When the season began and we built our college basketball model, Georgia Tech had not yet appealed its NCAA sanctions. Which, pertinently, included a postseason ban.
A few weeks later, Georgia Tech did appeal those sanctions. The appeal process is ongoing.
We made the mistake of not checking in on Georgia Tech’s status earlier. We missed that they appealed. We left them out of our model, and while it didn’t affect other teams’ numbers too dramatically, it did affect other teams.
Last night, upon learning of the appeal, we reached out to the Georgia Tech men’s basketball program, and received confirmation that until the NCAA rules on the appeal, the Yellow Jackets are eligible to be considered by the selection committees of the NCAA Tournament and, more relevantly in their case, the NIT. They’re slated, as of now, to play in the ACC Tournament.
This could, of course, change again—the NCAA could uphold the ban and announce that decision prior to the beginning of the ACC Tournament, and/or prior to Selection Sunday. But at least for the time being, Georgia Tech is postseason-eligible. Our apologies for not catching this sooner.
While we’ve got you here, a few notes on Georgia Tech, and a few notes on our model:
Georgia Tech
Georgia Tech enters the weekend 13-13, with a respectable strength of schedule, only one “bad” loss (at home to Ball State in December), and two more-than-respectable wins (they beat Louisville at home nine days ago, and beat N.C. State in Raleigh to open the season). They are certainly in the mix for the NIT, and if they can upset Syracuse on the road this weekend, I’d expect you’d see them in our NIT Bracketology on Monday.
The Model
I’ll have more on this in a few days when the process is complete, but we’ve begun trimming our model’s error margins. We’ve made reference to this in a couple places, but I wanted to explain what exactly it means:
Our model began the year with very wide error margins. It ended last year with very wide error margins. Too wide for our liking. It was calibrated correctly (20% of the things we said were 20% likely did end up happening, and ditto for other numbers), but it wasn’t very precise. In our hurry to get the model out, we did not do the work necessary to tighten the error margins when we built the model. It’s served us well enough so far, but with Selection Sunday rapidly approaching, we want more precision, so we’re making updates that will tighten those error margins. Most teams won’t see much of a change from this, but some will see a noticeable change (two or three seed lines in the bracketologies, for the biggest movers). This will be the last significant adjustment to our model for the season, so don’t expect this to be a regular occurrence, but still, our apologies for not getting this done in December or November. We intended to do this earlier, but this is not our full-time job at the moment, and we just didn’t make the time.
When the process is complete (we’re aiming to finish it Sunday so the changes will be reflected in Monday’s bracketologies), I’ll write up a better description of how our model works than the one I wrote last year and get that posted where it can be conveniently found.
***
Thanks for spending your time here at The Barking Crow. As always, hit us up with your questions and thoughts. We always appreciate the questions, and we almost always appreciate the thoughts.
