Last week, while perusing some Duke Basketball forums, I came across a sentence that I’m reasonably confident I had never read before on a college hoops message board. “What we really need here is a regression of individual efficiency, regressed against number of games into the ACC season and weighted by minutes played,” a user who goes by CrazieDUMB wrote on Duke Basketball Report. “Any ambitious Dukies still have access to STATA or SAS and want to tackle that one?”
Now, I’m not a statistics major, and I couldn’t distinguish STATA from Strat-O-Matic, but the response caught my attention. The comment was part of a three-page thread reacting to a column from Alex Fanaroff in this very space last Wednesday, in which he graphed data and even rolled out some scatterplots to explore whether there was proof of a downward trend in the latter half of Duke’s last seven seasons. The reaction to the column was robust. Fanaroff received letters far and wide—some with more data still—and ESPN and Basketball Prospectus, among others, examined the evidence in greater detail.
The subsequent discourse was generally one of reason and civility, not irrationality and vitriol, which was sort of stunning in itself. But it was the mere fact that a numbers-heavy column with an analytic bent provoked such a dialogue, regardless of its nature, that took me aback.
Before last week, I didn’t understand much about tempo-free statistics. I knew they existed and that it was generally naive for me to ignore them, but I also understood that I was intimidated by all those fancy numbers on Ken Pomeroy’s Web site. I decided that if DBR was capable of an intelligent, mild-mannered give-and-take about tempo-free statistics, then it was time for me to ditch my ignorance.
So I brushed up on the logic behind examining statistics adjusted to possessions. I read about the four factors of offensive and defensive efficiency—effective shooting percentage, offensive rebounding percentage, turnover rate and free throw rate—and I devoured blog posts about points per possession and pace. I learned that most of this thought dates back to the 1950s, when Dean Smith charted possessions and kept track of pace at that school down the road. Just for good measure, I checked around to confirm that Smith’s old squad still does have a 3-9 ACC record.
I called Basketball Prospectus’ John Gasaway, who indulged my naivete, and directed my unexpected education in further directions. “We’re not being terribly advanced or strange,” said Gasaway, who, just last week, offered his free analysis to the first mid-major team that contacted him and makes the NCAA Tournament. “We’re just trying to get to a point where other major sports have long been and taken for granted, and that is being able to point to a team and say, empirically, that they have a good offense or a good defense. It’s strange that in basketball—not just college, but basketball period—that we haven’t been able to do that, widely, as fans until the past few years.”
Wouldn’t it make sense, I asked him, for the general community of Duke fans to be especially progressive about espousing these types of statistics? For the most part, it’s a group of smart, educated diehards and, perhaps more importantly, the basketball team is consistently dominant. There is never a dearth of interest. We’re always looking for new ways to analyze, new data to pore over, new ways to win longstanding debates.
“I definitely hear a lot from that region of North Carolina,” Gasaway said. “I think the key might be smart and impassioned fan bases, even more than having a successful program, because I would also point to a community like Michigan, which has had next to nothing in the way of recent success.”
Still, I’m not sure Duke fans have reacted strongly to advanced statistical analysis one way or the other. It’s not every day we can walk through the Krzyzewskiville line and hear about Duke’s offensive efficiency. Mostly, we just talk about who looks good—who, to our naked and fairly uninformed eyes, seems to be playing well. Pomeroy’s ratings system, by contrast, ranks teams by numbers, not by feel. (His poll, driven by offensive and defensive efficiency, is the only one in the country that currently ranks Duke No. 1 in the country.) He, too, has the relatively rapt attention of Triangle hoops fans. In December, on his blog, he unveiled the analytics of his Web site, revealing that the most-trafficked team pages were UNC’s and Duke’s, and North Carolina sent the fourth-most traffic of any state—when adjusted for scale of population, naturally.
Matt Johnson, a graduate student in biology, is a Pomeroy reader who blogs about tempo-free statistics, specifically applied to Duke. He’s been blogging since January 2007, and while he does touch on other sports, Duke Basketball is his primary muse. Last year, he created an NCAA Tournament simulator, using expected winning percentages to predict individual outcomes; he’s been working on tweaking it for optimal performance ever since.
“I’m an empirical biologist, so looking at any kind of data is kind of exciting to me,” Johnson said. “To be able to use statistics and use the stuff I’ve learned from biology and apply it to being a fan—it brings together multiple facets of being a nerd.”
Of course, tempo-free statistics are most useful not for analyzing the team from the couch, but for actually making adjustments on the bench. Anyone can blog about the four factors; only a relative few can do something with it. Which is why I got in touch with Chris Collins, Duke’s associate head coach, to ask whether Duke’s staff was a member of the growing ranks that swear by statistical analysis.
Get Overtime, all Duke athletics
Signup for our editorially curated, weekly newsletter. Cancel at any time.
“The numbers we use a lot are turnovers and offensive rebounds,” Collins said, noting that he relies on statistics, especially those from the last five games, when he scouts opponents. “The other key is we try to get ourselves to the free-throw line. Those are probably the main ones that we look at—and obviously, well, shooting the ball.”
Turnovers, offensive rebounds, free throws and shooting percentage. Sounds like it's worth our attention, after all.