NET Rankings are Flawed

ISU_Guy

Well-Known Member
Jul 21, 2021
5,107
4,093
113
47
Des Moines
I think what will help Iowa get in is their win total and the fact they have a player of the year candidate.
It would be a total iowa move to get in with zero Q1 wins though.
 

BillBrasky4Cy

Well-Known Member
SuperFanatic
SuperFanatic T2
Dec 10, 2013
17,479
31,791
113
This is the ultimate question and a great 1. If you use the NET rankings fully I'd say today both should get in no question. Iowa State has beaten many many good teams bases off the NET Rankings so it's in. Iowa has a lofty NET ranking so it gets in. If you use it selectively to determine the opponents you've beaten then Iowa's in trouble. If you use it selectively by the individual ranking Iowa States in trouble. I'm really not sure what they'll do or which is best. I lean to be Iowa State and have the quality wins over ranking based off bracketolgists still having teams like Providence as high as a 2.

At some point in the season who you beat and who you lose to has to matter. I get that everyone is obsessed with analytics but they don't paint the full picture. Until they tinkered with the formula a year or two ago Q1 and Q2 carried much more weight than they do now.
 

BillBrasky4Cy

Well-Known Member
SuperFanatic
SuperFanatic T2
Dec 10, 2013
17,479
31,791
113
Very tough call. Both teams have highly unusual and historically unique resumes that could enter meme-territory on Selection Sunday if they keep up the trends (Iowa doesn’t win a Q1 game all year and ISU only wins 1 more game this year).

I think Iowa could slide A LOT if they struggle down the stretch given the lack of Q1 wins. Could fall into the 30s on NET and then become a 2019 NC State and get “snubbed” of a bid or be “just in” because of the lack of good wins.

ISU at 18 wins would be an interesting test case given I can’t imagine many teams have played 15+ Q1 games and won 8 of them as well as avoided any Q3/Q4 losses. NET is penalizing them for bad efficiency metrics vs our Q4 opponents too.

There are way too many data points, the models won't overreact now. Iowa could lose three out of the next four and would move less than 10 spots.
 

CyTwister

Active Member
Aug 30, 2019
124
102
43
29
There are way too many data points, the models won't overreact now. Iowa could lose three out of the next four and would move less than 10 spots.

When I said that I meant to say Iowa has a catastrophic collapse and goes 2-6/1-7 to finish the season with no Q1 wins. Could see that tanking them like our recent streak tanked our ranking.
 
  • Agree
Reactions: GoHawks

allfourcy

Well-Known Member
SuperFanatic
SuperFanatic T2
Apr 26, 2012
6,950
2,983
113
Iowa won't get in with 0 Q1 wins.

They probably get Mich at least once (I'm not totally impressed with Mich despite their win vs. Purdue), Northwestern and more than likely Mich. St at home (MichSt not playing well), plus Nebr on the road. No way they leave out that record w/ at or above .500 in the BigTen. Of course, I guess if they win what I'll think they will, they'll finally have some of those Q1 wins.
But why again is their NET so high? Based on their high scoring totals/differentials? Keegan Murray is so smooth. Wish we had a legacy recruit like that.
 

qwerty

Well-Known Member
SuperFanatic
SuperFanatic T2
Apr 3, 2020
7,682
11,626
113
60
Muscatine, IA
They probably get Mich at least once (I'm not totally impressed with Mich despite their win vs. Purdue), Northwestern and more than likely Mich. St at home (MichSt not playing well), plus Nebr on the road. No way they leave out that record w/ at or above .500 in the BigTen. Of course, I guess if they win what I'll think they will, they'll finally have some of those Q1 wins.
But why again is their NET so high? Based on their high scoring totals/differentials? Keegan Murray is so smooth. Wish we had a legacy recruit like that.
Michigan St in CHA is the only one of the four listed that is Q1 at this time, and they only need to fall 4 spots to become Q2 game. Very real chance Iowa ends regular season with 0 Q1 wins.
 
  • Optimistic
Reactions: VeloClone

allfourcy

Well-Known Member
SuperFanatic
SuperFanatic T2
Apr 26, 2012
6,950
2,983
113
Michigan St in CHA is the only one of the four listed that is Q1 at this time, and they only need to fall 4 spots to become Q2 game. Very real chance Iowa ends regular season with 0 Q1 wins.

Why isn't Michigan a Q1 win? (guess I left out a Q1 at OhioState- Iowa could win there if Murray is his usual very good and they let Jorbo go off)
 

BillBrasky4Cy

Well-Known Member
SuperFanatic
SuperFanatic T2
Dec 10, 2013
17,479
31,791
113
Why isn't Michigan a Q1 win? (guess I left out a Q1 at OhioState- Iowa could win there if Murray is his usual very good and they let Jorbo go off)

Beating Michigan at Carver would be a Q2 win but winning at Mich would be Q1. If Iowa loses to either Michigan or Michigan St at home they will need to either win at tOSU or at Illinois. Iowa resume is really really bad. If they were in the Pac they wouldn't even be in the bubble discussion right now.
 

khaal53

Well-Known Member
SuperFanatic
SuperFanatic T2
Apr 13, 2006
2,894
618
113
41
still would like some input from @khaal53 on his thoughts with our NET compared to others


*****

Rankings

The biggest thing when looking at rankings is we all need to understand what the ranking is for and how it is being calculated. There has been much consternation of Iowa State’s NET ranking but specifically their NET ranking in how it relates to Iowa. I get it, I personally don’t worry about it, but I get it.

In theory, the NET ranking is some unknown magical formula to homogenize performance rankings and resume rankings, but we really have no idea how. My two main thoughts on this…

It is clearly more tied to performance ratings that can be sorted out pretty quickly just by glancing at NET rankings compared to KenPom in contrast to how the NET rankings compare to a pure resume ranking like WAB (wins above bubble) is essentially the relative number of wins a team his against their schedule when compared to what an average bubble team would be expected to have. At this minute, Iowa State is #28 in NET, #25 in KenPom, and #20 in WAB.

What is interesting is that if you look at the top 30 NET teams the average variation in rank to KenPom is 3.03 while the average variation, when compared to WAB, is 8.07. So yes, NET leans heavier to a performance-based metric than a resume-based metric.

My 2nd beef is the quad win discussion. If a team has a win over the #1 team in the country on the road that is a quad one win just as a win over the #30 team at home is and those are two entirely different things that should be considered at more of a variable level. WAB on the other hand accounts for the variable data more effectively.

Whether we’re ranking team quality or team resumes when we get to March and deal with the selection committee these numbers can matter but we’re mostly at the mercy of a bunch of humans and whatever they want to rationalize, so, God speed!
 

BillBrasky4Cy

Well-Known Member
SuperFanatic
SuperFanatic T2
Dec 10, 2013
17,479
31,791
113

*****

Rankings

The biggest thing when looking at rankings is we all need to understand what the ranking is for and how it is being calculated. There has been much consternation of Iowa State’s NET ranking but specifically their NET ranking in how it relates to Iowa. I get it, I personally don’t worry about it, but I get it.

In theory, the NET ranking is some unknown magical formula to homogenize performance rankings and resume rankings, but we really have no idea how. My two main thoughts on this…

It is clearly more tied to performance ratings that can be sorted out pretty quickly just by glancing at NET rankings compared to KenPom in contrast to how the NET rankings compare to a pure resume ranking like WAB (wins above bubble) is essentially the relative number of wins a team his against their schedule when compared to what an average bubble team would be expected to have. At this minute, Iowa State is #28 in NET, #25 in KenPom, and #20 in WAB.

What is interesting is that if you look at the top 30 NET teams the average variation in rank to KenPom is 3.03 while the average variation, when compared to WAB, is 8.07. So yes, NET leans heavier to a performance-based metric than a resume-based metric.

My 2nd beef is the quad win discussion. If a team has a win over the #1 team in the country on the road that is a quad one win just as a win over the #30 team at home is and those are two entirely different things that should be considered at more of a variable level. WAB on the other hand accounts for the variable data more effectively.

Whether we’re ranking team quality or team resumes when we get to March and deal with the selection committee these numbers can matter but we’re mostly at the mercy of a bunch of humans and whatever they want to rationalize, so, God speed!

Since the committee switched to the NET rankings it has been used heavily though. Yes, they sort through some outliers but to say these rankings don't matter is incorrect. Also, the reason there has been a lot of Iowa and Iowa State comparisons is because they are two huge outliers on the opposite ends of the bell curve. In the past Q1 wins carried a lot of weight but this new formula has washed that out which is just wrong.
 

khaal53

Well-Known Member
SuperFanatic
SuperFanatic T2
Apr 13, 2006
2,894
618
113
41
Since the committee switched to the NET rankings it has been used heavily though. Yes, they sort through some outliers but to say these rankings don't matter is incorrect. Also, the reason there has been a lot of Iowa and Iowa State comparisons is because they are two huge outliers on the opposite ends of the bell curve. In the past Q1 wins carried a lot of weight but this new formula has washed that out which is just wrong.

I don't think I said they don't matter? Didn't mean to imply it either... as far as I know the formula is basically the same since it was introduced, but I could be incorrect there? I do tend to think that they will reward ISU for the Q1 wins.
 

BillBrasky4Cy

Well-Known Member
SuperFanatic
SuperFanatic T2
Dec 10, 2013
17,479
31,791
113
I don't think I said they don't matter? Didn't mean to imply it either... as far as I know the formula is basically the same since it was introduced, but I could be incorrect there? I do tend to think that they will reward ISU for the Q1 wins.

I wasn't necessarily saying you were implying that, there has just been a lot of that thrown around. Yes, the formula changed going into last year I believe. Clearly offensive efficiency has been favored actual wins and losses.
 

khaal53

Well-Known Member
SuperFanatic
SuperFanatic T2
Apr 13, 2006
2,894
618
113
41
I wasn't necessarily saying you were implying that, there has just been a lot of that thrown around. Yes, the formula changed going into last year I believe. Clearly offensive efficiency has been favored actual wins and losses.

My assumption is that it isn't just offensive efficiency but efficiency margin, though there is no way to know because the formula is so black box-y.
 

SEIOWA CLONE

Well-Known Member
Dec 19, 2018
6,793
6,989
113
63

*****

Rankings

The biggest thing when looking at rankings is we all need to understand what the ranking is for and how it is being calculated. There has been much consternation of Iowa State’s NET ranking but specifically their NET ranking in how it relates to Iowa. I get it, I personally don’t worry about it, but I get it.

In theory, the NET ranking is some unknown magical formula to homogenize performance rankings and resume rankings, but we really have no idea how. My two main thoughts on this…

It is clearly more tied to performance ratings that can be sorted out pretty quickly just by glancing at NET rankings compared to KenPom in contrast to how the NET rankings compare to a pure resume ranking like WAB (wins above bubble) is essentially the relative number of wins a team his against their schedule when compared to what an average bubble team would be expected to have. At this minute, Iowa State is #28 in NET, #25 in KenPom, and #20 in WAB.

What is interesting is that if you look at the top 30 NET teams the average variation in rank to KenPom is 3.03 while the average variation, when compared to WAB, is 8.07. So yes, NET leans heavier to a performance-based metric than a resume-based metric.

My 2nd beef is the quad win discussion. If a team has a win over the #1 team in the country on the road that is a quad one win just as a win over the #30 team at home is and those are two entirely different things that should be considered at more of a variable level. WAB on the other hand accounts for the variable data more effectively.

Whether we’re ranking team quality or team resumes when we get to March and deal with the selection committee these numbers can matter but we’re mostly at the mercy of a bunch of humans and whatever they want to rationalize, so, God speed!
ISU is not ranked #28 in NET, but #42, not even close to EIU's #19. Ken Pom has ISU is #41, while EIU is 18th.