It was February 2020 — and the source of Hall-of-Fame Syracuse Coach Jim Boeheim’s ire wasn’t an opposing player, fans or a referee.

It was Salt Lake City-based statistics guru Ken Pomeroy, who for years had run a popular college basketball analytics website.

“I don’t know where this guy… he’s making a lot of money, KenPom,” Boeheim railed. “I’m telling you right now: No one in this room, nobody that’s doing KenPom knows who’s at fault when somebody scores on us.”

Pomeroy, of course, hadn’t actually suggested this, but his nuanced stats on KenPom.com have been shaping key coaching decisions across the sport.

The Daily Coach caught up with Pomeroy recently to discuss his unconventional career path, effective vs. ineffective use of data, and the luck element of any performance.

This interview has been condensed and edited for brevity.

Ken, thanks a lot for doing this. Tell us a little about your childhood and how it shaped you.

I grew up in Northern Virginia. That’s really where my interest in basketball developed. It was the late 80s, early 90s in the middle of Big East/ACC basketball country in a market where the professional team wasn’t very good.

My dad worked for the government. My mom had a variety of jobs and took care of me and my sister, then joined my dad as an administrative assistant. I wasn’t very good at sports. I played baseball and basketball, but my career probably peaked in about the fifth grade. I quickly realized I wasn’t going to be involved in basketball at a high level. My genes directed me in a way where I was into math and numbers, not so much reading Shakespeare.

You go to Virginia Tech and major in civil engineering. What drew you there and can you take us through your career path after?

Engineering seemed like a natural fit for me based on my interest in math, and Virginia Tech was the best engineering school in the state. I wasn’t involved with the basketball team at all, didn’t even really know how to be. I was interested in stats and knew player averages, but I had no depth of knowledge in terms of watching a game.

After I graduated, I took a job designing roads in suburban Baltimore for a couple of years. It was a great job that I was fortunate to have, but what I learned was the more I worked, the better chance I had for a promotion, except the promotions were less for engineering and more writing contract proposals. The career path really turned me off. The more work I did, the less interesting it was.

I ended up going to graduate school at the University of Wyoming to pursue meteorology. I was really interested in weather and embarked on a second career at the National Weather Service. At Wyoming, I was still watching college basketball games as a casual fan, but I started working on sports ratings, too. There was some really insightful writing on the internet about baseball analytics, and I thought there had to be some good stuff about college basketball, but there really wasn’t. I’d watch games and start to think about the sport a little more analytically like, “There’s got to be something there.”

Were you sharing your data with friends or family or when did you realize you might really be onto something?

I wasn’t really sharing it with anybody. I didn’t think there were people who would care about this stuff in my inner circle. There were little steps along the way, but the big problem was how do I corral all this data and process it because I had done really poorly in my computer programming classes in college? There were some little moments along the way, and in the early part of it, media members started referencing my work.

When he was at Sports Illustrated, the late Grant Wahl was into my stuff and gave me my first national mention that these numbers are really good and provide insight. Then, it spread to coaches. The big moment was 2010, Butler made a run to the National Championship and Brad Stevens mentioned after one game that he used my work. He said after they’d win, he’d go to my website and look up the next team they were playing to get a high-level picture of what that team did well and what they didn’t. That was a huge moment for me.

You were balancing work between the National Weather Service and keeping basketball analytics. What advice would you give someone who has a serious passion but isn’t sure he/she can turn it into a career?

I really liked the weather service job. Eventually, (the basketball stats) just got to be so popular and successful that it was pretty obvious I could make a go of it. I was fortunate I had a boss at the weather service who really let me try it out for a year and gave me a friendlier work schedule to test it.

For me personally, I wish I had had more confidence in what I was doing. I didn’t know I could run a business or a website. I could’ve staked this out a couple of years before I did, but for people out there who are looking for their path, it took me three tries. I’m on my third career now. Looking back on it, it was unconventional, but I wouldn’t worry about whether things are conventional or not. If you believe in something, are good at it and have the drive to do it, take a swing at doing it. If you succeed, it’s an incredibly fulfilling experience.

You keep this interesting stat on your website called “Luck Rating.” Can you explain what that is and maybe the larger implications beyond college basketball of a luck element in performance?

The luck rating is something I borrowed from (revered basketball statistician) Dean Oliver. It looks at a team’s performance in close games, the idea being that a team doesn’t have total skill in winning a one-possession game. There’s a lot that’s out of a team’s control. It really gets people riled up. In the long run, if you have 10 one-possession games in a season, even if you’re a great team, you’d be expected to win six or seven of those. Some teams win 10, though, because they get some bounces or some calls, so the luck stat measures how well a team performs in close games.

It just gets back to the idea that that’s what we’re trying to do in the analytics field: Find the signal, find what’s meaningful, find what teams are skilled at, then find what’s luck and not measure that.

I think that’s a core principle of coaching or any life activity as well: You have to control what you can control and not really worry about what you can’t control. I think the next logical step to that is understanding “What can I control?” and working on that. If the process is good, it should lead to good results. It won’t do that every time, but if you’re the better team and doing the things that should lead to wins, you should win a lot of games.

With all the numbers you keep, is there anything you’d caution an audience about with your data or advise someone to be aware of with luck factors or general outliers?

You definitely don’t want to take all data at face value. People are always trying to sell that their numbers mean this or that. A lot of times, you can be led astray. We need a pretty large sample for data to mean something. If a player goes 0-for-3 from the free-throw line, it doesn’t mean that player is never going to make a free throw again or is even a bad free-throw shooter. That principle needs to be in everyone’s mind when they’re looking at data. The work I do I feel like is pretty useful, but I do try to be honest with the audience when some of the data is fluky or not meaningful or can change.

People in the field producing data are the experts and the audience are not experts and assume the people producing the data are and that everything is honest and true. But sometimes it’s not. Sometimes you do need to be skeptical of what people are telling you, not to the extent of never believing analytics, but you don’t always want to take everything at face value.

For a coach or business executive looking to incorporate more data into a work environment, are there one or two key questions you think they need to ask in advance?

I think it starts with a curiosity. The great thing about data is it’s an impartial third party. It can give you another opinion that is theoretically free of bias and emotion, and that can be helpful if you use it wisely.

But to me, it really starts with: “What do you want to know? What aspect of your business or your team are you concerned about or want more information about?” I think once you start there, everything follows, and you can figure out what matters. I said it’s important to be skeptical of data, but it’s also important to not just dismiss data as out of hand and use it as the tool it’s meant to be.

You’ve touched on skepticism of data. When you hear someone now like a Charles Barkley knock analytics, do you roll your eyes or do you understand where he/she may be coming from?

When Charles Barkley does it, I laugh. He’s hilarious, and I really laugh at the people who get offended by what he says because he’s really on TV to entertain people, and he’s very, very good at doing that.

You see athletes criticize analytics. I think of football guys like Troy Aikman, who starred in the 80s and 90s, who don’t have much use for modern analytics. I can understand if you’re an elite athlete why you wouldn’t care about the data. You probably got to where you are without looking at it and just being a very, very gifted athlete, so I don’t take offense for the most part. It’s more some of the journalists who dismiss it.

But the results kind of speak for themselves. Analytics are embedded in pro and college teams now. I’m pretty secure about the analytical field, and I don’t take offense to too many people who dismiss it.

Q&A Resources

Ken Pomeroy ― Website | Twitter | The Athletic