NET Rankings are Flawed

BillBrasky4Cy

Well-Known Member
SuperFanatic
SuperFanatic T2
Dec 10, 2013
17,496
31,818
113
One dispute I have with your comparison, Wisconsin's Q1 winning % is better than ISU's, even though ISU has more Q1 wins.

Also if we compare all 4 quads, the separation of 19th & 27th doesn't seem too far out of whack.

UW — 5-2 .. 4-1 .. 1-0 .. 4-0
ISU — 6-4 .. 1-1 .. 0-0 .. 8-0

--> So, W better % at Q1, better record at Q2, fewer Q4 games (more favorable SOS at the low end). (I didn't assess the teams in between, so maybe should be closer to each other)

Granted, other comparisons we could make, like Iowa being 20th with its Q results, are more glaring/puzzling.

He's saying Wisconsin and Iowa State should both be higher in the rankings. Both teams play defense and aren't putting up gaudy numbers against dog sh!t teams.

Edit: he beat me to it :)
 

dahliaclone

Well-Known Member
Mar 4, 2007
16,192
25,056
113
Minneapolis
No I mean Wisconsin should be higher than #19, not that ISU should be above them. A 5-2 Q1 win ratio and an overall record 15-3 should have that team close to the top 10. Instead they are almost 20.

Right! And right behind Wisconsin at #20 is Iowa with an 0-4 Q1 record. Makes no sense that those two teams are even close to each other.
 
  • Like
Reactions: RezClone

MJ271

Well-Known Member
SuperFanatic
SuperFanatic T2
Aug 9, 2012
2,127
2,757
113
Atkins
What is obvious at this point is style points are extremely valuable to your NET ranking. Us at 6-4 in Q1 wins at #27, Wisconsin at #19 with 5-2 record in Q1 wins, etc.

I think the formula could be tweaked a bit to favor actual performance over efficiency metrics, but I'm totally fine with the NET ranking. I want an efficiency component. Don't look now but if we were using the old RPI metric, ISU would be #46 which is absolutely ludicrous. Baylor is #17 in the RPI, and Iona is #18. You get the idea.

NET isn't perfect, but no ranking will ever be perfect. It's MUCH better than what we were using.

One question I wonder about is if there are any adjustments to the efficiency metrics themselves. Does it count garbage time/routs? Does it account at all for strength of opponent? If a team's offensive efficiency is 130 against a bad defense and 90 against a good defense, with no adjustments that would average the same as if it were 110 against both, even though I think most would argue that the latter is better.
 

BillBrasky4Cy

Well-Known Member
SuperFanatic
SuperFanatic T2
Dec 10, 2013
17,496
31,818
113
One question I wonder about is if there are any adjustments to the efficiency metrics themselves. Does it count garbage time/routs? Does it account at all for strength of opponent? If a team's offensive efficiency is 130 against a bad defense and 90 against a good defense, with no adjustments that would average the same as if it were 110 against both, even though I think most would argue that the latter is better.

I think that's the issue. I don't think margin of win matters anymore but the offensive efficiency numbers definitely do. It appears to me that this metric is weighted way too heavily.
 

Cyclones125

Well-Known Member
May 3, 2016
271
409
63
Seems like most rating systems have ISU rated significantly lower than what I would expect based on our resume. My guess is some of our narrow-ish victories against very bad opponents early in the season are the culprit.
 

NorthCyd

Well-Known Member
SuperFanatic
SuperFanatic T2
Aug 22, 2011
21,232
35,937
113
Seems like most rating systems have ISU rated significantly lower than what I would expect based on our resume. My guess is some of our narrow-ish victories against very bad opponents early in the season are the culprit.
I think its this. Both scoring margin and NET efficiency are not our friend I would bet, and go a long way to explaining why we're ranked where we are at.
 
  • Winner
Reactions: CascadeClone

NorthCyd

Well-Known Member
SuperFanatic
SuperFanatic T2
Aug 22, 2011
21,232
35,937
113
I think that's the issue. I don't think margin of win matters anymore but the offensive efficiency numbers definitely do. It appears to me that this metric is weighted way too heavily.
Margin of victory matters, but its capped at 10. It's listed as one of the 5 factors.
 

cyfan92

Well-Known Member
Sep 20, 2011
8,243
13,106
113
Augusta National Golf Club
I think 5 factors go into the rankings. Not sure how they are weighted. My guess Iowa has offensive numbers that help them a bunch. We play gritty defense against great offensive teams. They play PSU, Nebby and Northwestern. It may get worse.

Scoring margin is an area where we struggle and Iowa succeeds. This pillow soft non-con and B1G schedule provide plenty opportunity for large victories.
 

GoHawks

Well-Known Member
Jul 12, 2009
3,828
2,253
113
It's not Iowa. It's the NET rankings. There is a major flaw somewhere. 0-4 vs 6-4?
I definitely think it is flawed first and foremost. My eyes, the schedule, and the score between the 2 tells me Iowa State is better. They should be higher than Iowa in the rankings. I think it's unfair though that people say the net is flawed but then point out it's flawed by showing their quad wins based off their formula because I think that is also flawed. Kansas state could have a better record in the SEC but worse NET ranking they benefit too much by losing to good teams. The bad rises because of the good unjustly. It happened in the Big Ten last year. In the end the Big Ten was not as good as what the NET or committee told us.
 
Last edited:

BillBrasky4Cy

Well-Known Member
SuperFanatic
SuperFanatic T2
Dec 10, 2013
17,496
31,818
113
Scoring margin is an area where we struggle and Iowa succeeds. This pillow soft non-con and B1G schedule provide plenty opportunity for large victories.

On paper Iowa is a really efficient team but those numbers are certainly skewed especially when you compare their results against good defensive squads. At the end of the day who you beat and who you lose to has to matter.
 
  • Agree
Reactions: RezClone

knowlesjam

Well-Known Member
Oct 21, 2012
4,325
4,776
113
Papillion, NE
So much is weighted on the offensive efficiency numbers, along with the margin of victory in wins/losses, along with where the wins are coming from. Here are the latest for the teams...

Iowa: 4 (OE) 94 (DE) +13.7 (MoV)
Wisky: 60 (OE) 121 (DE) +5.8 (MoV)
ISU: 180 (OE) 10 (DE) +8.7 (MoV)

Iowa's efficiency numbers are very good, with the DE number improving by more than 50 over the past 2 weeks, while the ISU OE number has dropped nearly 30. The quality wins are what is saving ISU, while Iowa is being hurt. Wisky's efficiency numbers are a bit better than ISU, but they have 1 less quad 1 win. Given how net works, they teams are properly placed...it really is a result of the calculation.
 

BillBrasky4Cy

Well-Known Member
SuperFanatic
SuperFanatic T2
Dec 10, 2013
17,496
31,818
113
I definitely think it is flawed first and foremost. My eyes, the schedule, and the score between the 2 tells me Iowa State is better. They should be higher than Iowa in the rankings. I think it's unfair though that people say the net is flawed but then point out it's flawed by showing their quad wins based off their formula because I think that is also flawed. Kansas state could have a better record in the SEC but they benefit too much by losing to good teams. The bad rises because of the good unjustly. It happened in the Big Ten last year. In the end the Big Ten was not as good as what the NET or committee told us.

I can't say that's the case this year though. How can you say K-State is benefiting, they are #68 in NET. Last year the B1G was extremely top heavy and everyone knew that. This year the Big 12 has 10 teams that can beat you night in night out. It's a freakin grind and that should carry some weight.
 

knowlesjam

Well-Known Member
Oct 21, 2012
4,325
4,776
113
Papillion, NE
To follow on the previous Iowa, Wisky, ISU stats, Houston is top 10 in OE, DE, and MoV. Normally they would be an easy #1 in NET rankings, but they take a hit due to their strength of schedule and zero quad 1 wins.
 

GoHawks

Well-Known Member
Jul 12, 2009
3,828
2,253
113
I can't say that's the case this year though. How can you say K-State is benefiting, they are #68 in NET. Last year the B1G was extremely top heavy and everyone knew that. This year the Big 12 has 10 teams that can beat you night in night out. It's a freakin grind and that should carry some weight.
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.