NET Rankings are Flawed

BillBrasky4Cy

Well-Known Member
SuperFanatic
SuperFanatic T2
Dec 10, 2013
10,981
20,269
113
They don't drop or move even when they lose. You have enough good teams in a league and it makes the bad ones rise in their formula by just playing the games. Last Big Ten had it too. There was like 7-17 type teams in the top 100.

Iowa State has dropped at least two or three spots every time they've lost. Hell last night they picked up a Q1 road win and only moved up two spots. Iowa State has 6 Q1 wins and zero bad losses and that should matter more.
 
  • Like
Reactions: NCCisufan

GoHawks

Well-Known Member
Jul 12, 2009
1,521
602
113
Iowa State has dropped at least two or three spots every time they've lost. Hell last night they picked up a Q1 road win and only moved up two spots. Iowa State has 6 Q1 wins and zero bad losses and that should matter more.
Quad 1 wins are based off Net Rankings which you believe to be flawed is my point. I agree Iowa State should be higher in any rankings/seeding right now
 

NorthCyd

Well-Known Member
SuperFanatic
SuperFanatic T2
Aug 22, 2011
11,878
14,168
113
Iowa State has dropped at least two or three spots every time they've lost. Hell last night they picked up a Q1 road win and only moved up two spots. Iowa State has 6 Q1 wins and zero bad losses and that should matter more.
Hopefully it will. They highlight the quad wins and losses for a reason. This is from the NCAAs own web page:
"The number of Quadrant 1 wins and Quadrant 3/4 losses will be incredibly important when it comes time for NCAA tournament selection and seeding."
 

cyclones500

Well-Known Member
Jan 29, 2010
29,864
16,495
113
Michigan
basslakebeacon.com
Hopefully it will. They highlight the quad wins and losses for a reason. This is from the NCAAs own web page:
"The number of Quadrant 1 wins and Quadrant 3/4 losses will be incredibly important when it comes time for NCAA tournament selection and seeding."

And that's what people need to keep in mind -- the actual ranking should become more and more reflective as season progresses (it'll have some flaws, I'm sure) ... Quadrant results have more impact.

So in a way, we should want Iowa's # to remain fairly high -- the fact the relative placement seems skewed right now is mainly just an annoyance.
 

CascadeClone

Well-Known Member
Oct 24, 2009
5,511
5,182
113
I suspect with the OE number, they capped wins at 10 points, but that doesn't cap shooting %, TO%, et al when you are playing a Quad 7 team in early December. The reason you won by 35 is that the other team could guard a gallon of milk and you made 40 layups, shot 80% and only turned it over twice. That makes your OE amazing even adjusted for opponent, and even if you cap the win down to 10.

For ISU, their OE is not great, period - they are just not real efficient on offence. Last night they won, but the OE was far from amazing. Their OE early in the year against said Quad 7 teams was similarly bad, but then probably adjusts down below wet garbage level. I would guess that is what is holding ISU and maybe Wisky down.

I think it is good to use both predictive stats as well as actual results, this is a good compromise to "resume" vs "potential" argument that we see with the CFP all the time. But maybe some way to overweight recent results would be helpful. I know some of the computer systems use this concept, maybe NET does but who knows.
 

BillBrasky4Cy

Well-Known Member
SuperFanatic
SuperFanatic T2
Dec 10, 2013
10,981
20,269
113
And that's what people need to keep in mind -- the actual ranking should become more and more reflective as season progresses (it'll have some flaws, I'm sure) ... Quadrant results have more impact.

So in a way, we should want Iowa's # to remain fairly high -- the fact the relative placement seems skewed right now is mainly just an annoyance.

You won't see big swings in the model now, there are too many data points.
 

AuH2O

Well-Known Member
Sep 7, 2013
7,930
10,979
113
ISU and Wisconsin are the major underranked outliers. Iowa and Texas have been major ovverranked outliers.

Why I'm not sure. I don't think it's even net efficiency, as these are not different enough between the four teams to overtake the massive difference in Q1 performance.

The only explanation can be that it has a huge over weighting in offensive efficiency, still has a preseason bias built in or both. Either way, those are stupid things. Favoring a specific style of play to this extent is of course idiotic. Having preseason bias baked in 20 games into the season in a sport where massive turnover occurs is also idiotic.

In general I like the NET rankings, but there are some details of it that are clearly causing outliers. And that's OK, provided those outliers are treated as such by the committees. In the past it seems they have, but I do get annoyed when AP voters do things like interchangeably tweet about the #X BEST team then use overall KP rankings while there is a big preseason bias still included, a heavy weighting on how efficient you were in beating Jackson State, and it cares that you beat a terrible team ranked #230th instead of a terrible team ranked #340th.
 

CoKane

Well-Known Member
SuperFanatic
SuperFanatic T2
Oct 26, 2013
15,215
6,589
113
Cedar Rapids
I believe that it takes into account metrics, so our lack of offensive output it probably whats up there
 

Chitowncy

Well-Known Member
SuperFanatic
Jan 14, 2009
1,899
900
113
Ames
How about Houston at #3 when they're 0-2 in Quad 1. NET seems completely worthless to me.

"Completely worthless" is a bit of hyperbole, but there sure are some weird outliers, as the OP pointed out. Seems like we should be more in that hovering around 20 range right now. It seems mostly accurate and fairly well-designed, but then you get these teams like Houston, Iowa and Iowa State that seem like such outliers and are overvalued (UH) or undervalued (ISU).

I do hold out a hope and belief that the committee will look carefully at the team sheets and analyze each team more carefully. Can't imagine they'd seed us right now as a 7 seed if the Tourney started today and that's where we'd be if they seeded on the S-Curve based on exclusively the Net rankings. Got to think we're more like a 5 based on last night's game if the Tourney started today.
 

knowlesjam

Well-Known Member
Oct 21, 2012
4,078
4,421
113
Papillion, NE
ISU and Wisconsin are the major underranked outliers. Iowa and Texas have been major ovverranked outliers.

Why I'm not sure. I don't think it's even net efficiency, as these are not different enough between the four teams to overtake the massive difference in Q1 performance.

The only explanation can be that it has a huge over weighting in offensive efficiency, still has a preseason bias built in or both. Either way, those are stupid things. Favoring a specific style of play to this extent is of course idiotic. Having preseason bias baked in 20 games into the season in a sport where massive turnover occurs is also idiotic.

In general I like the NET rankings, but there are some details of it that are clearly causing outliers. And that's OK, provided those outliers are treated as such by the committees. In the past it seems they have, but I do get annoyed when AP voters do things like interchangeably tweet about the #X BEST team then use overall KP rankings while there is a big preseason bias still included, a heavy weighting on how efficient you were in beating Jackson State, and it cares that you beat a terrible team ranked #230th instead of a terrible team ranked #340th.
If you look at the actual calculations, the OE number always comes out higher than the DE number. For some reason, the calculation gives more weight to the OE number. To be honest, ISU being around 26 seems to be just about right for tourney seeding...a 6-7. Their OE is crap, by the DE and quad 1 wins raise the seed.
 

Hoggins

Well-Known Member
SuperFanatic
SuperFanatic T2
Sep 2, 2019
2,217
2,858
113
34
More fun. ISU has a net SOS of 12.

Average conference NET rank
Big 12 - 34
Big 10 - 59
SEC - 67
Big East - 69
ACC - 91
Pac 12 - 92
AAC - 107
 

FriendlySpartan

Well-Known Member
Jul 26, 2021
886
907
93
35
You guys have 11 games left plus the conference tournament, and you play in the toughest conference this year. Dont worry about some stupid Net rating, Win 5/6 more games and you are an easy 7/8 seed. Win more or get an upset of KU or Baylor and jump even higher. Way to much season left to be worrying about stupid computer ratings.
 

AuH2O

Well-Known Member
Sep 7, 2013
7,930
10,979
113
If you look at the actual calculations, the OE number always comes out higher than the DE number. For some reason, the calculation gives more weight to the OE number. To be honest, ISU being around 26 seems to be just about right for tourney seeding...a 6-7. Their OE is crap, by the DE and quad 1 wins raise the seed.
Yeah, I just never understood why efficiency numbers have much weight at all as the season winds down. After conf. tournaments, some of these cross-conference challenges, you've got a hell of a lot of data on, you know, actual wins and losses and SoS. The fact that Iowa keeps playing starters to bloat their offensive efficiency vs. Western Michigan, or throws the press on with their starters vs. Wisconsin's walk-ons to salvage an ass-kicking doesn't seem like something that should be helping a team.