An interview with the creator of Basketball-Reference.
Just like everything else in this (sometimes) over-dissected World Wide Web, basketball statistics have taken on a new life of their own. No longer is the casual fans’ knowledge of random tidbits regarding free throw percentages and offensive rebounds considered unusual, instead those attention-paying folks are fixated on Plus/Minus, Player Efficiency Ratings, and True Shooting Percentages.
Thankfully, to feed our crazy and insatiable appetite for the miscellaneous, Justin Kubatko, creator of Basketball-Reference.com, realized there were mouths to feed and acted accordingly. Kubatko is at the helm of the foremost hoops stats/history site, and his work and compilations are used by probably 98.7 percent of all basketball writers. Frankly, some of the features on his site make it cool to be nerdy: Boxscores dating back to 1986-87, player pages packed with lifetime stats, player salaries and elementary school game logs (I made that last one up but don’t be shocked), and you can’t forget the Hall of Fame probability, a formula Kubatko created himself that attempts to decree who is worthy of a Call to the Hall. To say those are the tip of the Basketball-Reference iceberg is a gross understatement—the site offers, for lack of a more exciting term, everything.
SLAM recently had an opportunity to catch up with Kubatko to chat about BBall-Ref and get an insight into the Internet’s most-used basketball statistical haven.
SLAM: You would have to be the ultimate hoops stat addict, right?
Justin Kubatko: I’m more of a sports addict rather than a hoop stats addict. Don’t get me wrong, I’m a big fan of both college and pro basketball, but I’m a big fan of college and pro football as well. I used to be a big baseball fan, but my interest has waned over the years.
SLAM: When you started BBall-Ref, did you feel the internet was lacking a site of this nature? Was there a need for a comprehensive hoop stats site?
JK: At the time I started working on Basketball-Reference.com there were a few sites on the Internet that had historical basketball statistics, but I felt that all of them were lacking in some way. My goal was not to simply match what was already out there, but to go far above and beyond what anyone else had ever done.
SLAM: Do you feel your site has played a role in the tremendous growth of basketball statistics in recent years?
JK: Probably, if for no other reason than it brought a lot of information that could only be found in the past using multiple sources and put it under one roof. I also wanted to place an emphasis on some of the more advanced statistics that are out there (I.e. Dean Oliver’s work), and I think that appealed to a small but extremely loyal group of users. The dynamic nature of the Internet also means my site is constantly changing, which is a huge plus. If you buy an encyclopedia if will be out of date before you know it, but a web site can be updated almost instantly.
SLAM: Does your site have any NBA affiliation and do NBA teams consult with you about any content on the site?
JK: My site is not affiliated with the NBA, and NBA teams do not consult with me regarding the content on the site. I am a statistical consultant for an NBA team, although my client would prefer that I not mention them by name.
SLAM: The Hall of Fame probability is a neat attraction on the site—often drawing citations from ESPN, NBA TV, etc.—what is the basis of that feature?
JK: I was just fooling around one day trying to see if I could build a model that would predict the Hall of Fame status for any eligible player with at least 400 games played. It turns out that I came up with something that works rather well. Of the 78 players in the Hall of Fame, 63 were correctly classified (80.8 percent) and 15 were not (19.2 percent). Of the 590 players not in the Hall of Fame, 583 were correctly classified (98.8 percent) and 7 were not (1.2 percent). Overall, 646 of the 668 players (96.7 percent) were correctly classified using the model.
SLAM: Your site has some unique features—old boxscores, head-to-head, individual game scores—is any stat beyond the reach of BBall-Ref?
JK: I guess the short answer is no, as long as it’s being tracked by someone, somewhere.
SLAM: What player page, historically or active, is the most viewed?
JK: Michael Jordan, by far; Kobe Bryant is second, but Jordan’s page gets about 50 percent more traffic.
SLAM: Where to now for BBall-Ref? What are the goals, plans for the site?
JK: There are two things that users constantly ask for: player transactions and more box scores from past seasons. We are doing our best to try to address those requests, but as you can imagine finding box scores in a usable format for seasons prior to the late 1980s is difficult. I think we can get something done, though. In the near future, I’m hoping to have a playoff probability report ready sometime after the All-Star break. Basically what I’ll do is simulate the remainder of the season a large number of times and report how often a team made the playoffs, won their division, etc.
SLAM: Last thing, what is your all-time favorite stat? There would have to be one thing, in all your research and compiling, that appeals to you most.
JK: I’m not a big trivia guy, so I don’t really have one. Here’s something I’ve always found odd and fascinating, though: Wilt Chamberlain played in 1205 career games (regular season and playoffs) and, despite playing almost every minute of every game, he never fouled out.