There’s an article that gets written this time of year, right during the heart of conference play. Duke’s defense could be a problem–the header in an article from two days ago. Duke’s defense needs to improve by March–published the last week of January 2016. Defense continues to plague Duke after a devastating home loss in mid-January 2015. (Last year, the discussion was mostly about Grayson Allen’s incessant horseplay, but the Blue Devils did get torched badly enough by NCSU and Dennis Smith Jr.–again at home in late January–that Coach K ceremoniously banned the team from the locker room.)
So this is a pattern. More importantly, it didn’t always used to be this way. I’m too young to be nostalgic about ’80s and ’90s Duke–the hey-day of “pesky gym rat” defenders like Steve Wojciechowski–but I recall watching Shane Battier and Shelden Williams patrol the paint with authority in the early-aughts. I remember Battier covering so much ground during a 2001 Final Four game, his name called so crisply by Jim Nantz (“re-JECTED by Battier!), that I imagined myself as the M.O.P. during the next day’s dodgeball game. It didn’t make sense, but hey, at least it was an ethos.
Nevertheless, that’s anecdotal. Fortunately, KenPom.com has data going back nearly two decades, data which includes the adjusted offense and adjusted defense ratings (which factor in pace and opponent quality) for every team in the NCAA. Here’s Duke’s rank (with 1 being the best in the country) for each of the past seventeen years. Keep an eye on the year 2008, and see if you can figure out why it’s interesting.
| Year |
Offense Rank |
Defensive Rank |
|
2002 |
1 |
1 |
|
2003 |
13 |
15 |
|
2004 |
2 |
3 |
|
2005 |
13 |
2 |
|
2006 |
1 |
16 |
|
2007 |
44 | 6 |
|
2008 |
13 | 7 |
| 2009 | 7 |
28 |
| 2010 | 1 |
5 |
|
2011 |
5 | 9 |
| 2012 | 8 |
79 |
|
2013 |
4 |
26 |
|
2014 |
1 |
86 |
|
2015 |
3 |
11 |
|
2016 |
4 |
86 |
|
2017 |
6 |
47 |
| 2018 (through 2/8) | 2 |
78 |
Did you get it? 2008–an entire decade ago–is the last year during which Duke had a better defense than offense. Every year since, including its championship runs in 2010 and 2015, the offense has led the way. The offense has never disappeared from the top-ten nationally, but the defense–especially in the last seven years or so, starting with the 2012 team that bowed out to Lehigh–has fallen off a cliff.
So why–and how–have things changed? There are a few culprits. Perhaps the most commonly-invoked explanation is Coach K’s adoption–or gravitation–toward recruiting “one-and-dones,” or top prospects who may only stick around for one year before jumping to the NBA. Duke’s first one-and-done was Kyrie Irving in 2011, so let’s split the data into two sets–2002-2010, and 2011 onward–and see how the numbers stack up.
|
Years |
Avg. Offensive Rank | Avg. Defensive Rank |
Avg. # “One-and-dones” |
|
2002-2010 |
10.6 | 9.2 | 0.0 |
| 2011-2018 | 4.1 | 52.8 |
1.4* |
*Doesn’t assume any one-and-dones for 2018, although there will almost certainly be multiple.
From the year after Duke’s 2001 championship run up through its fourth championship in 2010, the Blue Devils were fairly balanced–the offense and defense were around tenth-best. Since then, they’ve become bipolar: The offense is typically top-five, but the defense is tiers below what it used to be.
It’s worth noting that this doesn’t mean the glory days are over, by any means. Duke has flamed out in the tournament a few times in recent memory, but their overall degree of success hasn’t been far from its previous level. As the below table shows, their win percentage is identical; they’ve slightly underperformed in the tournament while also being less consistent there.
|
Years |
Overall Win % | Avg. # of Tournament Wins |
Variance in # of NCAA Wins |
|
2002-2010 |
80.0 | 2.4 | 3.0 |
| 2011-2018 | 80.0 | 2.0 |
4.3 |
But back to the matter at hand: defense. I had a few ideas for the decline. A roster full of one-and-dones has some tell-tale characteristics: First, they likely have a number of superstars who are deserving of playing big minutes. Second, their depth is savaged each year as players jump to the NBA or, as has happened recently, they transfer to a program that can give them more playing time. (Since 2011, six top recruits have transferred from Duke; another, Rasheed Sulaimon, was dismissed from the program.) My first hypothesis–triggered by the Duke-UVA game this year in which the Blue Devil starters played all but six minutes–is that Duke’s starting five gets overused in games and tires, leading to decreased effectiveness on the defensive end.
I looked at the minutes played for each of Duke’s top five most-used players over the season–would the lineups stacked with top recruits be more imbalanced?
Not really. Duke’s recent lineups are fairly top-heavy–their five go-to players take up roughly three-quarters of all available minutes–but that’s not a departure from how Coach K has typically deployed his lineups. Moreover, the correlation between defensive rank and this measure of player overuse is practically zero (-0.014).
|
Years |
% Minutes from Top Five Players* |
| 2002-2010 |
74.3% |
| 2011-2018 |
74.9% |
*Straight average, not weighted.
I also explored how regular seasons have expanded since 2006–with a change in how multi-game tournaments are counted, teams now play 2-4 more games over the regular season. Perhaps this is another added cause for fatigue, but the change predates the drop-off in quality by a few years, so it doesn’t hold up as an explanation.
The second argument is that all freshmen simply don’t know how to play defense, the logic being that one-and-dones roundly take the place of grizzled upperclassmen who–while lacking a silky smooth offensive game–know how to scrap and claw on defense. I’ve always found this argument a bit weak: Anthony Davis, as a freshman at Kentucky, was a monster; Justise Winslow, while at Duke, had no problem shutting down opposing wings.
But perhaps these were the exceptions, I figured. I ran the same analysis as before, but this time looking at the percentage of minutes allotted to Duke freshmen each season.
|
Years |
% Minutes from Top Five Players* |
|
2002-2010 |
23.3% |
|
2011-2018 |
35.5% |
*Straight average, not weighted.
This is what we’d expect: In the new one-and-done era, more freshmen are earning playing time year-after-year, and fewer upperclassmen are seeing the court. But what does it mean for the effect on defense?

It’s not a perfect relationship, but there is a trend: The more minutes freshmen play, the worse the defensive rating. (The correlation is .395.) This is a start, although it certainly doesn’t seem to be a silver bullet.*
*Let’s note that the one “good” defensive year in recent memory is 2015, Duke’s last national championship season. While their final defensive ranking was 11th, they entered the NCAA tournament ranked 62nd in that category before playing a solid (and somewhat lucky) six games against teams they were well-suited to stop (e.g., Utah, Gonzaga, and Wisconsin).
What I haven’t figured out, though, is why freshmen are worse defenders. My initial guess was that they foul more–and put the opposing team on the foul line more frequently–but in the eight full seasons for which we have this data, no trend shows up. (In fact, the correlation between freshmen minutes and opposing FT attempts is slightly negative.) We do see a clear relationship between freshmen minutes and opposing field goal percentage: The more freshmen play for Duke, the better the opposition shoots. (Correlation: .462.) But this is just a way of restating what we already know: The defense is worse when freshmen play because the defense is worse.
There’s also the delicate question of whether Coach K has lost his fastball. If we zoom in on the 2002-2010 period, before Duke added its first one-and-dones, the defensive rating was trending in the wrong direction…:

…while it was actually playing freshmen a little less frequently:

There are many questions remaining: Why are freshmen bad defenders? Have rule changes–particularly a recent de-emphasis on charge whistles that Duke notoriously has tried to earn, combined with an increase in overall fouls–played a role? And why, with the extended preseason that the NCAA approved starting in 2013, hasn’t Coach K been able to crack this code? Does he not know how to *gasp* communicate with millennials?
I’m interested in another question, though: Seeing this trend, what would Coach K think of it? I’m struck by that two-year offensive dip in 2007 and 2008 which, really, is the only blemish on that side of the ball in the past two decades. I wonder if K simply grew weary of coaching players like Greg Paulus and DeMarcus Nelson–players who couldn’t get enough buckets to down a not-good VCU squad in the NCAA first round–and decided he needed more offensive ammunition. Although his roster for the next few years was laden with upperclassmen, he soon snagged Kyrie Irving and then Austin Rivers, two unbelievable talents. The defense suffered more, but the offense was fantastic, and Duke has never looked back.
If you have the same job for years on end, even hard-ass disciplinarians get to a point where they say, “Screw it, let’s have some fun.” Is that where we are now with Coach K? Given how frustrated he looks on the sideline when his team can’t grab a single defensive rebound, probably not.
But the trend is clear: This is the new normal. Duke isn’t a top defensive team–and hasn’t been for a while. Fans and commentators would do well to stop acting so surprised.





