clock menu more-arrow no yes

Filed under:

Basketball: Georgia Tech's Chances in the NIT, per the LRMC

New, 22 comments

What do the in-house statistics gurus think about how the NIT will shake out?

Danny Karnik/Georgia Tech Athletics

As you may be aware by now, Georgia Tech ISyE professor (and chair of the senior design program when I completed it) Dr. Joel Sokol is the leader of a team of professors who developed a statistical model meant for predicting college basketball outcomes, especially when it comes to the NCAA Tournament. The model is named the "Logistic Regression/Markov Chain" (after the two primary mathematical techniques it employs), or "LRMC" for short. Basically, the system is used to "rank" each team in Division I based on a variety of metrics, and then those rankings are applied to the bracket in simple fashion -- whichever team is ranked higher in any prospective match-up gets the nod, and is predicted to move forward. (Thus, the team ranked #1 overall is the one predicted to win the tournament. Dr. Sokol would tell you himself that it's not a perfect system, but it's had a lot of success in the past as well.)

In conjunction with these calculations, the LRMC team also produces and releases a bracket each year using its numbers -- this year, Kansas is its champion. That said, there isn't one released for the NIT Bracket, but using their published numbers and methodology, we can form one ourselves.

So, what does the LRMC think about Georgia Tech's chances in the NIT?

Well, let's just say it doesn't like them very much. Below is the bracket along with each team's LRMC ranking. Remember, the higher-ranked team is to be chosen to win each matchup, going purely by LRMC standards.

REGION 1 (LRMC) REGION 2 (LRMC) REGION 3 (LRMC) REGION 4 (LRMC)
1 Valparaiso (34) 1 St. Bonaventure (98) 1 Monmouth (47) 1 South Carolina (52)
8 Texas Southern (145) 8 Wagner (170) 8 Bucknell (164) 8 High Point (142)
4 Florida St (44) 4 Creighton (49) 4 George Washington (59) 4 Georgia Tech (56)
5 Davidson (125) 5 Alabama (90) 5 Hofstra (77) 5 Houston (63)
3 Georgia (71) 3 Virginia Tech (72) 3 Ohio State (83) 3 Washington (73)
6 Belmont (124) 6 Princeton (102) 6 Akron (67) 6 Long Beach St (105)
2 Saint Mary's (27) 2 BYU (35) 2 Florida (33) 2 San Diego St (41)
7 New Mexico St (114) 7 UAB (113) 7 North Florida (161) 7 IPFW (138)

As you can see, LRMC picks Georgia Tech to win in the first round over Houston on Wednesday night, but its chances diminish from there. Afterwards, they would have to take on higher-rated South Carolina, where they would be picked to lose. For what it's worth, manually forcing them ahead in the tournament would have them losing in any given round moving forward, assuming that the standard methodology was used to pick all other games. (Further matchups would be projected against #41 San Diego State, #33 Florida, and #27 St. Mary's.)

The good news? Even though it's a 4-seed in its region, LRMC only ranks four teams higher than Georgia Tech on its half of the bracket, including Region 3. Also, the prospective second-round matchup with South Carolina is the closest prospective second-round matchup, with the two teams separated by only 4 spots in the rankings. It should be noted, too, that Georgia Tech is the 10th-ranked team in the tournament per LRMC.

Some other interesting notes:

  • The 2-seed in each region is ranked higher by the LRMC than that region's corresponding 1-seed, and the bracket's projected Final Four would be made up entirely of 2-seeds. Valparaiso is the only team in the tournament rated higher than any of the 2-seeds, but would be picked to lose to Saint Mary's in its regional final.
  • Even as a 1-seed in the tournament (and widely discussed as a major NCAA Tournament snub), LRMC ranks St. Bonaventure in the bottom-50% of the teams in the tournament. 19 of 31 other teams are ranked higher.
  • All of the 4-seeds in the tournament are rated better than all of the 3-seeds, with the closest margin being 11 spots (between George Washington and Georgia).
  • If remaining at a 4-seed, Georgia Tech's ideal position (per LRMC) would be to swap with Creighton. In that case, they would be projected to defeat Alabama and then St. Bonaventure, before losing to BYU in the regional final. That's the only region where Georgia Tech would make the regional final.

But, as we know, statistics don't always tell the whole story. How do you see things going?