Worst Duke Defensive Performances

It's been a while since we've done college basketball here, and I have a few things planned for the coming months, including a systematic re-do of last year's tournament simulation. In the meantime, I want to nip any "Matt's a Duke fanboy" criticism in the proverbial bud by doing a negative post about Duke.

The tempo-free era of college basketball began with the 2003-2004 season as Ken Pomeroy started putting posting his rankings based on offensive and defensive efficiency. Duke, for whatever reason, has always done pretty well during the regular season in Pomeroy's adjusted (for opponent offensive and defensive strength) rankings:

Year/Offensive Efficiency/ Defensive Efficiency
2010/1/18
2009/10/20
2008/11/9
2007/40/5
2006/5/13
2005/15/1
2004/2/4

Nevertheless, Duke has put up some hard-to-believe individual games on both sides of the ball. Defense, however, is so intricately tied to Coach K's philosophy that it is the better choice for a comprehensive breakdown. Defensive ability puts players on the floor for Coach K, and can limit the minutes of potentially explosive offensive perimeter players, if they cannot grasp Duke's defensive philosophy (c.f. Taylor King, Elliot Williams, Andre Dawkins). Usually, the result is a solid defensive gameplan that has consistently been in the top 20 the past six and a half seasons.

There have been great defensive efforts, and not always against inferior competition. In fact, the best defensive performance of the 228 games recorded on kenpom.com is this one, played December 19, 2009 against Gonzaga in Madison Square Garden. Duke held the Zags to an efficiency of 57.1, which is absolutely stellar when you consider that Gonzaga's 2009-2010 average is 111.3! But it has not been all good news for Duke this year, and it is an historic defensive lapse that is the reason for this post. Here are the top 10 worst defensive performances since 2003-2004:

#10 (118.4): January 14 2006 at Clemson (Duke W 87-77). We start with a surprising performance in that not only is it a Duke win, but also a win on the road. Although the 2005-2006 Duke team was frequently spotlighted for it's excellent defense, led by Shelden Williams' shotblocking ability, on this night at Littlejohn, things did not go well. Giving up more than a point per possession to a Clemson team that did not make the NCAA tournament and finished 94th in the nation in adjusted defense was simply not acceptable. Duke had some other defensive clunkers along the way, but eventually fell in the NCAA tournament due to their worst offensive performance in the tempo-free era. (But as Alton Brown might say, that's another post...)

#9 (119.0) November 16 2008 vs Rhode Island (Duke W 82-79). This was the night that RIU's Jimmy Barron almost got a permanent middle name from Duke fans. Barron hit seven straight three-pointers before Dave McClure came off the bench to get a hand in Barron's face, causing him to heave an air-ball with 1:24 remaining and Duke down by 2 (ESPN's play-by-play calls the shot a 2-pointer, but I was at the game and remember differently). This was by far the loudest I've heard Cameron in a non-conference game. I also had the pleasure of high-fiving Jimmy Barron as he ran off the court, and I have not washed that hand since.

#8 (119.1) March 8 2009 at North Carolina (Duke L 79-71). This was the closer of the two Duke-UNC games last year, and both teams were had worse efficiencies than the earlier matchup. The pace was 13 possessions slower as well. The biggest standout is that the Tar Heels rebounded over 40% of their missed shots, while Duke's offensive rebounding percentage was a dismal 18.8%. Eww.

#6t (121.3) February 18 2004 at Wake Forest (Duke L 90-84). Duke's best team during the Tempo-Free era only lost six games all season, and every time they did, they allowed more than a point per possession. It was Duke's second loss in a week (they lost at NC State earlier) and is perhaps the best lesson for Chicken-Little Duke fans and Duke haters alike, concerning this year's Duke team. Despite the efforts of the media and the common fan, "consistency" just doesn't mean anything. Especially in ACC road games. A team can put up clunkers in consecutive games in January and still get within an Okafor of the national championship game.

#6t (121.3) January 21 2006 at Georgetown (Duke L 87-84). From the second paragraph of that ESPN recap: "That's my child," the elder Thompson said. "I love my child. After all he's had to go through, he deserves this." Duke never led in a classic execution of the Princeton offense, and JJ Redick scored 41 points, and with Duke almost matching the Hoyas' offensive efficiency with a 117.3 rating. Free throws were the difference, as Georgetown got to the line more often than Duke. It is unclear what the elder Thompson said of Duke's destruction of Georgetown in 2009.

#5 (124.5) February 22 2009 vs Wake Forest (Duke W 101-91). Last year was Duke's worst from a defensive standpoint, and it shows in this top ten list. This game was the polar opposite of the game a few weeks earlier in Winston-Salem, a two point loss for Duke in which neither team topped 95 in efficiency. The rematch in Cameron however had a similar number of possessions and basically no defense.

#4 (125.8) February 20 2005 vs Wake Forest (Duke W 102-92). An eerily similar game to the one above, although this rematch played out exactly like the preceding game at Wake Forest (see below); the only difference was the result for Duke. Interesting point about the 2004-2005 Duke team; if you check here, you can see all the factors that were correlated to Duke's defensive performance that year. Interestingly, nearly all the factors are in bold, indicating a significant (at the 95% confidence level) correlation. Duke's defensive performance that year was strongly tied to their opponents' shooting, rebounding ability, turnover rate, and ability to get to the free throw line. Defense was also significantly correlated to Duke's own offensive ability that night, especially on Duke's offensive glass. It should come as no surprise, then, that Duke was knocked off in the Sweet 16 in a game to a Michigan State team that was among the best in offensive rebounding, among other things.

#3 (127.0) January 19, 2010 at NC State (Duke L 88-74). Ah, the inspiration for this post. Duke was outplayed on their defensive end in a big way, and NC State was helped by a healthy 62.7 effective field goal percentage. They also protected the ball, turning it over on just 13% of their possessions, and turnover rate seems to be a relative strength of this year's Wolfpack team. Offensive efficiency, however, has not exactly been consistent for NC State, and while they put up similar numbers against Georgia Southern and UNC-G, an effort like this against a top-20 defense was completely unexpected. I see no reason to see the game as anything other than a fluke for Duke, and while winning at Littlejohn tomorrow will be a tough task, if Duke loses it will not be due to another defensive calamity. For this, in K, I trust.

#2 (127.6) February 11 2009 vs North Carolina (Duke L 102-87). Easily the worst Duke game I have ever attended, as Duke was never really in this game. Unlike some of the other games, this was not a defensive lapse against an otherwise mediocre offense. This was a lashing at the hands of the eventual champs who could not stop two All-Americans from doing whatever they want on the court. Yes, I'm a little bitter that I slept outside in a tent to watch this game. Next.

#1 (131.6) February 2 2005 at Wake Forest (Duke L 92-89). So many surprising things about this list, including the number of home games (4) and the number of Duke wins (4) and the number of appearances by the Demon Deacons (4). There is quite a wide gap between #2 and this game, and Duke almost pulled off the win! This was a collision of elite squads, as Duke finished the 04-05 season with the best defense in the nation, while Chris Paul's team was the #2 offense. Wake's inability to stop anyone came back to bite them in March, as they couldn't shoot their way out of early exits in both the ACC and NCAA tournaments.

To be fair, Wake Forest dominated the the middle 36 minutes of the game and it was only the 3-point abilities of the talented Mr. Redick that brought the game close at the end: JJ hit three from beyond the arc and Sean Dockery made it 90-89 with two seconds left. Taron Downey hit both of his free throws and Redick missed a running three to end the game. Notably, Duke's offensive efficiency in this game was 127.3, and so one must conclude that it was indeed Duke's defense that prevented the win. As the rematch (see #4) showed, home court has a lot to do with the outcome of games like these.

The biggest absence from this top 10 list: games in March. For all of the hate heaped upon Duke for its struggles in the month of March, not once did they make an NCAA tournament exit due to a sudden lack of defense. In fact, the worst NCAA tournament performance is (predictably) the 2007 first-round loss to VCU, which is 27th on the list at 112.0, followed by last year's win over Texas (36th at 109.3). But those games are miles from the ones above, and allowing around 110 points/100 possessions is probably to be expected from some of the elite offenses in the nation. Coach K gets his teams prepared on the defensive end for March, combined with playing non-ACC teams unfamiliar with Duke's defensive style. There are relatively more games in ACC tournament on the list.

I have compiled the stats from kenpom.com from all years into one spreadsheet, which can be found here for your sorting pleasure. Tempo-free stats can tell us a lot about a team that we weren't expecting to hear. This would be not-so-subtle foreshadowing of a future post about Duke. For a further hint, see this and this.

Remember When...

...Duke had a good basketball team? Those were the days:



Stats taken from Ken Pomeroy's site, which are tempo-free statistics. Scroll over to the right for the defense stats. Efficiency is measured by Points/Possession x 100, and the other percentages are also per possession. The rankings are through games on February 15, showing the clear difference between teams Duke played in the calendar year 2008 compared with those played in 2009 (The ACC season plus Davidson and Georgetown).

Duke has also played about 10 percent less efficiently on both ends of the court in the last thirteen games. Depending on how much schedule played a part, this is either terrible or perhaps expected. Still, here is the defensive efficiency rating turned in by Duke in the last six games, along with that team's Pomeroy offensive rating (in parentheses):

@ Wake Forest (51): 92
vs Virginia (126): 77.8
@ Clemson (15): 118.3
vs Miami (26): 97.4
vs North Carolina (1): 127.6
@ Boston College (22): 116.2

Perhaps this performance is no accident; four of the five best offenses Duke has faced this season have been in the last four games! The only other defensive efforts worse than 1.00 points/possession came against Georgetown (19th ranked offense), at Michigan (66), and against Rhode Island (31). While it may be tempting to cut Duke some slack because of the level of offensive play they've been against, getting far in the NCAA tourney by hoping to play crappy offenses is not exactly a winning strategy.

Breaking it down, the component statistics that most closely parallel Duke's declining efficiency is rebounding. Duke is not a horrible rebounding team on the offensive glass, grabbing about 40% of their own missed shots, good for 15th in the nation. On the defensive glass, however, things are a lot less pretty: 31.8% of Duke's opponent's missed shots are grabbed by the other team (128th nationally). This particular statistic has gotten much worse in ACC play, where opponents are grabbing 35% of their missed shots.

I have not said much about the offense because it hasn't been as much of a problem in 2009, games versus Clemson and Wake Forest excepted. North Carolina is frequently seen to have a soft defense but it still ranks 14th in the nation, and the fourth best defense Duke has seen (behind Purdue, Wake Forest, and Florida State). Duke ran out a 109.9 offensive efficiency against UNC at Cameron, which is not too shabby-- it was the 127.6 stinker on defense (worst of the season for Duke) that was the deciding factor in that game. While Duke is indeed limited by some of the best defenses, such as at Wake Forest (89.3) and Florida State (98.0), it is not as stark a difference as on defense. What Duke's offense is not, however, is an offense that can carry a poor defensive performance to victory against top teams.

Curiously, none of the main component statistics stand out for Duke, which has the fifth best defense in the nation as ranked by Pomeroy. The only stat that kind of stands out is that Duke makes a steal on 12.9% of possessions, which ranks 19th nationally. The other stats, which Pomeroy calls the "four factors" are above average but none seem to scream "This why Duke has the fifth ranked defense." Indeed, just going by raw statistics Duke is ranked 29th, which seems a lot more in line with the components. Pomeroy weights each performance by the strength of schedule, which is how Duke ends up fifth. But if Duke plays exceptionally against poor offenses and poorly against exceptional offenses, weighting by schedule is going to skew the actual ability of Duke's defense.

The conclusion I am forced to draw is that Duke's high defensive efficiency was inflated by a mediocre non-conference schedule (ranked 91st by Pomeroy), and is incapable of playing at an elite level against top-25 offenses. Last Tuesday night, before the Duke-Carolina game, Mike Krzyzewski spoke to the Cameron Crazies. One thing he said stuck with me: "You guys don't need to chant 'Let's Get Hungry.' This team is hungry, believe me they are hungry." And so when I watched the dismantling of Duke's defense by UNC and by Boston College four days later, I couldn't help but think of this South Park clip:



If Duke is interested in keeping their Pomeroy ranking high, may I humbly suggest abandoning the Atlantic Coast Conference, with a move to the less elite A-10. Or at least the Big Ten. In the meantime, I'm looking forward to the Annual Duke Sweet Sixteen Bowout. Because at this point in the season, even that's a reach for the Blue Devils.