Statistical Analysis of the 2006-07 Duke Men’s Basketball Team

Last year’s loss to VCU in the first round of the NCAA Tournament has left a bad taste in every Duke fan’s mouth. I’ve avoided the loss and tried to block it out for a while now, but with the start of the new season rapidly approaching, it’s an appropriate time to analyze what went wrong.

I could give my opinion on why we faltered last year, but that wouldn’t be a good way to conduct an objective analysis. To eliminate any personal bias, let’s turn to something that doesn’t lie, the numbers. My friend James, who is one of those brilliant Duke kids who is going to win a Nobel Prize in quantum mechanics or something like that someday, put together a table analyzing the individual contributions of all of our players from last year’s squad. The table is based on an accepted metric to analyze the relative importance of basketball players in terms of wins and losses, a system created by The Wages of Wins .

I’d like to stress to all of you the theory behind this model. It’s tempting just to look at points scored as an indicator of a player’s ability, but the “Win Score” philosophy emphasizes efficiency. So a player that scores a lot of points but misses a lot of shots (i.e. Allen Iverson) is an inefficient player. The “Win Score” also emphasizes that getting rebounds and steals are in fact just as important as taking shots because they get possessions back for your team, which in turn leads to more scoring opportunities. Further, blocks prevent field goals and create rebound opportunities, while assists help indicate how much a player helps facilitate the offense. Thus, getting possessions back can be as important as getting shots. I would also like to stress the validity of this statistical design; the guys who came up with this metric achieved a 95% correlation between values for each player and actual team wins.

Before we delve into the stats, I’d like to define a few terms. If you’d really rather not get into all the specifics, you can skip the definitions (the bold terms) and just remember that the higher the score the better.

Win Score = Points + Rebounds + Steals + ½Assists + ½Blocked Shots – Field Goal Attempts – Turnovers - ½Free Throw Attempts - ½Personal Fouls

Win Score per Minute (a more accurate tool to compare players who have a significant disparity in minutes played) = Win Score/Min (duh)

PAWS (Position Adjusted Win Score) is used to compare a player to an average player at the position. Average Win Score/minute values for each position:

Center: 0.225
Power Forward: 0.215
Small Forward: 0.152
Shooting Guard: 0.132
Point Guard: 0.128

PAWS (Position Adjusted Win Score) = Win Score – (Average Win Score at Position Played * Minutes Played)

PAWS/min = Win Score per minute –Average Win Score at Position Played

From the Duke men’s basketball team statistics from last season, the following table was constructed:

Table 1- Duke Basketball Statistical Analysis/ ’06-‘07

The two most important values from this table are the WS/min and PAWS/min values. Only considering players who played at least 100 minutes (Jamal’s 30 minutes of garbage time at the end of games isn’t a large enough sample size to be considered), here is our ranking of WS/min (Win score/min):

1. Josh McRoberts: 0.262
2. David McClure: 0.216
3. DeMarcus Nelson: 0.173
4. Brian Zoubek: 0.162
5. Jon Scheyer: 0.136
6. Gerald Henderson: 0.099
7. Greg Paulus: 0.084
8. Lance Thomas: 0.056
9. Martynas Pocius: -0.016

Now let’s see the ranking for PAWS/min (Position Adjusted Win Score per minute) values for players who played at least 100 minutes:

1. Josh McRoberts: 0.037
2. DeMarcus Nelson: 0.021
3. Jon Scheyer: 0.008
4. David McClure: 0.001
5. Brian Zoubek: -0.063
6. Gerald Henderson: -0.029
7. Greg Paulus: -0.048
8. Lance Thomas: -0.159
9. Martynas Pocius: -0.168

So what does all this mean?

By both statistical measures, Josh was by far our best player and Marty was by far our worst. This makes sense since Josh was the only player in the NCAA to average at least 10 points, 7.5 rebounds, 2.5 blocks and 3.5 assists per game while Marty’s stat line included only a single block and 3 steals in 139 minutes on the floor.

Paulus was only our seventh best player last season in both statistical measures. Paulus was punished for committing the most turnovers, the most fouls, and for recording low rebound and block totals compared to other players. The average Win Score/min for the point guard position is .128 so even compared to the average point guard, Paulus still fares poorly.

Demarcus and Dave are both relatively high on the lists because they both rebound well, hit the boards hard, and get a lot of steals while still taking care of the ball on offense.

This list doesn’t reflect the relative worth of each player but it does highlight how efficient each player is when he is on the court. These rankings don’t explain everything. Perhaps Josh was rated so highly because he left his man to pick up extra blocks. And one thing this metric definitely does not measure is heart, chemistry, and all the intangibles a guy brings to the table.

The take home message here is that a lot of our players from last year’s team need to become more efficient on the court. Paulus, Zoubek, Henderson, and Pocius in particular, are all going to have to take care of the ball better, hit the boards harder, and be more active on defense for this team to be successful.