Our NIT Bracketology is updated. One last time: Our apologies for how long this took. We’ll be updating it consistently from here through Selection Sunday.
The highlights are fun: Old friend North Carolina is back in our projected field. Arkansas opens against Arkansas State. SMU opens against TCU. UC San Diego opens against USC. There are a couple more intrastate matchups, we have a potential UNC vs. Georgetown tilt in the second round, and Samford, MTSU, and Akron give the tournament a little Cinderella flair.
Again, here’s the link to the NIT Bracketology itself. We didn’t update our NCAA Tournament bracket today, but here’s that Seed List. Notes on today’s brackets:
Our “Lite” Model
We aren’t running our full model yet. That should arrive sometime in the next couple weeks, with NCAA Tournament probabilities and NIT probabilities and probabilities of winning each, plus conference tournaments, etc.
In the meantime, we’re running our “lite” model. What does that mean?
Four differences:
The biggest difference is that we aren’t simulating the remainder of the season, like our full model does. Instead, we’re taking every ranking system on the team sheet—NET, KPI, BPI SOR, WAB, kenpom, BPI, and Torvik—and projecting it to slide towards where kenpom currently rates each team, with the degree of slide proportional to how much season remains for the average bubble team. So, if a team currently enjoys a great KPI rating but has a poor kenpom rating, we project KPI to slip. If a team currently has a terrible BPI SOR but is great on kenpom, we project BPI SOR to improve. The basis of our bracketology is four selection/seeding scores, each of which is based solely on team sheet rankings. This gets us a projection for each of those rankings, thereby getting us those scores.
The second-biggest difference is that we aren’t adding any of what we term “exceptions” to our NCAA Tournament Selection Score. In the full model, we look at Q1 wins, non-conference strength of schedule, how close to .500 a team’s final record is expected to land, and a few other variables. Our lite model doesn’t look at those.
The third-biggest difference is that instead of calculating out the average cut lines and each team’s probability of receiving an NIT exempt bid, an NIT automatic bid, and an NCAA Tournament automatic bid, we’re eyeballing those numbers. This is not really a piece of the “model.” This is where the model ends. We use the lite model to give us selection/seeding scores, and then we use those to fill in the blanks.
The fourth-biggest difference is that our full model will allow selection/seeding scores to wiggle, reflecting uncertainty about which variables this year’s specific committee will value. In some of the full model’s simulations, NET will be a huge deal to one committee. In most, it won’t be. That is not a feature of our lite model.
What does all of this mean? The short version is that if there’s a bubble team with an unusual résumé, there’s probably a gap between our full model and our lite model. The longer version is that we also aren’t especially confident in our automatic bid/exempt bid projections, and it’s possible we should have the NCAAT/NIT cut line one slot higher.
3-Bid League?
We have all three of VCU, George Mason, and Dayton in today’s bracket despite those being the three best A-10 teams, per kenpom. I don’t know if the A-10 Tournament field would be favored over all three of these teams combined or not. For today’s bracket, we took the field, but that was a guess. Each of these teams is aimed at the NIT.
Opt-Outs, the College Basketball Crown, Etc.
We still don’t know which teams will go to the College Basketball Crown. We understand that some sort of contracts exist binding two teams each from the Big East, Big Ten, and Big 12 to that tournament. We don’t know whether these have to be the two best non-NCAAT teams in each of those conferences, or how “best” would even be defined. There are a lot of ways to get six teams from those leagues, make them sit around two weeks, then ship them off to Vegas. There are a lot of ways to get the other ten teams the tournament advertises it will have. Personally, I struggle to see how the College Basketball Crown won’t face an even bigger opt-out problem than the NIT. Thankfully, that’s one of Fox Sports’s problems, in addition to those two lawsuits recently filed against them for sexual misconduct.
We didn’t worry about NIT opt-outs in today’s bracketology. To deal with the College Basketball Crown, we removed the first two teams in line for the Big East, Big Ten, and Big 12’s respective NIT exempt bids, making a note of that at the bottom. Our full model will account for opt-outs in some way. It will also account for the College Basketball Crown. We don’t know how it will do either of those things just yet.
