The RTC Interview Series: One on One With Ken PomeroyPosted by Brian Goodman on July 8th, 2011
Rush The Court is back with another edition of One on One: An Interview Series, which we will bring you periodically throughout the year. If you have any specific interview requests or want us to interview you, shoot us an email at email@example.com.
One of the great things about college hoops is that with nearly 350 Division I teams, you can find any and every playing style under the sun. Some teams push the ball at frenetic paces in an effort to wear down the opposition, while others prefer to slow down and make every possession count. Smaller teams rely on outside shooting and bigger teams assert their dominance down low, so in comparing teams to one another, how do you account for such widely varying styles of play? This is the question that cult hero and statistician Ken Pomeroy longs to answer. The solution isn’t always simple, but it boils down to evaluating how teams fare on a possession-to-possession basis, rather than using the commonly-held method of measuring events from game to game. At KenPom.com, fans can track the performance of all 345 Division I teams in his tempo-free style. Over the last few years, his approach has moved from the underground into the mainstream, mentioned by media outlets such as ESPN.com and The Wall Street Journal during the college hoops season. He is also a regular contributor to Basketball Prospectus and you can follow him on Twitter (@kenpomeroy). In this interview, Ken took a few minutes to talk with us about his methodology and the growth of his website.
RTC: For our readers who are used to the more traditional “counting stats,” what makes your analysis different and worthwhile?
Ken Pomeroy: In all of the statistics I use, I’m trying to equalize opportunities. If you’re going to compare one offense to another, it’s not fair to look at raw points. North Carolina, for instance, has more opportunities to score (than an average team). It’s also not fair to compare defenses for the same reason. We look at rebounding percentage, for instance, which takes into account how many rebounds are available to a player when he’s on the floor to get an appreciation for whether he’s a good rebounder or not. Those are the things I try to do with all these stats.
RTC: One thing that makes your analysis easy to digest is that most of the teams that excel in traditional stats and occupy the top spots in the rankings also excel according to your tempo-free analysis. There are some exceptions, though – what are some schools in recent years that may not have had those alluring traditional stats but were more eye-catching in your analysis?
KP: Wisconsin played at the second-slowest pace in the nation (ed. note – 58 possessions per game, compared to the national average of about 67), but had a very effective offense. They weren’t effective all of the time — they obviously had that ugly game against Penn State in the Big Ten Tournament which called into question just how good their offense was. Georgetown in their Final Four run in 2007 had an outstanding offense, but played a very slow pace. North Carolina’s 2005 championship team was criticized for their defense based on the points they allowed, but tempo-free, their defense was one of the best in the country. If you win 90-75, it looks like you gave up too many points, but when you factor in that the game was 80 possessions, it reveals a better defensive performance.
RTC: A number of coaches have openly embraced tempo-free analysis, including Buzz Williams, Brad Stevens and Mike Krzyzewski, and you’ve recently added a coaches’ section to your site where fans can look at performance. Have you seen some coaches improve in a way that can be attributed to some acknowledgement of embracing the points you emphasize?
KP: For the most part, teams play the style that they have to with the players they recruit. If you’re a team that’s in the basement of your conference, you’re going to have to do things differently. You look at John Beilein, he has a very unconventional style that he took from Canisius to West Virginia and on to Michigan; he’s made adjustments to his style. Coach K’s teams vary, but there are some constants, like they seem to defend the three-point line well every year, but there are other things that fluctuate, like offensive rebounding and pace. He really adapts his system, especially offensively, to the personnel he has.
RTC: Some followers liken your analysis to Bill James’ statistical advancements in baseball in that you’re both trying to find what makes teams successful and how they identify undervalued skill sets and use them to gain advantages. Does that comparison work?
KP: In basketball, it’s much more difficult to figure out what those undervalued things are, because those things are driven by your personnel. I think offensive rebounding is generally undervalued, but if you’re a smaller team like Murray State, you have to rely more on the three-pointer because you can’t crash the boards as much. Belmont is another example of a team that values the three-pointer because they can’t bang with the big boys. I think each coach has to judge how they can best play to their strengths. Bill James made waves in baseball circles by advocating for on-base percentage over batting average, which has become universal, but there isn’t that kind of stat in basketball.
RTC: In your rankings, one element you use is a factor for luck. When we watch games, there are a number of variables: Teams will make shots that have no business going in, some teams will have uncharacteristically bad shooting nights, and officiating plays a role as well. How do you account for those variables and how much of a team’s performance in a given season can be attributed to luck?
RTC: How did you get into developing these tempo-based metrics?
KP: I’m a certified stats geek, and I don’t shy away from that. My two main sports growing up were baseball and college basketball. There’s tons of stats in baseball and a bunch of people have followed from what Bill James did in the 1980s. I discovered in the late 90s that nobody was doing that kind of analysis in college basketball, so I started a blog and wrote about it. Everyone was using points per game and minutes per game, and from that point in 2002, I started developing my stats. The site has grown little by little every year, pretty much through word of mouth. For two or three years, it was this underground hobby with a fervent fanbase, but it wasn’t very big. I’m really surprised at how it’s caught on and that some commentators actually talk about it during games.
Rush The Court: What are your plans for the site this year?
KP: Ideas just come when I least expect them to. I’d like to venture more into the graphical realm, which I started last season, instead of just giving numbers. I’m looking to refine that and I’m planning to do some preseason win projections, which I’ll start up in the next month or two.
Rush The Court: Thanks for your time, Ken! We’re looking forward to next season.