clock menu more-arrow no yes mobile

Filed under:

Texas Longhorns Basketball: Inside the Numbers, Week 10

If you buy something from an SB Nation link, Vox Media may earn a commission. See our ethics statement.

Texas earned its second conference win this week, as well as suffering its second conference loss. Texas has now entered its toughest stretch of the season. I don't need to tell you about it, as this has already been detailed by Peter, as well as well as our friend JC over at Barking Carnival.

Conference wins are always tough to come by, and should never be taken for granted. For an example of this, just look at the Big 10. Ohio State is one of the favorites this year to win the NCAA tournament. And yet they already have lost to Indiana and Illinois. Indiana has losses against Michigan State and Minnesota. Michigan State was beat by Northwestern. Minnesota has lost a handful of conference games. This stuff is hard to figure out, so my advice is to not dwell too much on any one game when trying to understand which team is best. A similar scenario could easily occur in the Big XII, where there are many high quality teams.

This week in Inside the Numbers, I review the Texas A&M and Missouri games, use play-by-play data to examine Julien Lewis' offensive game, talk about the Texas offense, and explain my favorite rating system for college basketball.

The Week In Review

Background information on the statistics is posted here and here.














FGA + 0.475 x FTA




Off Rebs

























The Aggies aren't the greatest basketball team this year, but they do a pretty good job of forcing teams to play their type of game. They slow down the tempo (this game had only 61 possessions) and they play defense. Thankfully, the Longhorns are well suited to play this style of game. The Texas offense played much better than it looked. After this game much of the scuttlebutt on BON was how bad the Texas offense was. But let's take a step back and consider a two things:

1) The Aggies play tough defense. In a later section, I will talk a bit more about how Texas has been doing lately on offense when we consider the quality of the opponent's defense.

2) Texas did well at a couple of "hidden" factors on offense. They didn't turn the ball over very much, only giving it away in 19% of their possessions. As I have written before, no one notices when you don't turn over the ball, but it is one of the most important things that an offense can do. Additionally, Texas got to the free throw line 31 times. Their FTA/FGA ratio was 0.76, which is a very high rate. When you get to the line that much, your offense may not look very pretty, but it can be pretty effective.

This game was decided based on the true shooting percentage difference, as both teams took the same number of shots. The Texas true shooting percentage of 0.547 against A&M is a pretty solid number. A number like that is good enough to win most games, provided that you defend, take care of the ball, and rebound. Using the points above median (PAM) measure of scoring efficiency, two Longhorns stood out. Julien Lewis had a PAM of 5.9, which is a fantastic number. Alexis Wangmene chipped in with a PAM of 2.2.

As good as the Aggie defense was at times, the Longhorn defense was better. Texas held A&M to a true shooting percentage of 0.459, and no A&M player had a PAM greater than 2. The Aggies just couldn't get to the basket against Texas, with only 17% of their field goal attempts occurring at the rim. On the season, 1/3 of their attempts come at the rim, so this is a pretty significant drop off.

Texas didn't have the greatest game on the defensive glass, only rebounding 61% of the available defensive rebounds, but that was probably the only major problem for the Longhorns.














FGA + 0.475 x FTA




Off Rebs

























Missouri is pretty good. Here is what I wrote about them in my Big XII preview:

Missouri has the #2 rated offense in the country in the rankings. They play at a fast pace (#39th highest adjusted tempo on but also protect the basketball. They turn the ball over in 14.4% of their possessions (#2 in Division I). And they lead Division I in effective field goal percentage, at 58.8%. They get 42% of their field goal attempts at the rim, and convert on 72% of these attempts. This results in an average 57% field goal percentage on two point attempts. To add to this, they take around 36% of their field goal attempts from three point range and hit them 40% of the time. Missouri has a terrifying offense.

I still stand by that. Missouri turned the ball over in 14% of their possessions against Texas, and managed a true shooting percentage of 0.657. It is really hard to beat a team that manages to do those two things. In the Inside the Numbers world, basketball is a really simple game. If you are more efficient with your shots than your opponent, and get more shots than your opponent, you will always win. It is just a basic mathematical fact. When you manage a 0.657 true shooting percentage and a 14% turnover rate, you are well on your way to accomplishing these two goals.

On top of that, Missouri forces turnovers on defense. In this game, Texas turned the ball over in 22% of their possessions. This is a little bit high for Texas, but not a completely disastrous rate. Texas managed to not give Missouri an advantage in the total number of shots by doing good work on the glass on both ends of the floor.

Like the A&M game, this one came down to the difference in true shooting percentage. The Texas offense was pretty good, but the Missouri offense was great. For Texas, J'Covan Brown had a PAM of 15, which is pretty easily the highest total for any Longhorn this year. Jonathan Holmes had a PAM of 4.2. Unfortunately, Julien Lewis' shooting difficulties were really hard on the Texas offense, and Lewis ended up with a PAM of -7.6. Missouri true shooting percentage superstar Ricardo Ratliffe had a PAM of 8.

Looking at Julien Lewis' offensive game

Julien Lewis had a strange week. He had a very good game against Texas A&M, and then an awful one against Missouri. Against Texas A&M, Lewis scored 16 points while only taking 10 field goal attempts and 1 free throw attempt. Against Missouri, Lewis again took 10 shots from the field, while only scoring 2 points. Let's take a look at where his shots came from, and the situations in which they occurred. In going through this exercise, we will get a pretty good understanding of Julien Lewis' game.

Lewis is primarily a jump shooter. On the season Lewis only has taken 22% of his shots at the rim, and seldom shoots free throws. Getting to the rim just isn't his game. As a jump shooter, his scoring will be expected to be somewhat volatile. What is troubling is that even averaging out over the season, Lewis' true shooting percentage is only 0.431. Much of the problem comes from his tendency to take what I consider to be bad shots.

The table below shows the breakdown of his shots taken from the play-by-play data for the two games. Against A&M, Lewis took and made one shot at the rim, while he attempted no shots from the rim against Missouri. In both games Lewis also took 3 attempts from beyond the three point line. He made all three against A&M, and all of these shots were assisted. In other words, against A&M he hit 3 shots from three point range in catch-and-shoot situations. These are good shots, and they are the kind of shots we want Lewis to take. He missed all three of his three point shots against Missouri, but that is going to happen from time to time. Just because a shot was missed doesn't make it bad. When Lewis shoots a three that misses, it is still usually a good shot.

Opponent vs. A&M vs. Missouri
FGA 10 10
TS% 0.764 0.100
FGA at Rim 1 0
FG% at Rim 100% ---
%Assisted at Rim 0% ---
FGA 2pt Jumpers 6 7
FG% 2pt Jumpers 33% 14%
%Assisted 2pt Jumpers 50% 0%
FGA from 3pt 3 3
FG% 3pt 100% 0%
%assisted 3pt 100% ---
FTA 1 0
FT% 100% ---

That leaves us with Lewis' two point jump shot attempts. Against A&M, Lewis went 2 for 6 on these shots, with one of them being assisted. Against Missouri, Lewis went 1 for 7 on two point jump shots, and his one made shot was unassisted. On the season, Lewis has not shot well overall on two point jump shots, hitting only 24% of them. It is also interesting that Lewis' made two point jump shots are rarely assisted; his assisted rate of 17% on two point jump shots is the second lowest total on the team. This suggests that a high proportion of Lewis' two point jump shots are not catch-and-shoot situations.

I think when evaluating a player like Lewis, it is important to differentiate between good shots that miss, and bad shots that miss. Based on watching him play, and on studying the numbers, most of Lewis' two point attempts are not particularly good shots. When he puts the ball on the deck and creates his own shot, it is usually a poor shot with a low chance of going in. When he has his feet set, and he is in a catch-and-shoot situation, then things are better.

Lewis missed his three looks from the three point line against Missouri. No big deal, this will happen from time to time. The problem was that many of the rest of Lewis' shots against Missouri were not good shots. He took some bad shots against A&M as well, but he was able to make enough of the good ones to cover the bad ones up.

The Texas offense against A&M was better than it looked

I want to take a minute to compare the Texas offense in the A&M game and the Missouri game. If you are like me, after watching the A&M game you probably thought the Texas offense looked terrible. And you also might have though that the Texas offense looked much better against Missouri. What do the numbers say about this?

The Texas offense did do better substantially better against Missouri than it did against A&M. The 115 points per 100 possessions against Missouri is significantly better than the 99 points per 100 possessions Texas scored against A&M. But the shooting efficiency difference in these two games wasn't that much different, and I suspect that what you felt was most ugly in the A&M game was the difficulty Texas had getting and making good shots. Against A&M, Texas managed a true shooting percentage of 0.547, whereas they had a true shooting percentage of 0.556 against Missouri. This is a pretty small difference. But watching the game, didn't it feel like Texas shot the ball much better against Missouri?

Texas definitely shot much better from the floor against Missouri, with an effective field goal percentage of 0.534, compared with an effective field goal percentage against A&M of 0.476. And in terms of raw numbers, the difference seems even greater. Texas made 27 field goals against Missouri, and only 18 against A&M.

But this ignores free throws. Against the Aggies Texas was 22 of 31 from the free throw line. 31 free throws is a lot of free throws. Against Missouri, Texas was 10 of 14 from the line. When you shoot 31 free throws, it tends to make the game and offense appear disjointed and ugly. But the free throw line is usually the most efficient place to shoot from in basketball. With so many free throws, Texas' ugly shooting against A&M ends up being nearly as efficient as their more aesthetically pleasing shooting against Missouri. The thing that made the Texas offense significantly more efficient against Missouri, when compared with the game against A&M, was the offensive rebounding.

How is the Texas offense doing?

With the start of conference play, the Texas offense just hasn't felt as smooth and productive as it did earlier in the season. Of course, this is to be expected -- Texas is now facing much tougher competition then they had been during the non-conference season. This is particularly true when it comes to the quality of defense Texas has been facing.

We need to remember to consider defensive quality when we consider how well an offense is playing. One example I feel is helpful to think about is the Kentucky vs. Louisville game from New Year's Eve. In that game, Kentucky's offense only managed to score 84 points per 100 possessions. Does this mean Kentucky is bad on offense? Of course not. Kentucky averages 116 points per 100 possessions for the season, which is the 7th highest total in Division I. It is just that Louisville is really good defensively. Over the season, they have allowed an average of 88 points per 100 possessions. This average includes their disastrous performance at Providence, where they gave up 131 points per 100 possessions.

So what can we say about how Texas has done on offense, when we consider the quality of the defenses of their opponents? The plot below compares how many points per 100 possessions Texas has scored in each game with their opponent's defensive average over the entire season. It is a pretty interesting graph. (Click on it to make it bigger.) The Texas offense was really humming for much of the non-conference schedule, scoring 120 points per 100 possessions against UCLA and 130 points per 100 possessions against the hapless Nichols State. Much of this early success is probably attributable to playing against some really bad defenses. Still, Texas was destroying these bad defenses, which is exactly what you are supposed to do.


With the start of conference play, Texas played Iowa State. Iowa State is just OK defensively, although my subjective opinion is that they played really well on defense against Texas. Texas' offense performed about as well as Iowa State's average opponent. Against Oklahoma State, Texas' offense was significantly below average when compared with Oklahoma State's other opponents. Against A&M Texas' offense performed pretty well when we consider just how tough the Aggies are defensively. And against Missouri, the Texas offense did really well considering the quality of competition. It helps when you track down nearly half of the available offensive rebounds.

So is the Texas offense struggling since the start of conference play? I don't really know. The numbers seem inconclusive. But the level of competition is certainly higher, and I suspect that accounts for much of what we might perceive as Texas' offensive struggles.

Explaining the Simple Rating System

I love the Simple Rating System (SRS). It tends to closely match the ratings and the Sagarin predictor ratings. I have nothing against these other two rating systems, but when you consider just how simple and natural SRS is you start to wonder if these other systems are worth the trouble.

The Ken Pomeroy rankings are worth the trouble, particularly because they are based on per possession statistics. But given just how close the overall and SRS ratings end up being, how much value do you really get from using per possession numbers for ranking teams? I think per possession statistics are essential when comparing team offenses, or team defenses, but when it comes to ranking teams overall they are probably not that important.

College basketball SRS ratings are available at, going back to 1980. Here is a direct link to the rankings for this season. The beauty of the SRS system is that it works for pretty much any sport. Pro football, college football, hockey, it doesn't really matter. You could probably set up SRS ratings for your kid's soccer team or your fantasy football league. Of course, if you did this you would probably need to be an even bigger dork than I am. You also should probably get a hold of a linear algebra software package or library. Hit me up on Twitter, and I can make some recommendations.

SRS attempts to tell us how many points per game above or below an average level each team is. To give the numbers some context, the current Division I SRS leader is Ohio State, with an SRS of 26.6. That means that the SRS approach rates them as on average 26.6 points per game better than an average Division I team. Grambling has an SRS of -29.0, meaning that they are almost 30 points per game worse than the NCAA average over the course of a season. As of my writing this, Texas has an SRS of 15.0, which is the 27th highest total in Division I. While it varies from year to year, typically the top 5 or so schools have an SRS greater than 20, and the top 25 or so schools have an SRS greater than 15.

When we look back on a season, SRS rankings tend to match pretty closely with NCAA tournament outcomes. I did a study of this last summer. It is pretty rare for an NCAA championship to be won by a team that is not one of the top few teams in the SRS rankings. Only 7 of the last 32 NCAA champions have been less than 20 points per game above average according to SRS. Here is a list of the NCAA champions with an SRS less than 20, going back to the 1979-1980 season:

2011 Connecticut (SRS = 17.95)

2003 Syracuse (SRS = 19.01)

1988 Kansas (SRS = 15.71)

1985 Villanova (SRS = 11.99)

1984 Georgetown (SRS = 18.75. In 1984 only 2 teams had an SRS > 20. Georgetown had the forth highest SRS in 1984.)

1983 NC State (SRS = 15.22)

1980 Louisville (SRS = 15.57)

So how does SRS work? The easiest way to understand this is to create a simple example. Let's imagine that Team A has played games against Team B and Team C. We will assume that Team B has an SRS rating of 5 points above average and Team C has a rating of 5 points above average. For this first simple example, we will imagine that we miraculously know the Team B and Team C SRS ratings. We will worry about where these come from later. Let's further assume for this first example that Team A has beaten Teams B and C each by a 5 point margin. We can easily calculate Team A's SRS rating given this information. SRS rating comes from the formula

{SRS Rating} = {Average margin of victory} + {Average SRS Rating of opponents}

In our simple example, Team A has an SRS rating of 10 points above the league average. That is pretty logical, right? If you win by 5 against teams that are 5 points better than average, it is pretty reasonable that you should be considered 10 points above average.

Of course, reality isn't always so simple. Let's think through another example. What if Team A beats Team B by 5, and then loses to Team C by 5. In this situation, Team A will end up with an SRS rating of 5 points above average. Again, this is pretty logical. Based on this, SRS sees these three teams as being equals.

Now, in the real situation, we don't know the SRS ratings of Team B and Team C when we calculate Team A's rating. So how do we figure these out? Linear algebra holds the answer. Division I college basketball has 344 teams. For each of these teams, we know their average margin of victory (or defeat), and who they have played against. You could write out equations like the one that I have above for each team in Division I treating the all team SRS ratings in each equation as unknowns. The opponent SRS ratings show up in the "Average SRS Rating of opponents" portion of the equation. If you did this, you would end up with 344 equations that contain a total of 344 unknowns. Provided certain technical requirements are met, you can use linear algebra to solve this system of equations and come up with an SRS rating for each team.

So give the SRS rankings a look from time to time.