NET Rankings are Flawed

[knowlesjam]

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.

Huh, as much as we like to smack Fran around, maybe he has broke the code and is gaming the system. Then again, maybe that is according him too much credit...regardless, Iowa is hitting the sweet spot on how to maximize the Offensive Efficiency number.

 

[CascadeClone]

Huh, as much as we like to smack Fran around, maybe he has broke the code and is gaming the system. Then again, maybe that is according him too much credit...regardless, Iowa is hitting the sweet spot on how to maximize the Offensive Efficiency number.

Well, he knows a thing or two about CHEATING MOTHER *******

 

[AuH2O]

Huh, as much as we like to smack Fran around, maybe he has broke the code and is gaming the system. Then again, maybe that is according him too much credit...regardless, Iowa is hitting the sweet spot on how to maximize the Offensive Efficiency number.

Lots of those years they were something of a KenPom darling also despite not being that good of a team. I just attributed it to them running it up big in wins and managing to piss themselves in their close games. Which kind of makes sense when you have a coach that loses his **** so easily.

 

[Beyerball]

Posted this in another thread...

Someone posted the NET Rating formula but i think it was the outdated one.

They changed the NET rating system from 5 things to 2 things in 2020.

With the changes announced in May 2020, the NET will no longer use winning percentage, adjusted winning percentage and scoring margin..

1- Adjusted net efficiency...O-D efficiency etc..
2- Team Value Index..Wins and losses..location etc..

ISU looks great with #2...like Top 10 great. But our O-D efficiency sucks. Our entire non conf was beating some bad teams but like 60-53..every night.

Iowa on other hand beat those same bad teams like 90-60..

Iowa continues to score alot on offense but they are playing Quad 2 teams mostly who are also not all that good offensively so it skews Iowa's D.

Bottom line is the way Iowa plays is favored far more by NET algorithms than most teams.

Edit: look at last year...although bit anomoly with Covid.

Last 4 in :
The NET ranking below is after NCAA tourney over...

Drake- 43rd NET...1-3 Quad 1
Mich ST- 76th NET....5-11 Quad 1
UCLA- 15th NET...1-7 Quad 1...? I think. Ended 5-7 obviously bc made FF.

UCLA was like 17-10 going into selection day? with (1) Quad 1 win..I think.

My Point is since the NET created I don't think a high major team has made NCAA tourney without a Quad 1 win?

I could be wrong though.

 

[FriendlySpartan]

Posted this in another thread...

Someone posted the NET Rating formula but i think it was the outdated one.

They changed the NET rating system from 5 things to 2 things in 2020.

With the changes announced in May 2020, the NET will no longer use winning percentage, adjusted winning percentage and scoring margin..

1- Adjusted net efficiency...O-D efficiency etc..
2- Team Value Index..Wins and losses..location etc..

ISU looks great with #2...like Top 10 great. But our O-D efficiency sucks. Our entire non conf was beating some bad teams but like 60-53..every night.

Iowa on other hand beat those same bad teams like 90-60..

Iowa continues to score alot on offense but they are playing Quad 2 teams mostly who are also not all that good offensively so it skews Iowa's D.

Bottom line is the way Iowa plays is favored far more by NET algorithms than most teams.

Edit: look at last year...although bit anomoly with Covid.

Last 4 in :
The NET ranking below is after NCAA tourney over...

Drake- 43rd NET...1-3 Quad 1
Mich ST- 76th NET....5-11 Quad 1
UCLA- 15th NET...1-7 Quad 1...? I think. Ended 5-7 obviously bc made FF.

UCLA was like 17-10 going into selection day? with (1) Quad 1 win..I think.

My Point is since the NET created I don't think a high major team has made NCAA tourney without a Quad 1 win?

I could be wrong though.

Also I dont think sparty gets in last year without the history. Even though we had a good end to the season and were crushing UCLA for awhile we probably didnt deserve to get a bid.

 

[cykadelic2]

Why does NET continue to devalue Iowa States resume? I know there has been talk about offensive efficiency being a significant weight but at some point who you beat and who you lose to has to matter. Iowa State now has SIX Q1 wins and zero bad losses. I just don't get it. How are these two resumes even comparable? Yet Iowa is SEVEN spots higher than Iowa State? I really don't get it.

View attachment 94925

NET is primarily used as a sorting tool for the Selection Committee to ID W-Ls for each of the Quads. Which is why ISU's 6 Q1 wins vs TOE's zero are far more relevant than the rankings themselves and which is why ISU is currently seeded higher than TOE by rational bracketologists.

If you want to use a singular ranking metric to analyze potential seeding, use ESPN's SOR. It has correlated most closely with final Committee seeding and selection over the past couple of seasons. ISU is #17 there and TOE is #38.

 

[EnkAMania]

NET is primarily used as a sorting tool for the Selection Committee to ID W-Ls for each of the Quads. Which is why ISU's 6 Q1 wins vs TOE's zero are far more relevant than the rankings themselves and which is why ISU is currently seeded higher than TOE by rational bracketologists.

If you want to use a singular ranking metric to analyze potential seeding, use ESPN's SOR. It has correlated most closely with final Committee seeding and selection over the past couple of seasons. ISU is #17 there and TOE is #38.

Agreed. The committee doesn't blindly follow the NET ranking. Last year #6 NET ranked USC was a 6th seed and #10 NET ranked Loyola was an 8th seed.

 

[isufbcurt]

NET is primarily used as a sorting tool for the Selection Committee to ID W-Ls for each of the Quads. Which is why ISU's 6 Q1 wins vs TOE's zero are far more relevant than the rankings themselves and which is why ISU is currently seeded higher than TOE by rational bracketologists.

If you want to use a singular ranking metric to analyze potential seeding, use ESPN's SOR. It has correlated most closely with final Committee seeding and selection over the past couple of seasons. ISU is #17 there and TOE is #38.

Didn't ESPN use the Net rating as the official ranking in it's first year?

I remember they did a weekly unveiling for it kind of like the football playoff rankings.

 

[cykadelic2]

Didn't ESPN use the Net rating as the official ranking in it's first year?

I remember they did a weekly unveiling for it kind of like the football playoff rankings.

IIRC, that weekly unveiling was provided by the Committee to media outlets (e.g. ESPN) and based on NET Quad results and other analytics they utilize for the final bracket.

 

[CoachHines3]

Need @khaal53 to explain

 

[qwerty]

NET is primarily used as a sorting tool for the Selection Committee to ID W-Ls for each of the Quads. Which is why ISU's 6 Q1 wins vs TOE's zero are far more relevant than the rankings themselves and which is why ISU is currently seeded higher than TOE by rational bracketologists.

If you want to use a singular ranking metric to analyze potential seeding, use ESPN's SOR. It has correlated most closely with final Committee seeding and selection over the past couple of seasons. ISU is #17 there and TOE is #38.

I feel dirty but just glanced at HN and someone there thinks Iowa should be a 6 seed but will get screwed by having zero Q1 wins and end up a 8-9 seed.

 

[QBEagles]

With the changes announced in May 2020, the NET will no longer use winning percentage, adjusted winning percentage and scoring margin..

1- Adjusted net efficiency...O-D efficiency etc..
2- Team Value Index..Wins and losses..location etc..

The problem is that NET is trying to measure how good a team is supposed to be (efficiency) and how good their resume is (Ws and Ls) with the same ranking when they're two different things. The result is just some arbitrary rankings based on their (apparently efficiency heavy) weighting of the two.

They should just keep the two separate, use an average of KenPom-type ratings for efficiency, and use strength of record metrics (not just ESPN's) to measure resume. We'd at least know what each ranking is supposed to mean then, and have an idea on how much the committee is weighting each.

 

[SEIOWA CLONE]

The problem with the NET rankings is it rewards teams for playing Quad 1 schools, more than actually beating them. Any system that has EIU ahead of ISU with zero Quad 1 wins is flawed.

We are past the days when a 20-win season got you into the tournament, we now need a system that actually rewards you for not only playing a difficult non-conference season, but rewards teams that actually beat those team.

 

[awd4cy]

So Iowa dropped only 2 spots for getting smacked at home by Purdue.

 

[NorthCyd]

People can complain about Iowa's ranking in NET all they want, but they are ranked about the same in Torvik and Kenpom as well. It seems most power rankings like there numbers for whatever reason.

 

[isufbcurt]

People can complain about Iowa's ranking in NET all they want, but they are ranked about the same in Torvik and Kenpom as well. It seems most power rankings like there numbers for whatever reason.

Scoring points but not winning quality games seems like a ****** metric

 

[NoCreativity]

I feel dirty but just glanced at HN and someone there thinks Iowa should be a 6 seed but will get screwed by having zero Q1 wins and end up a 8-9 seed.

They also thought Iowas defense would be alot better this year because Perkins and Ulis were alot more athletic guards.

 

[awd4cy]

People can complain about Iowa's ranking in NET all they want, but they are ranked about the same in Torvik and Kenpom as well. It seems most power rankings like there numbers for whatever reason.

And yet, it still makes no sense.

 

[dahliaclone]

People can complain about Iowa's ranking in NET all they want, but they are ranked about the same in Torvik and Kenpom as well. It seems most power rankings like there numbers for whatever reason.

It's not 'whatever reason' it's their OE. I thought this was already discussed ad nauseam lol

 

[NorthCyd]

It's not 'whatever reason' it's their OE. I thought this was already discussed ad nauseam lol

Hey, thay was just discussion about NET, not Kenpom and Torvik. So put that in your pipe and smoke it!