NIT Bracketology: Monday, March 4th

Weekend? Over. NIT Bracketology? Updated. Here’s our model’s latest NIT Bracketology. Here are our model’s latest NCAA Tournament/NIT probabilities. Here’s where the NIT picture stands:

Last Four In, First Four Out, Etc.

On the low side:

• Last Four In (from the cut line going upwards): Yale, Maryland, Loyola Chicago, Washington
• First Four Out (from the cut line going downwards): NC State, San Francisco, UNLV, LSU
• Others in the mix (>10% NIT chance, below NIT field, not in our field or in the First Four Out): Indiana, Florida State, UMass, Rutgers, Miami (FL), Duquesne, Georgia

On the high side:

• Last Four In (from the cut line going downwards): Wake Forest, Utah, Drake, Texas A&M
• First Four Out (from the cut line going upwards): Virginia, New Mexico, TCU, Colorado
• Others in the mix (>10% NIT chance, above NIT field, not in our field or in the First Four Out: Florida Atlantic, St. John’s, Indiana State, Oklahoma, Villanova, Saint Mary’s, Northwestern

As always, we account for bid thieves by looking at the median cut lines in our model’s simulations. Today, those remain 45^th and 82^nd, meaning if there was a master seed list for both tournaments, the 45^th overall seed would be the last NCAAT at-large and the 82^nd seed would be the last NIT at-large.

Conference Tournament Favorites

Because of where those cut lines stand and how many teams sit between them, we’re only including one conference tournament favorite in our NIT Bracketology right now. Overall, yes, it’s likelier than not that one of JMU and App State wins the Sun Belt Tournament. Each, though, is individually likelier to make the NIT than miss the NIT. So, they’re both in there, but JMU could just as easily be Princeton, Grand Canyon, or McNeese. They’re a filler for that kind of team.

Here are all the conference tournament favorites more than 50% likely to make the NIT if they lose that conference tournament:

• Princeton
• Grand Canyon
• James Madison (in NIT Bracketology to keep cut line at 82)
• McNeese

We have Indiana State more likely to make the NCAA Tournament as an at-large than the NIT as an at-large. We have Samford narrowly more likely to miss the NIT than make it if they don’t win the SoCon Tournament.

Where We’re an Outlier (Seton Hall & Colorado Talk)

Only a moderate percentage of bracketologies have updated to account for Saturday’s results, and fewer have responded to Seton Hall’s 30-point loss yesterday at UConn. So, we don’t have a perfect sense of how our model compares to the industry. Still, in our subjective Seed List, we’ve moved Seton Hall into the NCAAT field, albeit in a bid thief seat. Let’s talk about that.

What’s happening here (apologies to regular readers, we’re rehashing some things) is that there are two ways our model can disagree with the industry.

The first is in a predictive/reflective sense. Our model only looks at where it expects the brackets to end up, based on each team’s median set of outcomes from here. It doesn’t reflect where things would stand “if the season ended today.” This is a big distinction between our model and the industry. It’s not apples to oranges, but it’s red apples to green apples. They’re different. Which makes some of the differences not disagreements as much as differences of preferred approach.

To use poker terminology, Colorado has a few outs remaining. They’re about 50/50 to beat Oregon on the road, something which would give Colorado a second Q1 win. Washington State could climb into the NET top 30, something which would give Colorado a second or possibly third Q1 win. Colorado might meet Utah in the Pac-12 quarterfinals, something which would give Colorado a second, third, or fourth Q1 win if Utah holds onto a top-50 NET position. Colorado would be favored over Washington State in the Pac-12 semifinals, something which…you get the picture. Right now, Colorado is 1–5 against Q1, 7–4 against Q2, and 12–0 against Q3 and Q4, having played through the absences of significant contributors and currently boasting a stronger NET ranking than anyone who’s missed the NCAA Tournament in the NET era. The issues are Q1 wins (which our model does account for, in a few ways) and Q1A wins (which our model doesn’t account for directly).

It’s possible Colorado’s surge could come too late, like Texas A&M’s in 2022, and that the committee will have made up its mind before the season is actually over. Still, looking at all these possibilities, we think our model’s difference of opinion on Colorado has more to do with the predictive/reflective difference of preference than an actual difference of opinion between our model and the industry. We’re uneasy about it, because our model has a below-average track record of predicting the NCAAT/NIT bubble, but we’d still bet on Colorado to make the field, if forced to make a choice.

The second way our model can disagree is about the criteria the committee will use, and how heavily and in what ways they’ll emphasize different pieces. We like having an objective model because it keeps us from following groupthink too closely, but we’ve yet to get the coding right around the NCAAT/NIT bubble. We trust the industry more than ourselves.

With Seton Hall, this is relevant, because the thing I assume the industry likes about Seton Hall is that they’ve beaten both UConn and Marquette, teams likely to wind up as a 1-seed and a 2 or 3-seed, respectively. Our model doesn’t consider Q1A wins. It doesn’t consider Best Win or Second-Best Win. We’ve tried adding those, but we haven’t found a precise enough way to do it.

It’s possible that with Seton Hall’s blowout loss yesterday, the industry will turn on the Pirates. This is why our subjective list only has them in one of the two Bid Thief Seats and not taking a full-on NCAA Tournament bid. It’s a bit of a hedge. (Context: To determine NCAA Tournament automatic bids, our model looks at conference tournament favorite, not conference tournament 1-seed. It expects two bid thieves. Bracket Matrix looks at conference tournament 1-seed, so it already has two bid thieves in there in the form of Richmond and South Florida. Teams in our Bid Thief Seats are, right now, equivalent to the first two teams out on Bracket Matrix.)

Overall, if you push Bid Thief Seat occupants into the NIT (as our NIT Bracketology does), our differences of opinion with Bracket Matrix as it currently stands (it’s currently a mix of bracketologies from Friday to yesterday) are about Seton Hall, Providence, Wake Forest, Villanova, St. John’s, and Colorado. Of those, we think the Villanova/Providence split will resolve itself once every bracket updates to account for Villanova beating Providence at the Dunk. That leaves us with four teams left. Our model expects St. John’s and Colorado to play their way in, and we don’t disagree (we’ll take a better look at St. John’s tomorrow if the Matrix still has them out). This leaves Wake Forest and Seton Hall. Would those two teams be the ones bumped out? We’re comfortable saying yes to Wake Forest. Our model is only one spot off of Bracket Matrix on the Deacs. Again, though, we’re unsure about Seton Hall. New Mexico and Virginia are also possibilities to take the fall, both in our model’s eyes and those of the Matrix.

There are 14 days of basketball left before the brackets are unveiled. A lot will change. But our model is far from 100% accurate on this stuff, and since we can predict some of its inaccuracies, we think they’re worth discussing. Just because our model thinks something will happen doesn’t mean we personally expect it.

Our current plan is to look at our model’s differences with Bracket Matrix tomorrow, and to then compare its NIT/CBI cut line to John Templon’s on Wednesday. Circumstances may change that approach.

For the Archives