I'm intrigued by the NET rankings. I guess this is the first year I really ever paid attention to them. I find it interesting that the formula is secret, that NET efficiency is just another way to say scoring margin, and how they define scoring/classification of quad 1,2,3 etc wins/losses (note these don't really directly impact NET rankings)
With all this data analysis in College Football/Basketball, I think the biggest flaw is the lack of connection points between teams. Each conference essentially stays within its own bubble for most of the year, which can artificially inflate a team's metrics. This was evident in the Big12 football where their entire conference had virtually no meaningful wins against other P5 teams. BBall is a little bit better since there are interconference games, but NET rankings can still be inflated.
I took a closer look at the comparison between Mizzou and Utah St to see how they ended up with such a higher ranking.
Here is a breakdown for each:
Utah State: (record and +/- differential per game)
Net 0-50 Record 4-5 (-1.4 pts)
Net 50-100 Record 8-0 (+14.5)
Net 100-200 Record 11-2 (+10.8)
Net 200+ Record 4-0 (+14.5)
Total +/- 8.8 pts/game
Mizzou:
Net 0-50 Record 7-8 (-4.7pts)
Net 50-100 Record 2-1 (-1.3 pts)
Net 100-200 Record 6-0 (+11.3 pts)
Net 200+ Record 8-0 (+21 pts/game)
Total +/- 5.01 pts/game
I always had thought we did not blow out the sub 100 teams, but that doesn't seem to be the biggest issue. I think the reason why Utah St has a higher NET is they stayed closer in the 0-50 range and really did well in the 50-100 range.
In teams ranked 0-50, Utah State had about the same record as us, but were only outscored by an average of 1.3 pts vs our 4.7 pts. However, here is actually who they played:
San Diego St (3 times - lost all 3)
Boise St (3 times)
Oral Roberts
Nevada (2 times)
Here are Mizzou Top 50 teams:
Alabama (2 times)
Tennesse (2 times)
Kansas
Texas AM (2 times)
Arkansas (2 times)
Kentucky
Auburn
Illinois
Iowa St
Miss St (2 times)
In teams ranked 50-100, they were 8-0 and outscored them by an average of 14 pts.
However, a closer look, here are the teams they played 50-100:
New Mexico (2 times)
Bradley
Utah Valley
Santa Clara
UNLV (2 times)
For comparison, here are the teams we played in the 50-100 ranking:
Florida
UCF
Vanderbilt
I know it is just one metric, but Vegas does put stake in it and obviously it has some influence over the committee. Next year we need to schedule more teams in the 50-100 range (like UCF etc) and be sure to blow those types of teams out. Also probably no need to play top 50 teams out of conference.
With all this data analysis in College Football/Basketball, I think the biggest flaw is the lack of connection points between teams. Each conference essentially stays within its own bubble for most of the year, which can artificially inflate a team's metrics. This was evident in the Big12 football where their entire conference had virtually no meaningful wins against other P5 teams. BBall is a little bit better since there are interconference games, but NET rankings can still be inflated.
I took a closer look at the comparison between Mizzou and Utah St to see how they ended up with such a higher ranking.
Here is a breakdown for each:
Utah State: (record and +/- differential per game)
Net 0-50 Record 4-5 (-1.4 pts)
Net 50-100 Record 8-0 (+14.5)
Net 100-200 Record 11-2 (+10.8)
Net 200+ Record 4-0 (+14.5)
Total +/- 8.8 pts/game
Mizzou:
Net 0-50 Record 7-8 (-4.7pts)
Net 50-100 Record 2-1 (-1.3 pts)
Net 100-200 Record 6-0 (+11.3 pts)
Net 200+ Record 8-0 (+21 pts/game)
Total +/- 5.01 pts/game
I always had thought we did not blow out the sub 100 teams, but that doesn't seem to be the biggest issue. I think the reason why Utah St has a higher NET is they stayed closer in the 0-50 range and really did well in the 50-100 range.
In teams ranked 0-50, Utah State had about the same record as us, but were only outscored by an average of 1.3 pts vs our 4.7 pts. However, here is actually who they played:
San Diego St (3 times - lost all 3)
Boise St (3 times)
Oral Roberts
Nevada (2 times)
Here are Mizzou Top 50 teams:
Alabama (2 times)
Tennesse (2 times)
Kansas
Texas AM (2 times)
Arkansas (2 times)
Kentucky
Auburn
Illinois
Iowa St
Miss St (2 times)
In teams ranked 50-100, they were 8-0 and outscored them by an average of 14 pts.
However, a closer look, here are the teams they played 50-100:
New Mexico (2 times)
Bradley
Utah Valley
Santa Clara
UNLV (2 times)
For comparison, here are the teams we played in the 50-100 ranking:
Florida
UCF
Vanderbilt
I know it is just one metric, but Vegas does put stake in it and obviously it has some influence over the committee. Next year we need to schedule more teams in the 50-100 range (like UCF etc) and be sure to blow those types of teams out. Also probably no need to play top 50 teams out of conference.
