This is why Pre-season rankings are a joke

[texascpa]

Kentucky is now 5-4 after being destroyed by Gonzaga last night. Their other three losses were all to ranked teams. Their wins were against non-Power 5 (including Big East) teams. One of those losses was a blowout and the other was by 8, yet they were 18th coming into last night's game. Essentially, their season tracks exactly with Cuse's up until last night and they were STILL in the Top 20!

 

[sufandu]

Especially in an era with so much change year to year. Thankfully, not only are they a joke, they're also meaningless beyond fan and media attention.

 

[Cusefan0307]

UK is overrated, but they have been dealing with 3 players having injuries although one did come back last night.

 

[Charbroil]

The new rankings don’t come out until Monday. Kentucky was 5-2 when the last rankings came out, which is why they are still in the top 20. They won’t be ranked Monday now that they’ve lost 2 more games, 1 in embarrassing fashion.

They fell to the 39th in the NET. Although that’s still too high, the rankings and metrics are catching up to Kentucky based on their performance thus far.

 

[Sbucketball]

The Kentucky fanbase has turned on Pope. They want to fire him.

 

[MadNY3]

Yep. But I’d add that isn’t is ridiculous to be staring down KenPom rankings every other day in November and early December. And yet, we do

 

[DomeHolmes]

The new rankings don’t come out until Monday. Kentucky was 5-2 when the last rankings came out, which is why they are still in the top 20. They won’t be ranked Monday now that they’ve lost 2 more games, 1 in embarrassing fashion.

They fell to the 39th in the NET. Although that’s still too high, the rankings and metrics are catching up to Kentucky based on their performance thus far.

This also illustrates what a joke NET is. They have mostly the same resume as us (with a bad loss to top 15 team, and 4 cupcake wins). But we have a top 15 win and 3 top 25 losses and they have a 4 top 25 losses.

They are 39. We are 70.

As the flaws in the net rankings become more and more apparent, hopefully the committee starts to refer back to “who did you play? Who did you beat?”

I saw the same thing with RPI. The first few years, it was “the” metric and was not questioned by the committee. Then the flaws were too many to ignore. Same with NET. last year there was more emphasis on quality of schedule according to the committee.

 

[Cusefan0307]

This also illustrates what a joke NET is. They have mostly the same resume as us (with a bad loss to top 15 team, and 4 cupcake wins). But we have a top 15 win and 3 top 25 losses and they have a 4 top 25 losses.

They are 39. We are 70.

As the flaws in the net rankings become more and more apparent, hopefully the committee starts to refer back to “who did you play? Who did you beat?”

I saw the same thing with RPI. The first few years, it was “the” metric and was not questioned by the committee. Then the flaws were too many to ignore. Same with NET. last year there was more emphasis on quality of schedule according to the committee.

Barely beating Monmouth and getting blasted by ISU is basically why we are 70th. UK destroyed every cupcake. I agree they’ve sucked otherwise.

 

[Orangezoo]

Barely beating Monmouth and getting blasted by ISU is basically why we are 70th. UK destroyed every cupcake. I agree they’ve sucked otherwise.

Which means if we don’t have another Monmouth like performance we should rack up some big wins margin wise and have a much better number going into the Clemson game.

 

[DomeHolmes]

Barely beating Monmouth and getting blasted by ISU is basically why we are 70th. UK destroyed every cupcake. I agree they’ve sucked otherwise.

Kentucky got blasted worse. And we have a quality win. I’m not saying by the “metrics” we shouldn’t be 70th, I’m saying the metrics are stupid.

 

[Orangezoo]

Kentucky got blasted worse. And we have a quality win. I’m not saying by the “metrics” we shouldn’t be 70th, I’m saying the metrics are stupid.

They are and assuming we take care of business the next 5 games the margins will send us back up those rankings quickly against a bunch of weak teams

 

[jncuse]

Good thing about basketball compared to football is that preseason biases generally wash away
- Teams play 31 games
- Power Conferences play 200-230 games against other conferences, and about 70-80 against each other or quality games.

The metrics may not be perfect, but they don't carry bias (within P5 anyway). The tournament is shaped more by what actually happened and then measured.

With so few head to head OOC in football, biases can linger because small samples will be ignored.

 

[jncuse]

This also illustrates what a joke NET is. They have mostly the same resume as us (with a bad loss to top 15 team, and 4 cupcake wins). But we have a top 15 win and 3 top 25 losses and they have a 4 top 25 losses.

They are 39. We are 70.

As the flaws in the net rankings become more and more apparent, hopefully the committee starts to refer back to “who did you play? Who did you beat?”

I saw the same thing with RPI. The first few years, it was “the” metric and was not questioned by the committee. Then the flaws were too many to ignore. Same with NET. last year there was more emphasis on quality of schedule according to the committee.

For the most part teams are not getting selected based on their NET. They are getting in based on win/loss, quality of wins.

Usually that correlated fairly well but not always. And that is why we have seen teams with NET's in the 20's such as in 2024 miss the NCAA Tournament.

We do well in the ACC, and the NET should get into the reasonable range. And it will then come down to quality of wins and losses.

 

[jncuse]

Which means if we don’t have another Monmouth like performance we should rack up some big wins margin wise and have a much better number going into the Clemson game.

Yes / No .. it depends I guess.

It's still margin relative to opponent. So will need to get past expected levels and avoid those single digit wins.

 

[jncuse]

Kentucky got blasted worse. And we have a quality win. I’m not saying by the “metrics” we shouldn’t be 70th, I’m saying the metrics are stupid.

The NET metrics are a little flawed, imperfect, but I don't think stupid.

NET is still ultimately created to measure teams after 31 games. After 6 or 7 games, one or two outlier games can really skew a TEAM's net ... after 31 games those outlier games won't have that huge an impact.

But it's out there because people want to see them.

 

[Yeti]

UK is overrated, but they have been dealing with 3 players having injuries although one did come back last night.

I mean.. same. Donnie has missed a bunch of games. JJ missed games.

 

[DomeHolmes]

The NET metrics are a little flawed, imperfect, but I don't think stupid.

NET is still ultimately created to measure teams after 31 games. After 6 or 7 games, one or two outlier games can really skew a TEAM's net ... after 31 games those outlier games won't have that huge an impact.

But it's out there because people want to see them.

I would say more than a little flawed. For example, this week:

Kentucky got destroyed by Gonzaga in what was essentially a home game even though it’s “neutral” and they actually moved up from 39 to 30! This defies every so-called rule of how you move up and down in the NET.

Gonzaga got destroyed by 40 by Michigan and their NET changed from 2 to 4! 2 spots.

Purdue got blown out at home by Iowa State and they went from 4 to 9.

We got blown out at a neutral site on 20 hours rest by the same Iowa State team, and we dropped like 30 or 40 points?

We actually dropped more places today because Tennessee lost yesterday, than Gonzaga or Purdue drops for getting destroyed in games they actually played.

I maintain that it’s a stupid system that’s flawed beyond belief.

 

[Fly Rodder]

I maintain that it’s a stupid system that’s flawed beyond belief.

Should be super easy for you to put together a better system.

I will wait.

 

[ScriptCuse]

I've been a STAUNCH believer there shouldn't be rankings until week 3 or 4 in Football or Basketball for many reasons. Most of them being it sets an unfair ranking that teams have to climb that were arbitrary from a previous season to begin with and that is incredibly unequitable. I feel very strongly about this

 

[General20]

Should be super easy for you to put together a better system.

I will wait.

It really is quite easy. Institute a commissioner who sets a schedule for every team. Stop teams from making their own schedule and trying to game the system. Create a somewhat level playing field. Every team does not have to have a schedule of equal difficulty, it just needs to be fair enough and unbiased. Then the teams with the best records make the playoffs. Basically just do what every other sport on earth does.

 

[jncuse]

I would say more than a little flawed. For example, this week:

Kentucky got destroyed by Gonzaga in what was essentially a home game even though it’s “neutral” and they actually moved up from 39 to 30! This defies every so-called rule of how you move up and down in the NET.

Gonzaga got destroyed by 40 by Michigan and their NET changed from 2 to 4! 2 spots.

Purdue got blown out at home by Iowa State and they went from 4 to 9.

We got blown out at a neutral site on 20 hours rest by the same Iowa State team, and we dropped like 30 or 40 points?

We actually dropped more places today because Tennessee lost yesterday, than Gonzaga or Purdue drops for getting destroyed in games they actually played.

I maintain that it’s a stupid system that’s flawed beyond belief.

I'm not sure where you are getting some of your numbers from, but unfortunately they are not accurate.

#1. Kentucky was #15 in NET at the start of last week when they were first released. They have fell down to #31. They never increased in the rankings... not sure where you are getting their rise from #39 to #30 after a loss. For sure that would have been crazy - but it never actually happened. (I will comment on their #31 ranking in a separate post)

#2. When Gonzaga played Michigan the NET had yet to be even be released.
When Syracuse played Iowa ST the NET had yet to be revealed.
I'm not sure where you are getting how much they rise and fell from.

#3. Regarding Purdue's fall versus our fall. Think of a Standard Distribution Curve. Gaps between teams at the end of the curves are longer than those towards the middle.

Using KP.
From #1-#10, teams ADJ EM's range from 27 to 36 (a difference of nine)

From #40-93, teams range from 8 to 17 (a difference of nine). When you are closer to the middle teams are a bit more bunched up.

Purdue was hurt just as much in terms of their "NET Score" after losing to Iowa St as we did losing to Iowa St. But we are in the more packed area towards the middle of the curve, with many more teams with a closer score to us due the factor noted above. So our rank will fall more spots in a bad result. That being said we can also rise much farther when we do good things than Purdue. I have to use "NET Score" because they don't put a number for NET like KP. But you can observe it in KP.

There is also the sample factor... we are still only at 8 games so extreme results still get magnified more than they would after 31 games.

#4. The differences between teams ranked in any range of 10, as we get towards the middle are not that big. You will move a few spots back and forth as teams play around you especially when its still a sample of games played. (8 of 31). For example in KP, team #59 is 13.8, team #67 is 13.2... teams are going to move around there at a whim.

Again I'm not a 100% defender of the NET, but i try to take a balanced view of understanding its impacts, its "good", its weaknesses (Q4 margin outliers). Philosophically its more important that conference mates take advantage of their NET than a team does its own.

 

[jncuse]

It really is quite easy. Institute a commissioner who sets a schedule for every team. Stop teams from making their own schedule and trying to game the system. Create a somewhat level playing field. Every team does not have to have a schedule of equal difficulty, it just needs to be fair enough and unbiased. Then the teams with the best records make the playoffs. Basically just do what every other sport on earth does.

The solution is "easy", if you ignore the inability or desire to implement it.

If there was a balanced schedule that would be great. But that will never happen - so until then you need a system to be able to rank not 10, not 20 or not even all the P5 teams. But adequately rank 365 teams and its impossible to rank 365 teams without brining margin into play.

It would help if the P5 teams didn't avoid Q3 teams like the plague, and actually play them instead of running to play the worst in the NCAA.

 

[DoctahLexus]

Should be super easy for you to put together a better system.

I will wait.

Easy. Completely disregard the score of games. You play to win the game. Not to win by 20 plus or whatever it is. So many factors can go into the final score. Did you win or did you not? Keep it simple.

 

[Blanko97]

It really is quite easy. Institute a commissioner who sets a schedule for every team. Stop teams from making their own schedule and trying to game the system. Create a somewhat level playing field. Every team does not have to have a schedule of equal difficulty, it just needs to be fair enough and unbiased. Then the teams with the best records make the playoffs. Basically just do what every other sport on earth does.

360 teams makes that impossible.

 

[jncuse]

Easy. Completely disregard the score of games. You play to win the game. Not to win by 20 plus or whatever it is. So many factors can go into the final score. Did you win or did you not? Keep it simple.

We had the RPI that tried to do that before and it spit out a lot of junk. Its hard to measure 365 teams without bringing in margin into play, and things like KP and NET are probably ranking teams better at the end. But its still flawed.

I have a few potential solutions for a revamped NET that I think would be better based on what I have learned as the flaws and strengths of various systems. I'm intrigued by Bartorvik's Quality of Record metric that only focuses on W-L, but even that one has to have a margin based system influencing it to a degree from behind it to not spit out total jibberish.. but that's OK if sone properly.

I'll expand on those tomorrow or tonight.