pretty self explanatory, this is commonly called a "roll and replace" and has become one of the more popular concepts in modern basketball. When the man replacing the roller can shoot, it is particularly dangerous.

A the play progresses, we advance to the image below, where the defense is no longer in good shape. Roach has aggressively attacked the secondary defender and is in the process of blowing by him to make his turn to the basket. This is one of the things that Roach is really good at. He is hard to stay in front of, even when the secondary defender is a senior and a high-level athlete, as is the case here.

That is all well and good. Roach was beating guys off the dribble plenty often last season; it is what happened next that usually got him into trouble. So let's see how things turned out in this specific example. As the play progresses, we find that Roach has penetrated into the defense, as shown below.

Roach now has a number of options available to him.

1. He can pull up for a jump shot.
2. He can continue his attack at the basket and take a contested shot at the rim, attempt a runner, or potentially draw a foul. Another possible outcome here would be an offensive foul.
3. He can come to a jump stop, pivot, and throw the ball back to Hamm, who is wide open with his defender helping in the lane. (This is the shot that Connor Lammert took a lot, and that someday will be there for Kamaka Hepa. This is the power of the roll and replace.)
4. He can find Bamba rolling to the basket.
5. He can throw a cross court reversing pass (probably some sort of jump pass) to Coleman on the far side of the court. Coleman would then have a chance to either shoot or attack off the dribble against a closing defender.

Given the circumstances Roach made what is probably the best decision. He lobbed the ball to the player on the court who can out jump everyone else. Bamba ended up drawing a foul on the play.

The Texas offense still has some work to do

The Longhorn offense is far from perfect, but it is vastly improved from last year. Texas is getting better shots, taking care of the ball, and getting to many more offensive rebounds than we saw a season ago. Things would be really good if a few more guys would start to knock down a few more shots.

The development of Roach over the last 12 months has a lot to do with Texas' overall improvement. He is completing plays he just didn't complete all that often last year, and has become a catalyst for a lot of the good things that are happening for the Longhorns.