After I presented my early Top 25 last week, Kansas PG Naadir Tharpe left the team for personal reasons, and PG recruit Devonte Graham committed to the Jayhawks. I’m not going to fully re-run the projections again until after the Spring signing period, but based on the early numbers, my model seems to be a tad higher on Kansas than most experts. My model’s confidence is based on Bill Self’s per-possession track record (as reflected in 10 straight conference titles.) In particular, the model expects Kansas’ defense to bounce back substantially this year. But the more I thought about it, the more I realized there was plenty of nuance to the defensive projection.

Before I get to the defensive discussion, let me say that there is a high probability that Duke will have a better offense than Kansas next year. Duke’s returning players (Quinn Cook, Rasheed Sulaimon and Amile Jefferson) were more efficient than Kansas’ returning players. Perry Ellis was very efficient, but Wayne Selden was not a natural scorer last year, and the Kansas PG situation is still a little unsettled. Both teams bring in multiple impact freshmen, but based on the entire roster, Duke has more offensive weapons. The more interesting question is whether Kansas or Duke will have the better defense in 14-15.

Unfortunately, our current methods for predicting defense are not great. Individual defense is extremely poorly measured with current statistics. I can tell you that Duke’s Matt Jones had a higher steal rate than Rodney Hood. But that doesn’t mean Jones was a better defender. Hood’s size might have altered more shots. We also don’t know who Jones was matched-up against defensively. We don’t have a statistical measure of whether Jones took more chances and gave up more drives to get those steals. And we don’t know if Jones got a few more deflections because he played sparingly and didn’t need to conserve his energy for the offensive end of the floor. Perhaps if Jones played starters’ minutes, his defense wouldn’t be nearly as good.

Team defense is also astonishingly unpredictable. North Dakota St. brought everyone back last season, and yet NDSU’s defense fell from 59th to 131st. NDSU’s offense was good enough that Saul Phillips’ squad won the Summit League and defeated Oklahoma in the NCAA tournament. But even for teams with little roster turnover, the small sample of college basketball games in a season means we do not always have a true barometer of a team’s defense.

The larger sample of games for each coach may be more reliable, and that’s why my model includes coach effects on defense. The next table shows the national rank of Bill Self and Mike Krzyzewski’s defense in each season over the last 10 years, measured on a per-possession basis, adjusting for opponent and venue.  The table makes a pretty compelling case that Bill Self has been a better defensive coach over the last 10 years, and that Kansas will have a better defense in 14-15.

Def Rank

Bill Self

Mike Krzyzewski

2013-14

31st

116th

2012-13

5th

31st

2011-12

3rd

81st

2010-11

11th

21st

2009-10

9th

8th

2008-09

9th

36th

2007-08

1st

8th

2006-07

1st

7th

2005-06

3rd

18th

2004-05

25th

3rd 

But the personnel situation for Duke makes it a little less certain. Duke’s biggest problem on defense last year was that they did not play a true low post defender. Duke essentially played Jabari Parker out of position at the center spot for much of the season, rather than give Marshall Plumlee a chance to develop in the paint. Mike Krzyzewski clearly thought he had a better chance to outscore teams with that small lineup, but the defense was bad all year. This year with Jahlil Okafor in the paint, and a more mature Marshall Plumlee playing more minutes, Duke will almost certainly improve its interior defense. An argument can be made that Okafor is a uniquely talented center, and Duke’s defense will look more like it did in 2010, when Duke had Brian Zoubek in the middle. Meanwhile Kansas, which has often featured 7 foot shot blockers in the middle, may end up playing a slightly smaller front line of Cliff Alexander and Perry Ellis next year.

The above table also raises another issue. 13-14 was the worst defensive season for both Bill Self and Mike Krzyzewski since Ken Pomeroy began tracking the per-possession stats. And there was something else that changed in 2013-14 besides each team’s personnel. The NCAA also changed its rules about hand-checking perimeter players and drawing charges. And as the next table shows, Bill Self and Mike Krzyzewski’s teams both struggled with sending players to the free throw line after this change. Bill Self’s team particularly struggled in this area. While the average NCAA team sent players to the line 4 to 5 more times per 100 shots, Kansas sent players to the line 13 more times per 100 shots last year.

And this may not have been the only impact of the rule changes. If Duke’s defenders could no longer step in front and draw charges as they did in previous seasons, that may be more of a permanent concern for Krzyzewski’s team. Both teams were worse at creating turnovers in 2013-14, but both showed similar drops to the national average.

Def Free Throw Rate

Bill Self

Mike Krzyzewski

Bo Ryan

National Average

2013-14

45

41

27

41

2012-13

32

33

26

36

2011-12

33

33

30

37

2010-11

32

30

33

38

2009-10

31

34

36

38

2008-09

35

31

34

37

2007-08

31

32

25

37

2006-07

34

30

28

37

2005-06

32

28

31

36

2004-05

38

32

28

37 

It is obviously too soon to tell whether the rule changes have a unique effect on these two coaches. A strong argument can be made that the primary reason Kansas fouled so much last year was because of the team’s excessive youth. 2013-14 was the youngest team of Bill Self’s career.

But it is worth asking whether coaches that are particularly good at teaching their teams to play defense without fouling (Bo Ryan), coaches that play zone defense (Jim Boeheim), or coaches that play pack-line defense (Tony Bennett) will be better prepared to play elite defense going forward. Bill Self is clearly one of the best teachers of physical man-to-man defense in the nation, but it is harder to bank on his historic track record in an environment where the rules have changed.

Overall, I remain optimistic about both team’s defense next year. If you have watched any of Duke or Kansas’ recruits play, or read the scouting reports, it is clear they have the type of athletes coming in to be a real threat defensively.

Paul Biancardi says Duke’s Justise Winslow might be the best defensive player in the country. Duke’s Tyus Jones has such an impressive ability to drive and create using his low-center of gravity, I will be shocked if he doesn’t have the ability to keep opposing guards out of the lane. Kansas’ Cliff Alexander is an aggressive defensive rebounder. And Kansas’s Kelly Oubre has an impressive wingspan, the kind of arms that should cause a lot of deflections this year. Certainly, the incoming personnel are capable of playing elite defense.

But the changes in the rules make me nervous about including defensive coach effects in my model. An unusual number of great defensive coaches had mediocre defensive seasons in 13-14. In next fall’s projections, I may need to lower the impact of coach effects on the predictions. I’m projecting Duke’s offense to be better than Kansas’ offense, and I’m quite confident in that prediction. I’m also projecting Kansas’ defense to be better than Duke’s defense, but that is far from guaranteed.