Over the last few days Clemson Men’s Basketball fans have been befuddled by the NET Rankings, which, according to the NCAA are the “primary sorting tool for Division I men’s basketball”.
You can read more than you ever cared to about how these rankings came to be and the “justification” for them if you’d like, but I’ll give you a brief below.
First, a couple of primers:
- I’m not against using algorithms to evaluate teams, as a general rule. Heck, I used to do it myself, creating my own that was moderately successful at predicting football games and horrible against the spread.
- While I’ll admit to some Clemson bias (duh), I don’t actually think the Tigers should be ranked exceedingly high in this metric. Perhaps 40ish right now, but that’s just spitballing.
With that out of the way, let’s take a look at the NET and point out a couple of things and explain why Clemson is in trouble.
"The NET rankings system was improved in the summer of 2018 after consultation with the Division I Men’s Basketball Committee, the National Association of Basketball Coaches, top basketball analytics experts and Google Cloud Professional Services.Late-season games from the 2017-18 season, including from the NCAA tournament, were originally used as test sets to develop a ranking model that used machine learning techniques. The model was used to predict the outcome of games in test sets and it was optimized until it was as accurate as possible.That model is the one used for the NET.Game date and game order were not included in the NET rankings so a team’s first game counts the same as its 30th.With the changes announced in May 2020, the NET will no longer use winning percentage, adjusted winning percentage and scoring margin. The change was made after the committee consulted with Google Cloud Professional Services, which worked with the NCAA to develop the original NET."
So, a couple of things here that stand out to me:
- The NET is designed to predict outcomes of games, not determine who is more deserving to be in the NCAA tournament. As a matter of fact, as of 2020 they don’t even include winning percentage in the formula, so being 18-4 means nothing. See No. 4 below.
- It seems strange that the “primary sorting tool” and one that is used as a tool to evaluate teams for the NCAA Tournament is not designed to determine which teams are more deserving, but rather predict outcomes of games. In my mind, just because Team A might beat Team B, doesn’t mean Team B is any less deserving to be in the tournament.
- December losses that you and I may be discounting live forever.
- The claim is the model is results driven AND predictive. That has proven to be false for Clemson as the Tigers are 5-1 against teams ranked above them (not sure where they were ranked at time games were played). But just because it’s not predictive for Clemson doesn’t mean it’s not predictive for college basketball as a whole. Predictive doesn’t mean perfect.
Despite the tweet below being from 2018, it was included in the article linked above from December of 2022, so I assume it’s still accurate.
https://twitter.com/MarchMadnessMBB/status/1067063960753573889?s=20&t=0CWNrmXURKdR65mu2nCZJQ
Clemson Men’s Basketball will have a tough time improving their NET Ranking much moving forward
The Tigers are in a tough spot as far as the NET Ranking goes. They can’t afford to lose to ANY “bad” team on their schedule and they don’t play enough “good” ones to improve their rankings that much and even if they do, they will likely be offset by the “bad” teams.
For me, I get it, those two losses in December are part of your record, but that’s not traditionally how college basketball teams have been judged, which is tough to swallow when you see how this team has played since mid-December.
As far as the NET is concerned, those December losses might as well have happened this week and that’s a problem for Clemson.
The hope is the humans in the room on selection Sunday will see what the computers cannot.
I’m not holding my breath.