Monday, January 01, 2024

NET Ranking Black Magic

There are several elements to the NET ranking.  The Quad element is supposed to be a heavily weighted portion, but the current results for Syracuse are baffling.

Here are the Orange's NET ranking compared to two teams they have beaten, and a third team. The data is their record for each quadrant, as of 12-31-2023.

Syracuse #78 rank  0-2 Q1, 2-1 Q2, 6-0 Q3, and 1-0 Q4
Pittsburgh  #41 rank  0-2 Q1, 0-1 Q2, 4-1 Q3, and 5-0 Q4
Oregon  #71 rank 0-1 Q1, 0-1 Q2, 6-1 Q3, and 4-0 Q4
SMU  #45 rank 0-3 Q1, 0-1 Q2, 2-0 Q3, 6-0 Q4

In theory Q1 losses don't hurt you, and Q4 wins don't help.  Meanwhile Q1 wins are huge, and Q4 losses hurt.

None of these teams have Q1 wins.  Syracuse has played only one Q4 game, as opposed to the 5, 4, and 6 of the other teams. Pitt and Oregon both have a Q3 loss (bad); Syracuse has not bad losses. 

Syracuse is 2-3 in Q1&Q2 games. Pitt is 0-3, Oregon 0-2, and SMU 0-4.

Syracuse is 7-0 in Q3&Q4.  Pitt is 9-1, Oregon 10-1, SMU 8-0.

Someone countered with me the other day that the large margin of victories for the other teams must be coming into play.  But the NCAA said in 2018 they cap margin of victory at 10 points, so that should have no impact on the Orange.

Here's the Orange margin of victory so far this season after 12 games (Chaminade doesn't count for the NET).

Point margin: +11, +12, +4, -17, -19, +23, -22, +11, +12, +20, +12, +8

All of the Orange wins except for Colgate and Pitt hit the cap.  Hard to believe the opposition is doing more.

I know they take computer modeling into play, but that's all secretive, but also supposed to be a smaller portion of the ranking.  The black box magic going on is baffling.  Perhaps small sample sizes are coming into play, but we are over a third of the way through the schedule.

The Orange have more quality wins, no bad losses, a tougher schedule, better overall record, and are capping the point spread margin, yet are lower ranked, significantly in 2 cases, than Pitt, Oregon, and SMU.



1 comment:

Anonymous said...

It seems that they must be weighing the per possession stats more than it appears. If you look at Ken Pom, teams mostly get set in their place after about 7 to 9 games played and moving significantly after that is rare. Also, while margin of victory is capped, the per possession stats can be heavily skewed with a few blowouts to teams that latter struggle (Gonzaga, Virgina).