GT Baseball in USI Projected Tournament Field

After weeks of being on the bubble, we have Tech in our projected NCAA Tournament field

For those familiar with our ramblings on Scions of the Southland, you’ll know we often refer to the “USI” when talking about Georgia Tech Volleyball, Baseball, and Softball. Finally, we’re going to actually make that ambiguous term practically mean something for the masses here on fromtherumbleseat.com. This is the Universal Selection Index, Jake Grant’s proprietary NCAA Tournament field predictor.

I am no expert when it comes to creating tools like this, but at the very least I can tell you what it spits out and what seems right/wrong. Essentially, USI is a bunch of weighted averages that incorporates strength of schedule, adjusted RPI, and records across multiple categories of opponents. Here’s Jake with a more detailed explainer:

Jake here, with some clarifying thoughts. The short answer is, yes, it is a big weighted metric that aggregates multiple different consideration to rank teams. Given I am endeavoring to be consistent in how we conceive of the college postseason, I wanted to look at a publicly available dataset that is consistent across sports. For us, that is available via the NCAA’s record-keeping. Taking the available record data (adjusted RPI, conference and non-conference records, various quadrant ratings, etc.), we can reasonably calculate a composite indicator assessing a team’s body of work. The result is a rating between 0 and 1, with 1 being a perfect season and 0 being the opposite (and associated rank among eligible Division 1 teams). There was some consideration given to not over-weight any particular parameter, but in the calibration process, I found this generally took care of itself.

In the period since I have been regularly collecting this data, excluding 2021 due to pandemic-affected schedules, the calibrated USI metric correctly predicted 227 of 240 teams from multi-bid leagues, with a few very obvious outlier conference tournament champions excluded. This equates to a 94.6% success rate, with 95.3% of cumulative USI represented by those multi-bid league teams included, indicating that places where it misses are comparable bubble teams, rather than obvious misses. In as few words as possible, I believe this is due in part to bid-stealers, nuance in tie-breaking that a straightforward formula misses, and a touch of power conference bias.

For the sake of prognostication, I find it an effective way to assess the body of work of teams throughout the season, and that the approximately 95% success rate is perfectly suitable for our needs.

Coming into the weekend against Duke, Tech sat in the First Four Out along with Auburn, South Alabama, and Coastal Carolina. While they were ranked #50 overall in USI, the number of automatic bids from conference champions shoved out a number of potential at-large teams, including Tech.

After the 2-1 weekend vs. Duke, Tech climbed about seven spots in our rankings to the Next Four In category at #43 overall:

One disclaimer here, we are extremely confident Auburn is not going to make the tournament field unless they win the SEC Tournament. While they are in the projected field below as well, take that with the requisite grain of salt. Here’s Jake again with why we’re making this disclaimer:

Given that the SEC generally sends all or all but one of its softball teams to the postseason, this is one place where I am still tinkering sport-specific minimum conference record thresholds, including when in the season it should kick in. As ever, I am open to feedback, but am currently thinking that a minimum record of .333 in conference would be a good place to start (and could generally be safely applied across all sports).

Current Projected Field (sorted by regional location*)

We do not expect this to be a perfect prediction of what the tournament field will look like when the time comes to create the bracket, but merely a reflection of who would be in the field if the tournament began today. At the moment, USI only takes conference, top 16 status, and ranking into account when creating the regionals, so the groups are going to look a little wonky until that update comes. For example, Site 9 below is absolutely not happening.

Jake here for another time. I’ve got an optimized regional selection method in the working, it is just not ready for primetime yet. It’s a hobby, as much as anything, so I try not to break anything in the middle of the season, in case a few weeks pass between available time for updates. Of note, the numbers really like UC Irvine, and it makes sense to have at least one West Coast regional. I could definitely see them as a host, but given the nuance of human selections, I could see their national seed being in the 9-16 range.

Site 1 (College Station): Texas A&M, Oklahoma, Oregon, Bethune-Cookman

Site 2 (Fayetteville): Arkansas, Southern Miss, UNCW, Oral Roberts

Site 3 (Knoxville): Tennessee, Northeastern, St. John’s (NY), Farleigh Dickinson

Site 4 (Athens): Georgia, Louisiana, Nebraska, Arkansas-Little Rock

Site 5 (Lexington): Kentucky, Lamar, Troy, Army

Site 6 (Clemson): Clemson, LSU, Maryland, Bryant

Site 7 (Chapel Hill): North Carolina, Louisiana Tech, UCF, Columbia

Site 8 (Irvine): UC Irvine, NC State, Auburn*, Bowling Green

Site 9 (Charlottesville): Virginia, San Diego, Hawaii, Fresno St.

Site 10 (Tallahassee): Florida State, UC Santa Barbara, TCU, Wright St.

Site 11 (Greenville, NC): East Carolina, Mississippi St, Virginia Tech, VCU

Site 12 (Winston-Salem): Wake Forest, Vanderbilt, Cal (last 4 in), USC Upstate

Site 13 (Terre Haute): Indiana State, Alabama, Georgia Tech, Niagara

Site 14 (Dallas): Dallas Baptist, Oregon St., Kansas St. (last 4 in), Austin Peay

Site 15 (Durham): Duke, South Carolina, Creighton (last 4 in), Wofford

Site 16 (Stillwater): Oklahoma State, College of Charleston, Coastal Carolina (last 4 in), Grand Canyon

*As noted, Auburn is almost certainly out of the tournament. Adjust expectations accordingly.

At-Large bids by conference

SEC: 11

ACC: 8

Big 12: 4

Sun Belt: 3

Big West: 2

CAA: 2

Pac-12: 2

Big East: 1

Big 10: 1

CUSA: 1

What this means for Georgia Tech

Sitting at the Next Four In and facing Florida State on the road this weekend, their adjusted RPI and USI can make a positive jump if they get a win or two. By not playing Mercer this evening (cancelled due to weather), they avoid a massive potential resume blemish which none of the metrics would’ve liked.

There is still the ACC Tournament to come as well. In the worst case scenario where Tech doesn’t win again, that would likely put them back on the wrong side of the bubble.

We’ll be back next week after the FSU series with our next USI update.