As I did my final NCAA prediction ballot yesterday, for the first time since I started doing this, I had no confidence in my results. The final number of missing three teams (Michigan St, Florida Atlantic, and Virginia) doesn’t sound bad, but it’s more than that. The reality is that there are 32 automatic bids, and at-large teams in the first 8 or 9 seed lines should be automatic. That normally accounts for close to 60 teams being non-controversial. So, for someone who rarely misses more than one team, three misses are not good. Additionally, one if my misses was an 8 seed and another was a 9 seed. Even with an abnormally high number of automatic bids filling the 11-16 slots, that is surprising. The final team I missed is in the first four, but I didn’t even have them under consideration by Sunday.
The other factor that identifies the disconnect is in the seeding of the teams. In addition to the misses identified above, I originally had Gonzaga in the First Four, but bumped them to a 9 seed. They ended up (somehow) being a 5 seed. I also had Kentucky, BYU, Boise St, Nevada, and Washington State off by multiple seed lines. Lastly, from the 5-10 seedings, I only had a total of 5 teams on the exact lines. This means either I forgot how to do this after thirty years, or the committee reacted differently, so I decided to investigate.
I started by listening to multiple interviews with the selection committee chair, and to be honest, I am more confused than clear. For example, when evaluating at-large teams, the committee always makes the point that conference affiliation does not matter, and they view the totality of each team’s schedule regardless of conference. When the chair was asked about certain teams, however, he appeared to contradict that. He mentioned that Iowa State, who was thought to be a top seed candidate for some (including me), was dinged because of their non-conference schedule. This is unfair, since their overall strength of schedule was higher than top seed North Carolina and fellow two seed Arizona. He similarly noted Seton Hall’s non-conference schedule, even though they beat the best team in the country, a number 2 seed in Marquette, and won 13 Big East games. Gonzaga was also praised for scheduling well in non-conference, which is a little easier when half of your conference is equivalent to somewhere around Georgetown and DePaul. Their overall strength of schedule is roughly 60 spots worse than Providence or Seton Hall.
Another area I heard mentioned when Virginia was discussed was their strength of record, including their 10-10 record in Quad 1-2 and 8-3 record in Quad 2 games. Looking into the numbers shows a distinction between the Big East and ACC. Quad 2 homes games are those against teams ranked between 31-75 in the NET, while Quad 2 road games are between 76-135 in the NET. The problem for the Big East is they have no teams between 76-135 in the NET as 9 of the 11 teams are in the top 65, while 40% of the ACC (6 of 15 teams) are in this group. This means that Virginia gets “credit” for Quad 2 wins against teams that would never be considered for an at-large bid. Maybe there is more to this, but it appears that Virginia got more credit for a schedule with best wins of Florida and Clemson than the Friars got for beating Marquette, Creighton twice, and Wisconsin.
One last point that bothered me, but most people wouldn’t notice, is the seeding of the automatic bids. For lack of a better word, it seems like pure laziness. I have mentioned here before that I am a fan of Grand Canyon University, and they were given a 12 seed. Teams that only made the NCAAs because of upsets in their conference tournaments, such as North Carolina State, Oregon, and most egregiously Duquesne, were seeded ahead of the 29-4 Antelopes despite GCU having better metrics. It appears that teams were simply plugged into spots because of their names (NC State and Oregon), or because it was easier due to their late game start (Duquesne). This should be unacceptable. If they took shortcuts on something as simple as this, what other shortcuts were taken?
In conclusion, a member of the pcbb1917 Slack group, Adam, said that he did not think the committee would be able to go through the data they were given and would end up cutting corners in the selection process. I disagreed and noted that it really isn’t that much data, especially since 55-60 of the 68 teams are automatic or obvious. As someone who works with voluminous data, I still maintain that there is not an overwhelming amount of data for a 12-member committee that began this process in November. Given the results of this process, I apologize to Adam because I am convinced the committee cut at least some corners. The only other explanation I can fathom is that many of the committee members do not have a basketball background. While their positions at their schools or conferences involve basketball, only a few (including retiring Butler AD and former coach Barry Collier) came from basketball positions. Next year, at least one more basketball person, WAC commissioner and former Xavier basketball player Brian Thornton will join the committee. Let’s hope it helps.