Basketball Strategy: Should the Florida State Seminoles Basketball Team Develop an Up Tempo Offense?

Last year, the young Florida State Seminoles basketball team played down tempo basketball, running a motion offense and lock down man to man defense that ran seconds off the clock. As a result, the Seminoles averaged 66.8 possessions per game, or a possession every 18 seconds. The Virginia Military Institute averaged 81 possessions per game, or one possession every 14 seconds, earning them the number one spot in the nation for pace according to the Ken Pomeroy statistics. The Seminoles, however, ranked 150th in the nation, just above average.

During this off-season, there has been a lot of talk about the Seminoles increasing the tempo of their style of play. Corey Clark, from the Tallahassee Democrat, stated the following after the Seminoles' trip to Spain:

The offense played at a break-neck pace in Spain and it wasn't an accident.

Hamilton wants his team to push the ball much more this upcoming season than it did a year ago. He said he would like to average around 80 points a game in 2009-2010. Last season, the Seminoles were last in the ACC in scoring at 68.3 points per game.

...

"There have been times where we scrimmaged each other for a game and a half and couldn't score 116 points (combined)," Hamilton said with a laugh.

The more up-tempo style is just fine with sophomore Luke Loucks, who along with junior Derwin Kitchen helped orchestrate the century-busting point totals.

"I love it," Loucks said. "For me and Derwin, it's a point guard's dream to be able to do that. No one wants to walk the ball up the floor, even though sometimes it's good to do that. This gets everybody way more involved."

...

"All our point guards did a good job of kicking it ahead," Dulkys said. "We were told, 'Guards, run the floor.'

What impact does pace have on a basketball game? Does it make a team better? Is defense sacrificed?

Should the Seminoles push the pace and depart from their old ways? In this piece, the impact of pace will be examined and the argument as to why or why not the Seminoles should push the pace will be made.

Photo from here.

Wisconsin and North Carolina were at the opposite ends of the spectrum for the pace at which a basketball game is played. Florida State was right in the middle. The Tarheels' offense was dependent on scoring in transition of their opponents mistakes. They averaged 75.7 possessions per game. Wisconsin, on the other hand, averaged 58.8 possessions per game, which placed them at 341st in the country. There are only 344 Division I basketball teams in the country. Despite the variations in the pace at which Florida State's opponents typically played, the Seminoles managed to take control of games and play the game at their pace, relatively deliberate with little acceleration of pace or sense of urgency. This likely resulted from an inexperienced group of talented players who didn't quite gel with the a senior leader who was more than capable of taking over a game.

With a year under their belt, and the addition of a stellar guard, the 2009- 2010 Seminoles have the opportunity to change their style of play. While in Spain, Florida State played at a blistering pace, scoring at will while putting up some impressive numbers. Many times during the past season, the argument was made that the Seminoles should push the pace and use their athletic ability to their advantage.

Pace, possessions per game and tempo can be defined in any number of ways, but the terms are typically used interchangeably. For the sake of this discussion, the definition of possession that will be used is the one presented by Dean Oliver in his book Basketball on Paper: "In a game, a team alternates possessions with its opponent so that, at the end of the game, each team has just about the same number of possessions on which to try to score."

Oliver uses the following formula to calculate possessions:

Possessions = FGA - (OffReb/OffReb+OppDefReb) x (FGA -FGM)x1.07 + Turnovers + 0.4xFreeThrowAttempts

This based on the argument that a possession can end with a field goal attempt that is not rebounded by the offense (meaning a made shot or a rebound by the defense), a turnover or free throws. In any game, a team will have about as many possessions as their opponent. For example, the Seminoles had 66.8 possessions per game, so did their opponents. As you can see, the formula above takes into consideration the potential statistical measures to identify a change in possession.

There are a number of ways to increase possessions in a basketball game and ultimately increase the pace of the game. The Seminoles offensive set is based on a motion offense, which is a patient offensive set that depends on a team reading a defense and the way it reacts to your offensive movement. This is a dynamic offense in that it requires constant motion but it develops shots late in the shot clock. Recall the way that Wisconsin's offensive set. They moved the ball around the court until late in the shot clock. As mentioned previously, the Badgers were one of the "slowest" teams in the country last year.

As we move forward in the discussion, keep in mind that a team can increase the pace of the game on offense in two ways. This is a relative over simplification, but it will help develop a frame work in which to analyze the Seminoles. A team can either increase the pace of the game by scoring in transition, meaning taking advantage of their opponents mistakes and pushing the ball up the court quickly in order to score easy layups. Think about the way North Carolina played last year. They wanted to run the court. They wanted to get into open space and run the score up on you without you having the opportunity to play defense.

The other way is to shoot early and often in the shot clock. A classic modern example of this is the way the Phoenix Suns played under the watch of Mike D'Antoni. His 'system' is based on taking a shot within the first seven seconds of the shot clock. This requires his teams to push the ball up the court off an opponents basket and to make quick decisions. If you have a team that is designed to play within such a system and plays efficient basketball, it can be a successful.

If a team can fly up and down the court but can't put the ball in the basket, who cares how many possessions they have? Last year, Georgia Tech was one of the 'fastest' teams in the ACC, But, they only scored on about 50% of their possessions. Their offensive efficiency, which is defined as how many points a team will score given 100 possessions, was 100. They scored about 1 point per possession. Or, they, on average, scored a 2 point field goal on every other possession they had. Their eFG% was only 47.9%, which was 214th in the country, meaning they got up and down the court but they couldn't finish.

Compare that to Wisconsin who played at a brutally slow pace of 59.9 possessions per game, 334th in the country. The Badger's offensive efficiency was 112.6, 36th in the nation. They took their time to get an open shot and did not rush their possessions. By doing so, they took good shots and were very effective and efficient with their scoring. The top team in the country for offensive efficiency was the University of North Carolina who scored 117.7 points per 100 possessions. But, they played at a blistering pace of 75.7 possessions per game (5th in the country), suggesting they scored a lot of easy baskets in transition. UNC took advantage of opportunities and pushed the ball up court. This provides a nice example of how the statistics match what was witnessed on the court.

The discussion so far is intended to develop a framework with which to understand the impact of pace on success. The following section will look at the fastest teams in the country, the slowest teams in the country and the pace of the ACC. Given the new additions to the Seminoles and Coach Hamilton's comments, should the Seminoles play at a higher pace? Hopefully, by the end of this article, you will have the tools to make the decision for yourself.

Pace Analysis

One would think that teams that play at a higher pace naturally score more points and therefore are more likely to be ranked higher. However, pace does not automatically translate to points. To look at the impact of pace on success, offensive efficiency and defensive efficiency, I looked at the top 60 and bottom 60 teams in college basketball as ranked by pace.

The top 10 teams in pace include the following: (The numbers in the brackets next to the team corresponds to the conference rank)

Virginia Military Inst (BSth) 81.3 ( 1)

Northwestern St. (Slnd) 76.1 ( 2)

Longwood (ind) 76.0 ( 3)

Texas St. (Slnd) 75.8 ( 4)

North Carolina[1] (ACC) 75.7 ( 5)

Chicago St. (ind) 75.3 ( 6)

Wake Forest[4] (ACC) 75.0 ( 7)

Alcorn St. (SWAC) 74.9 ( 8)

South Carolina (SEC) 74.1 ( 9)

Providence (BE) 73.1 (10)

The bottom 10 teams in pace include the following:

Northwestern (B10) 59.9 (335)

Samford (SC) 59.6 (336)

Western Illinois (Sum) 59.4 (337)

Air Force (MWC) 59.2 (338)

Princeton (Ivy) 58.9 (339)

Oregon St. (P10) 58.8 (340)

Wisconsin[12] (B10) 58.8 (341)

Washington St. (P10) 58.6 (342)

Denver (SB) 57.4 (343)

Iowa (B10) 56.7 (344)

This gives you some sense of the types of teams that we are discussing. These are big names programs who play very different styles of basketball. The first interesting point is that if you compare the overall rankings, the Pythag winning percentage as developed by Ken Pomeroy, there is absolutely no statistical difference between the two groups. Meaning that simply looking at the two groups, the top 60 and bottom 60 teams as ranked by pace, are no different with regards to their success as identified by Pythag. Immediately, this shows us that you don't need to be fast to be good, and this should come as no surprise. All this says is that just because one team plays at a higher pace than the other, doesn't necessarily mean that they will be 'better.'

To pursue details about these relationships further, intergroup analysis was performed, meaning how did pace related to offensive efficiency, defensive efficiency within the top 60 and the bottom 60. To to this, a correlation coefficient was generated between all of the variables. If you remember from some of the pre-game analysis during the year, the correlation coefficient essentially looks at how strong a relationship there is between two groups of variables, with 0 being no relationship and 1 or -1 being a perfect relationship. If the value returned from the analysis is 1, for example, it means there is a direct relationship between the variables, as one goes up the other goes up. If the values is -1 one, they variables have an opposite relationship, as one goes up the other one goes down.

Fastest Tempo Teams

If you look at the correlation coefficient comparing pace to offensive efficiency for the top 60 teams as ranked by pace, there is no relationship (Correlation Coefficient = 0.07). If you compare their pace to defensive efficiency, there is no relationship (CorrCo = 0.05). If you compare it to their Pyth Ranking, there is no relationship (CorrCo = -0.05) If you, however, compare offensive efficiency to the Pyth ranking the value is 0.77. The defensive efficiency correlation coefficient is - 0.68. This makes sense as the better a team is on defense the lower the defensive efficiency will be.

The relationship between Pythag Ranking, offensive efficiency, defensive efficiency and possessions can be demonstrated graphically. Note how the dots fit the lines very nicely in the case of defensive efficiency and offensive efficiency and not so much for possessions:

From this data, one can generate a multiple regression analysis, which is a fancy way of saying we can create an equation that uses the offensive efficiency, defensive efficiency and possessions to calculate the Pythag value. As you create the model, you are able to see what kind of impact that each variable (OffEff, DefEff and Poss) has on the outcome measure, meaning which one has the most weight.

For the fastest 60 teams in the nation, the equation that is generated is the following:

Pyth = 0.4412 - 0.02261(DefEff) + 0.025626(OffEff) - 0.00362(Poss)

The actual numbers aren't important. But what this shows is that possessions have little weight in this equation and offensive efficiency has the most. This equation has an R-squared value (similar to the correlation coefficient) of 0.77 (again, the closer to one the better). If you take the possessions out of the equation the value drops to 0.76, which is minimal compared to the drop of 0.46 if you take out the offensive efficiency.

Slowest Tempo Teams

The same analysis was done for the 'slowest' 60 teams in the nation. Before we go further into the analysis, it should be noted that there was no difference in the average Defensive Efficiency, Offensive Efficiency or Pythag rating between the top 60 and bottom 60 teams as ranked by pace. The correlation coefficients for Offensive Efficiency, Defensive Efficiency and Pythag and possessions were again not significant. The relationship between OffEff and Pythag was 0.68 and DefEff and Pythag was -0.53. These correlations were slightly less than those of the fastest 60 teams.

The relationships are shown here graphically:

Possessions have nothing to do with success in teams that play at a slower pace. This can be shown in the equation that is generated in the multivariate analysis model:

Pythag = 0.5863 - 0.02048(DefEff) + 0.024644(OffEff) - 0.00812(Possessions)

The fit of this model isn't as good as it is for the fastest 60 teams. But, when you take out the possessions variable, it has little impact on the fit of the equation, reinforcing the point that there is something else going on for these teams that equates to success. I suspect that it is the impact that the slower paced game has on their opponents.

The Elite Teams

We have looked at the impact of possessions on success for the 'fastest' and 'slowest' teams as defined by pace. But what about the top teams in the country? What makes them good? Is it the offensive efficiency, defensive efficiency or their pace? This again is an oversimplification, but to create a model with all of the variables would huge and doesn't relate to our question about Florida State: should they try to play at a higher pace. Once we have looked at the top teams in the country we will look specifically at the ACC and how all of this relates to the Seminoles. Unfortunately, trying to create a model for these teams is difficult and that is because the numbers are clustered and there is little variation or scale. Trying to modify to spread the data points out (such as log regression or square root analysis) makes no difference either. Simply put these teams are good at everything.

The graphs highlight this nicely (I honestly don't know where that rogue point is coming from, combed through the data and it is still shows up despite not having a value, so just ignore it):

Impact of Pace in the ACC

Before we look at the impact of pace on the ACC here is a reminder of the final standings with the respective statistical measures:

Table adapted from Pomeroy.com

The amazing thing to notice upon first glance at this table is the overall strength of this conference in offensive and defensive efficiency. There are relatively few outliers.

How do these teams do in terms of pace? Are there general trends and associations we can make from the data and does it fit what was witnessed on the court last year?

Clearly, there is a lot of information to digest from that chart. Again, all of the information in that chart is from kenpom.com. The CorEff columns are the respective correlation coefficients for pace as compared to Offensive or Defensive Efficiency.

First, the range of pace in the ACC is remarkable. UNC and Wake and Georgia Tech pushed the ball up and down the court, but look at their respective ranks in conference and then look at their offensive efficiency. This is the key piece to the entire argument of this piece. If you're not efficient on offense, it doesn't matter how fast or slow you play.

Next, look at the correlation co-efficient to pace on offensive efficiency. If North Carolina was out running its opponents, it typically there was a positive correlation to offensive efficiency. It's not a great correlation but it's there. However, when UNC got out and ran, they forgot to play defense as it had a negative impact on their defensive efficiency. Florida State, however, saw a negative impact on it's offensive efficiency as the pace went up. Again, it's not a great correlation, but it's appreciable. Reflect on the season. When we played 'fast' it was hurried, mistake filled, inefficient basketball and I think the increased 'pace' was a result of increased turnovers. Increasing turnovers will increase possessions as you can see from the equation listed at the beginning of this piece. 

To look into the relative positive or negative effect of pace on OffEff, I looked at what percentage of a team's shots were three pointers. Clearly, a team can push the pace and shoot a three pointer at any time in transition or early in the shot clock. I made the assumption, whether it is good or bad, that teams that shoot more three pointers are likely to play at a slower pace or that an increased pace will have a negative impact on their offensive efficiency. First, ACC teams are not as reliant on 3-pointers as we think they are. At first, it looks like this model fits. Wake runs up and down the court all day, pace has little impact on their offensive efficiency and they shoot very few three pointers. Miami played a slower tempo game and their offensive efficiency was negatively impacted by pace and they shot a relatively large number of threes compared to the rest of the conference. But, unfortunately, this argument didn't hold up throughout the conference. Something to keep in mind.

Lastly, I wondered what impact height, or having a true center on a team impacted pace, thinking if you have a big man, you're not likely to press as much if he is the key to your offense. Well, as your effective height increases so does your pace, but the CorEff was only 0.3, not a strong correlation and this was a small sample set. I also found out that increased pace is typically associated with more points scored by your power forward and a taller center. This may be due to good centers pulling down rebounds and teams getting out in transition off the rebounds. The scoring at the PF data is likely team and personal dependent.

Should Florida State Up the Tempo?

Three out of the top 5 teams in the ACC last year placed at a relatively uptempo pace for the conference. Now, the differences in possessions may be subtle, 1 or 2 possessions here or there, but they can have a significant impact on a game, particularly if a team is efficient on the offensive end.

Last year, the Seminoles struggled for a good portion of the season on the offensive end as turnovers plagued the team early (and late) in the season. In fact, the Seminoles would have played at a slower pace if it wasn't for the turnovers.

With improved chemistry and experience, the Seminoles should take advantage of their athletic ability and get out in the open court and score in transition. I do not think they should push the shot clock and try the D'Antoni method of basketball; we are not that type of team. An uptempo style of game suits the athletic abilities of the team. However, unless you're efficient, pace doesn't matter.