Seed List

The seed list below is our best current prediction of the final NCAA Tournament seed list. It is predictive of where things will end up. It is not a reflection of where things currently stand.

We’re running our full model again. Here’s how it works. One thing to know about this seed list: It deviates from our model in a few places. Our bracketology is built from our prediction model’s objective outputs. This seed list starts with those outputs and then makes small changes. We’ve listed those changes and the reasons behind them below the list.

If you’re looking for a reflection of where the bracket currently stands, we’d recommend Bracket Matrix. It is a great resource. It has all the brackets.

If it’s bracketology you want, or if you’re looking for more detail on teams’ chances to make and/or win the NCAA Tournament and NIT, those links are here:

Automatic bids are marked with an asterisk (*).

Last Updated: Sunday, Mar. 16 – FINAL

Seed	Team	Team	Team	Team
1	Auburn	Duke*	Houston*	Florida*
2	Alabama	Tennessee	St. John’s*	Michigan State
3	Texas Tech	Maryland	Iowa State	Kentucky
4	Clemson	Texas A&M	Wisconsin	Purdue
5	Louisville	BYU	Michigan*	Arizona
6	Saint Mary’s	Missouri	Oregon	Mississippi
7	Memphis*	Illinois	Gonzaga*	UCLA
8	Kansas	Marquette	Creighton	Mississippi State
9	UConn	Georgia	Baylor	Oklahoma
10	Arkansas	Vanderbilt	New Mexico	Utah State
11 (First Four)	West Virginia	Texas	San Diego State	Indiana
11	VCU*	Drake*
12	Colorado State*	UC San Diego*	Liberty*	McNeese*
13	Yale*	High Point*	Akron*	Grand Canyon*
14	UNC Wilmington*	Lipscomb*	Troy*	Robert Morris*
15	Montana*	Wofford*	Bryant*	Norfolk State*
16	Omaha*	American*
16 (First Four)	SIU Edwardsville*	Alabama State*	Mount St. Mary’s*	Saint Francis*

Changes From Our Model

As we said above, we build this seed list by making some changes to our model’s original outputs. We do this for two reasons: First, this list is what feeds in to Bracket Matrix, and we want to make a worthwhile contribution, since so many people rely upon Bracket Matrix for accurate projections. Our model is imperfect. We’re not too proud to admit that. Second, the process of finding potential misses by our model and correcting them gives us a list of features to explore adding to the model in the offseason. For instance: Last year, we realized our model was undervaluing Q1A wins in the selection part of the bracketing process.

If you want to see our model’s raw outputs, they’re what make up our NCAA Tournament Bracketology. As for what’s different on this list…

Clemson: Our model has Clemson as a 3-seed. We’ve bumped them down to a 4-seed over concerns that within the seeding process, our model might be overweighting the AP Poll. The penultimate AP Poll has had some predictive value in recent years with seedings, but we think our model could be overdoing it with a few teams. Moving Clemson down pushes Kentucky up to a 3-seed.
Louisville: Our model has Louisville as a 4-seed. We’ve bumped them down to a 5-seed over the same AP Poll concerns, as well as our concerns that while our model gives predictive ratings a heavier weight in the seeding process, it might still be weighting résumé ratings too heavily. Moving Louisville would push BYU up to a 4-seed, but we have AP Poll concerns about BYU, so we instead push Purdue up to a 4-seed.
Arizona: Our model has Arizona as a 6-seed. We’ve bumped them up to a 5-seed over the same AP Poll concerns.
Mississippi: Our model has Mississippi as a 7-seed. We’ve bumped them up to a 6-seed because we believe our model might be underweighting their large number of Q1A wins.
Memphis: Our model has Memphis as a 5-seed. We’ve bumped them to a 7-seed because of those AP Poll concerns, plus the concerns about our model underweighting the predictive ratings.
UCLA: Our model has UCLA an 8-seed. We’ve bumped them up to a 7-seed over concerns that our model isn’t properly accounting for the Bruins’ winning record in Q1A and/or large number of Q1 wins. This pushes Kansas down to an 8-seed.
VCU: Our model has VCU a 10-seed. We’ve bumped them to an 11-seed because our model goes off of median simulation, and in the median simulation, VCU beats George Mason today. The committee probably won’t fully consider this result in their seeding process. VCU’s ratings will aggregately improve if they win, but they won’t get credit for that because the game is happening so late. This pushes Utah State up to a 10-seed.
San Diego State: Our model has San Diego State out of the field. We’ve moved them into the field and ahead of Indiana in selection priority (in case that matters) over concerns that our model is undervaluing their win over Houston, the lone win over a projected 1-seed by a bubble team.
Ohio State: Our model has Ohio State in the field. We’ve moved them out of the field over concerns that our model is underestimating the consequence of being only two games over .500.