ShockerHoops.net sat down with Erik Haslam, the brain behind Haslametrics.com. Erik has rated Wichita State highly in his algorithm all season, and he explains how the Shockers achieved that rating and what he expects against Arizona in the first round of the NCAA Tournament. Without further adieu, our interview with Erik.
ShockerHoops.net: Could you give our readers a rundown of what Haslametrics is all about?
Erik Haslam: Haslametrics.com is a website I started up a couple of years ago to rank the true strengths of D1 college basketball teams and forecast the results of future games. For the last quarter of a century, I have been a March Madness junkie. I’ve always had a love for the NCAA Tournament and have been a huge fan of predicting its outcome since back when I was in high school in the early 1990s. About five or six years ago, I grew tired of making selections based on my own “eye tests” and wanted more information at my fingertips to determine which teams were truly superior. I decided to begin mining data from various locations, piecing it together in my own unique, constructive manner, and kicking out results in an attempt to create a “one-size-fits-all” algorithm that I can apply to each and every team in college hoops. The process has evolved over many years of work – it actually began within a simple Microsoft Access database – but things have now obviously progressed. My analysis now primarily focuses on shot selection – how often teams shoot, how close to the hoop each shot is, how well teams shoot from different spots on the floor, and how often steals and offensive rebounds affect shot selection and shooting percentage. The whole process has now reached a point where I’ve become very satisfied with the results my algorithms deliver. Those results are what you see updated on my website every morning of the college basketball season.
SH: What separates your calculations from other metric sites such as KenPom?
Erik: I can’t speak to the methodology each basketball analytics site utilizes because they vary so much. One of the reasons I love analytics is because everyone’s work is totally unique. As I typically say, analytics are like wine tasting. There’s no 100% right or 100% wrong answer. It’s whatever works for you.
That being said, what makes my site unique is that it is, to my knowledge, the only college basketball analytics site that uses play-by-play data instead of box-score data. The code that I have authored over the years parses play-by-play data for each contest and organizes everything neatly on a possession-by-possession basis. I do this for two reasons. First, play-by-play data gives me information about the locations on the floor from where teams shoot. Box-score data provides three-point shooting data, but it doesn’t give you anything pertaining to mid-range shooting and near-proximity shooting. That data can be sufficiently extracted from play-by-play data. Second, play-by-play data allows me to determine when a game is mathematically out of reach and then truncate any data after that point. Once a game has been mathematically decided, coaches typically go to their bench and give the “scrubs” some floor time, thereby degrading the performance of the team in that game. By truncating this data, I can obtain a better feel for the team’s true performance at full-strength.
My site also focuses on the “how,” not just the “who.” In other words, many sites largely discount games where powerhouses score wins over lesser teams. My algorithms focus on how teams win those games. Did North Carolina beat Central Connecticut State by four points or 54 points? How efficient were they on each possession when the outcome was still in doubt? It’s not just about who they play….it’s HOW they play against that opponent and HOW they fared as compared to other teams that played against that opponent. That’s precisely how you form my spider-web of comparisons amongst all 351 D1 teams.
SH: Wichita State has been highly rated in your rankings nearly all year, what about the Shockers makes them deserve your classification as “one of the very elite teams in college basketball.”
Erik: Wichita State is highly rated for a couple of reasons. First is the aforementioned “how.” When you look at Wichita State’s results, you’ll see they opened up large leads very quickly against several opponents. Many of their wins, especially after the start of 2016, were dominant performances. Starting with the Nevada win on December 22nd, Wichita State won 12 straight contests with an average margin of victory of 21.1 points per game. That’s impressive regardless of which D1 opponents you’re playing. Compared to those teams that had common opponents with Wichita State, the Shockers really outclassed a great majority of them. Wichita State did far more damage far more quickly than other teams did playing the same opponent.
The other reason that Wichita State rates so well is because my algorithms greatly de-weight games near the start of the season, almost to the point that games on the first day of the season are rendered weightless. I’ve learned that college basketball teams grow by leaps and bounds over the span of five months. I’m more interested about who Wichita State is today, not who Wichita State was back on November 13th.
SH: What are your thoughts on this year’s committee and their seeding of the bracket?
Erik: My personal opinion is that the committee needs to reevaluate both its selection process and its seeding. There were more than a few head-scratchers during the bracket reveal this year. Wichita State is a good example. There is no way that they have played at the level of an 11-seed. They have easily performed above the 11-line from my point of view. In a way, the draw is very unfair to Arizona, who has to compete against a far more difficult opponent in the first round than other six-seeds do. I don’t necessarily have all the answers to solve the committee’s problems, but I think we’ve reached a point where a discussion needs to be had to address them.
SH: Who would you have left out that made the cut, and who would you have put in that didn’t make the cut?
Erik: I had my reservations about Temple. They really scored poorly in the “how” and ranked 87th in the country at Haslametrics.com on Selection Sunday. That being said, they did win a share of their regular season conference title, so I don’t take much issue with their inclusion. I think the team that most analysts agree should not have been in there is Tulsa. We had them squarely on the bubble right before they were absolutely trounced by Memphis in the AAC Tournament. That was the final nail in the casket for them in my opinion. Obviously, the selection committee thought differently.
As for inclusion, I would have gone with St. Mary’s. You can make the argument that they didn’t play anyone, but they did score two wins over a very good Gonzaga team. Additionally, they lost a grand total of five games (out of 33) all year. The WCC may not be the ACC, but losing five games all season while competing in a conference like the WCC should be impressive enough to score a bid. In our power rankings, St. Mary’s landed at #31 on Selection Sunday. They had the numbers equivalent to a team that would earn a six-seed. Instead, the committee excluded them altogether.
SH: Where would you have seeded Wichita State and Arizona?
Erik: I thought the combination of Wichita State’s resume – which obviously could have been better from a win-loss/strength-of-schedule standpoint – with its overall strength warranted an eight-seed. I had them matching up with nine-seed Cincinnati in the first round, with the winner likely getting Oregon in the second round. I also had Arizona seeded higher. I placed them as a five-seed on Selection Sunday, a slot that would have set up a contest with 12-seed South Dakota State in Round One.
SH: What are your thoughts on Wichita State vs Arizona on Thursday?
Erik: I’ve been a huge fan of Wichita State’s defensive effort this year, and they really brought it against Vanderbilt in the play-in game. Going into that contest, Wichita State was ranked third in the country in defensive efficiency and was facing an opponent ranked 20th nationally in offensive efficiency. The Shockers held that offense to just 30% shooting and 50 total points on the night. If that’s not impressive, I don’t know what is. After that outing, Wichita State, not surprisingly, is now rated #1 in the country in defensive efficiency at Haslametrics.com.
The Shockers now face a team with an offense even more capable than what Vanderbilt’s was. Arizona ranks 12th in the country in offensive efficiency, and the Wildcats land in the top-50 nationally in all offensive shooting percentage categories. They are 14th in field goal percentage, ninth in three-point field goal percentage, 43rd in near-proximity field goal percentage, and 49th in free-throw percentage.
That being said, Wichita State is a team with a chip on its shoulder, since so many experts did not believe they belonged in the dance. Combine that with the team’s vast senior leadership and the Shockers should be a very tough out for any foe this postseason.
SH: What is your prediction for Wichita State vs. Arizona tonight?
Erik: My calculations tell me that Wichita State has the advantage despite the lesser seed. Using the All-Play Estimates from this morning at Haslametrics.com – where you can simulate the outcome of any of the 61,425 possible D1 matchups – we are currently forecasting a 68-66 victory for the Shockers over Arizona in Round One.