While in the line to get into the Duke-Carolina game, I had a conversation with a group of grad students about Duke’s chances to win the NCAA Tournament. Since I’m a sports columnist for The Chronicle (and Coach K had just called me out in front the entire student body), they asked me how I thought Duke would do in the postseason. When I told them that I’d give Duke a 40 percent chance to reach the Final Four and a 10 percent chance to win the entire thing, they told me I was crazy.
But it’s not like they’re alone. Almost everyone I’ve talked to has a lower opinion of this year’s Duke team than I do. The commentators on ESPN and CBS praise the Blue Devils only begrudgingly. No one thinks they’re on the same level as Kentucky, Kansas or Syracuse. When I watch the Blue Devils play, I don’t even believe my own positive assessment of the team, if that makes sense. They just don’t pass the “eye test.”
But I believe in data. I believe that a computer can tell me things about college basketball that I can’t see with my subjective eyes. I believe that subjective eyes see what they want to see, instead of seeing reality. And so I believe the mathematical models of Ken Pomeroy and Jeff Sagarin when they tell me that Duke is the best team in the country and the favorite to take home the national championship trophy.
For centuries, philosophers like Thucydides, Francis Bacon and Leo Tolstoy have understood that seeing isn’t necessarily believing. Studies by cognitive psychologists in the 20th century obectively identified the notion of confirmation bias—that humans have a tendency to believe only information that confirms their pre-existing theories about the way the world works.
In college basketball, we tend to believe that the best teams are always hyper-athletic and chock-full of future NBA All-Stars. We believe that the best teams run up and down the court and swat opponents’ shots into the seventh row. We believe that the best teams always have five players on the court that can score, and that they can always dump the ball into the post to get an easy bucket in the half court.
And then we look at this year’s Duke team—without superior athletes, without a ton of NBA talent, without a fast-paced slam-dunking style, without more than three threats to score on most possessions, without a dominant post player—and we say that there’s no way they can be the best.
So when we see the Blue Devils get blown out by Georgetown and lose to N.C. State, that confirms our opinion and we believe the Duke team that showed up for those games is “real.” We ignore blowout wins over Gonzaga, Florida State, Wake Forest, Clemson, Georgia Tech and Maryland because dominant wins don’t support our idea that the Blue Devils are not a dominant team.
So instead of focusing on Duke’s strengths—that it has three players that shoot a high percentage and almost never turn the ball over, that Brian Zoubek is the best offensive rebounder in the country, that opponents almost never make a 3-pointer against Duke’s defense, that they score more points per possession than any other team—we focus on the weaknesses. The Blue Devils don’t have a dominant athlete. They’re too reliant on three scorers. They don’t have enough of a post presence on offense. They’re susceptible to aggressive driving guards.
And maybe that’s all true, but it hasn’t mattered this season, because the goal of basketball is not to get the most dunks or block the most shots or run the fastest fast break. Instead, the goal is to score more points than your opponent over the same number of possessions. And in scoring points and preventing their opponents from doing so, the Blue Devils’ strengths more than make up for their weaknesses.
On offense, Duke does rely heavily on three players. Kyle Singler, Jon Scheyer and Nolan Smith take just about 75 percent of Duke’s shots. But any team would kill to have them because they never turn the ball over and average just about one point per shot attempt. It makes sense to let your best offensive players shoot the most.
And while Zoubek isn’t the flashiest big man, he’s among the most effective at creating scoring opportunities for his higher-scoring teammates. Nearly a quarter of Duke’s missed shots end up in Zoubek’s hands as offensive rebounds. A full 40 percent of Duke’s missed shots turn into offensive rebounds for someone on the team. It seems obvious, but an offensive rebound is hugely valuable; it gives Singler, Scheyer and Smith another opportunity to shoot.
A typical Duke possession might go like this: Smith shoots a mid-range jumper, which rims out. Zoubek awkwardly controls the rebound and passes back out to Scheyer. Four passes later, Smith shoots another mid-range jumper, which goes in. The end result is two points. Meanwhile, a typical Kentucky possession might start with a steal at halfcourt by John Wall and end in a reverse alley-oop by Eric Bledsoe. The end result is still two points, but any observer would probably peg Kentucky as the better offensive team.
But the computer doesn’t pick and choose possessions to analyze. The computer doesn’t use an eye test. The computer sees one possession for Duke and one possession for Kentucky, and two points for each team. And the computer sees that over the course of the season, Duke scores more points per possession than anyone else, largely because grabbing offensive rebounds and avoiding turnovers allow the Blue Devils to take so many shots. Duke has shot the ball 2,045 times, seventh-most in the country, despite only playing at the 158th-fastest pace.
The same thing goes for defense. The subjective observer sees the blocked shots of Kansas’s Cole Aldrich and the steals of Syracuse’s Andy Rautins. Then he watches Duke get abused by backdoor cuts and dribble penetration by Georgetown and N.C. State. But the computer sees only possessions and points allowed over the course of the whole season, and it knows that Duke is ranked fourth in the country in points allowed per possession because the Blue Devils prevent their opponents from taking and making 3-pointers.
It doesn’t conform to our preconceived notions of what makes a college basketball team great, but when you remove those preconceived notions, it becomes obvious that Duke’s style has worked all season long. There’s no reason to believe that all of a sudden it’ll stop working in the Tournament.
So don’t believe your eyes. Do believe the numbers. There’s no guarantee that Duke will win the national championship—anything can happen in a single elimination tournament—but the data says the Blue Devils should be the favorites.
Get The Chronicle straight to your inbox
Signup for our weekly newsletter. Cancel at any time.