2023-24 NET ratings thread

[SmellyCat]

Lunardi has us as a 9 seed.

Men's Bracketology: The fact that UConn even has a chance to three-peat is amazing

The last time anyone saw an NCAA basketball three-peat the heroes were John Wooden, Bill Walton and the UCLA Bruins. Enter the current UConn Huskies.

www.espn.com

As much as I hate to say it, Purdue running the table is probably what we have to root for. With only two losses on their resume, our two games against them speak volumes and more than make up for any Chicago States. Every time a Nebraska comes along to smack Purdue, it lessens the impact of how great our games were. Most wonks have NU safely in the tourney, and it isn't because they beat Ohio State and Maryland and Nebraska.

 

[PurpleWhiteBoy]

After a helpful win over Nebraska, I've got Northwestern as the #30 team in the country or the #36 team in the country, based on my two methods. #30 on quality of wins/losses and #36 based on "points" (aka ""how well you run up the score"). That's essentially where we were before we lost at Minnesota.

The NET is an inferior algorithm.

 

[Styre]

After two expected wins, not much change for NU at all, sitting currently at #56.

QUAD 1 (4-5)
H #2 Purdue - W
A #2 Purdue - L
H #14 Illinois - W
A #14 Illinois - L
H #18 Dayton - W
A #21 Wisconsin - L
H #23 Michigan State - W
N #36 Mississippi State - L
A #53 Nebraska - L

QUAD 2 (3-1)
H #53 Nebraska - W
H #71 Ohio State - W
A #85 Minnesota - L
A #92 Penn State - W

QUAD 3 (4-0)
H #81 Maryland - W
H #92 Penn State - W
N #127 Arizona State - W
N #179 Rhode Island - W

QUAD 4 (6-1)
H #279 Binghamton - W
H #283 Northern Illinois - W
H #287 Jackson State - W
H #290 Chicago State - L
H #294 Western Michigan - W
A #316 DePaul - W
H #360 Detroit Mercy - W

This week features two Quad 2 road games at Rutgers and Indiana.

 

[olshin]

Of note, we dropped a spot for beating Penn State, while Penn State went up 6 spots for losing at Northwestern. NET is a complete joke.

 

[olshin]

Also of note ... it's not just that we have 7 Q4 games (and yikes lost 1) ... it's that these are like the Q4 OF Q4 (Q16?) games that just destroy us in metrics. 7 games 279 or below??

That really just can't happen. If those games were against teams 175-225 we'd probably be much higher, even if we went 5-2 instead of 6-1.

 

[UpsetAlert]

Of note, we dropped a spot for beating Penn State, while Penn State went up 6 spots for losing at Northwestern. NET is a complete joke.

Michigan State is example A. Have a couple good wins but biggest asset is they lose to good teams by not a lot of points. Maddening.

 

[curdog]

Michigan State is example A. Have a couple good wins but biggest asset is they lose to good teams by not a lot of points. Maddening.

Random Shitty teams ranked higher than NU:
-St. John’s 14-10, 2-7 quad 1
-Xavier 13-11, 2-8 quad 1
-MSU 15-9, 3-7 quad 1
-wake forest 16-7, 0-3 quad 1

How in the world did Net become a thing? It’s just a broken metric.

 

[hdhntr1]

Random Shitty teams ranked higher than NU:
-St. John’s 14-10, 2-7 quad 1
-Xavier 13-11, 2-8 quad 1
-MSU 15-9, 3-7 quad 1
-wake forest 16-7, 0-3 quad 1

How in the world did Net become a thing? It’s just a broken metric.

Not just a thing but a thing that is important as it has become

 

[PurpleWhiteBoy]

Random Shitty teams ranked higher than NU:
-St. John’s 14-10, 2-7 quad 1
-Xavier 13-11, 2-8 quad 1
-MSU 15-9, 3-7 quad 1
-wake forest 16-7, 0-3 quad 1

How in the world did Net become a thing? It’s just a broken metric.

Specifically looking at the MSU Spartans... Their 3 biggest outperformances of the season are a 24 point blowout of Baylor, a 31 point blowout of Penn State and a 44 point annihilation of Stony Brook... all in East Lansing.
That Stony Brook win is the most impactful score on their schedule.

On the negative side, they got drilled twice by Wisconsin (losing by 13 and 15) and lost to NU by 14. Those are their worst performances of the season.

All of the 3 big wins are about 50% more positive for their NET rating than the 3 losses are negative. That little oddity probably gets them 8-10 spots higher than they should be.

 

[BosCat]

Just a reminder about the great aggregating site: http://www.bracketmatrix.com/

It hasn't been updated since Sat (as of 2:30ET) -- but has links to a lot of sites that you can check individually.

 

[SimpsonElmwood]

Specifically looking at the MSU Spartans... Their 3 biggest outperformances of the season are a 24 point blowout of Baylor, a 31 point blowout of Penn State and a 44 point annihilation of Stony Brook... all in East Lansing.
That Stony Brook win is the most impactful score on their schedule.

On the negative side, they got drilled twice by Wisconsin (losing by 13 and 15) and lost to NU by 14. Those are their worst performances of the season.

All of the 3 big wins are about 50% more positive for their NET rating than the 3 losses are negative. That little oddity probably gets them 8-10 spots higher than they should be.

They also have only 4 Q4 games and one of them (Stony Brook) has a net of 212. Cats have 7, with one loss, and the best one is Binghamton at 279.

 

[PurpleWhiteBoy]

They also have only 4 Q4 games and one of them (Stony Brook) has a net of 212. Cats have 7, with one loss, and the best one is Binghamton at 279.

they have definitely played a far more difficult schedule than NU.
With no ugly underperformances.

I just don't think creaming inferior teams should be rewarded. Nor should comfortable, lazy wins against bad teams be penalized.

KenPom and the NET make the assumption that every point in the margin of victory is equal.

Nobody believes that.

Those methods also make the implicit assumption that each team is playing at 100% effort every time.

Also blatantly false.

And the capability of the team is constant throughout the season... Obviously not true.

So... a system based on false assumptions is going to produce... the truth?

 

[phatcat]

Also of note ... it's not just that we have 7 Q4 games (and yikes lost 1) ... it's that these are like the Q4 OF Q4 (Q16?) games that just destroy us in metrics. 7 games 279 or below??

That really just can't happen. If those games were against teams 175-225 we'd probably be much higher, even if we went 5-2 instead of 6-1.

It's hard to gauge when scheduling who will be Q4 and who will be Q16. Who knew Detroit and DePaul would be THIS bad? Or directional Michigan? NIU?

 

[Jonny2TheP]

they have definitely played a far more difficult schedule than NU.
With no ugly underperformances.

I just don't think creaming inferior teams should be rewarded. Nor should comfortable, lazy wins against bad teams be penalized.

KenPom and the NET make the assumption that every point in the margin of victory is equal.

Nobody believes that.

Those methods also make the implicit assumption that each team is playing at 100% effort every time.

Also blatantly false.

And the capability of the team is constant throughout the season... Obviously not true.

So... a system based on false assumptions is going to produce... the truth?

This is pretty well said. A few key things I think that make most metrics flawed and potential solutions:

- When you play a team really needs to be taken into consideration. Say a star player gets hurt on a team and the team struggles after the injury occurs. A team that gets a win against that team before the injury should get more credit than one that does after. But in all metrics that I'm aware of, that is not a factor.

- Margin of victory should be weighted less significantly than it is in some of these metrics. I'm not saying it should be irrelevant, but some of these models heavily weigh the margin of victory. For many of the reasons you said above, this should not be weighed heavily.

- Someone on these boards recently said that a team's best game and a team's worst game should be removed from any metrics (not completely erased when considering a team, but just from the metrics). I think this is a fantastic idea that should be adopted.

 

[SmellyCat]

- When you play a team really needs to be taken into consideration. Say a star player gets hurt on a team and the team struggles after the injury occurs. A team that gets a win against that team before the injury should get more credit than one that does after. But in all metrics that I'm aware of, that is not a factor.

I like your point, but just to play devil's advocate...how is it determined who a "star" is? Is Ty Berry a "star?" Does NU get as much credit for gutting out a win over Penn State as Illinois did when they won games without Shannon? Is Barnhizer a star? He hasn't missed much time at all, but he played through an injury and the computers aren't going to pick that up because it didn't affect his minutes played, even as it affected his effectiveness.

Illinois destroyed NU without Shannon, so maybe they aren't so bad when he doesn't play. But then someone else beat them during Shannon's absence (I think...I'm not looking at their schedule right now). Do they get less credit for beating a Shannon-less Illinois because Illinois proved it could destroy a good NU without him?

I guess my point is: this is what the committee is for. They'll presumably seed Illinois appropriately next month and provide their own weights on various games and other low sample sizes. Maybe they'll know enough to say that Chicago State beat NU when a couple of guy guys were hurt (probably not), or that they'll give NU less credit for beating Illinois because it was in overtime, I have no idea.

All these metrics are fun to look at and use, but ultimately humans have to do the seeding. They'll make mistakes and bring their own biases in, but they're better than most in determining how much weight to give to players being in or out or hurt or whatever.

 

[elgatoloco]

Here's an insane aggregated page of computer rankings from Massey. NU is 42, about where it should be:

Massey Ratings - College Basketball Rankings

College Basketball computer rankings and polls, weekly composite

masseyratings.com

 

[FeralFelidae]

This is pretty well said. A few key things I think that make most metrics flawed and potential solutions:

- When you play a team really needs to be taken into consideration. Say a star player gets hurt on a team and the team struggles after the injury occurs. A team that gets a win against that team before the injury should get more credit than one that does after. But in all metrics that I'm aware of, that is not a factor.

- Margin of victory should be weighted less significantly than it is in some of these metrics. I'm not saying it should be irrelevant, but some of these models heavily weigh the margin of victory. For many of the reasons you said above, this should not be weighed heavily.

- Someone on these boards recently said that a team's best game and a team's worst game should be removed from any metrics (not completely erased when considering a team, but just from the metrics). I think this is a fantastic idea that should be adopted.

The thought of "throwing out" our Purdue win breaks my heart. No.

 

[PurpleWhiteBoy]

The thought of "throwing out" our Purdue win breaks my heart. No.

Then you're in luck!
Our worst three games of the season were
Chicago State (-25)
@Illinois (-22.5)
Western Michigan (-19)

Our three best games were
Ohio State (+18)
Michigan State (+13)
Arizona State (+11)

The Purdue games are both pretty normal results for us. About +10.

 

[phatcat]

Then you're in luck!
Our worst three games of the season were
Chicago State (-25)
@Illinois (-22.5)
Western Michigan (-19)

Our three best games were
Ohio State (+18)
Michigan State (+13)
Arizona State (+11)

The Purdue games are both pretty normal results for us. About +10.

So you are saying a win vs WMU was so unimpressive that it counteracts our PU win and ASU win combined. That's ridiculous

 

[PurpleWhiteBoy]

So you are saying a win vs WMU was so unimpressive that it counteracts our PU win and ASU win combined. That's ridiculous

Yes that is essentially how it works.
The way to get a bogus good rating is to have big positive outliers while controlling your negative outliers.
If you have big negative outliers and don't run up the score on bad teams, the NET will underrate you.
(This is one reason I think "conference record" should be important in picking teams - and the selection committee obviously ignores that. I'd give each of the Power 5 or 6 at least 3 teams, no matter what - based solely on conference standings - as long as they have a winning record. Take some of the BS out of it)

 

[SmellyCat]

This is one reason I think "conference record" should be important in picking teams - and the selection committee obviously ignores that. I'd give each of the Power 5 or 6 at least 3 teams, no matter what - based solely on conference standings

I suppose as long as you're saying just three that get automatic invites, I suppose that's ok, but generally big conferences get that much anyway, and the only difference would be like a weird fluke of a team having a strong conference record but a bad overall record that would otherwise keep them out of the tourney.

Still, seeing someone like Iowa or Michigan (obviously not talking about this year) get a bid instead of someone like Grand Canyon kind of takes the fun out of the tourney.

 

[phatcat]

Yes that is essentially how it works.
The way to get a bogus good rating is to have big positive outliers while controlling your negative outliers.
If you have big negative outliers and don't run up the score on bad teams, the NET will underrate you.
(This is one reason I think "conference record" should be important in picking teams - and the selection committee obviously ignores that. I'd give each of the Power 5 or 6 at least 3 teams, no matter what - based solely on conference standings - as long as they have a winning record. Take some of the BS out of it)

Thank goodness football doesn't work like that. We won 8 games with a negative point differential. We would have been passed over for a bowl

 

[Styre]

A 1-1 road trip has NU exactly where they were a week ago, #56. The home win over Maryland has become a Quad 2 win.

QUAD 1 (4-5)
H #2 Purdue - W
A #2 Purdue - L
H #12 Illinois - W
A #12 Illinois - L
H #19 Dayton - W
H #20 Michigan State - W
A #21 Wisconsin - L
N #38 Mississippi State - L
A #49 Nebraska - L

QUAD 2 (5-2)
H #49 Nebraska - W
H #64 Ohio State - W
H #74 Maryland - W
A #77 Minnesota - L
A #89 Rutgers - L
A #99 Penn State - W
A #101 Indiana - W

QUAD 3 (3-0)
H #99 Penn State - W
N #135 Arizona State - W
N #193 Rhode Island - W

QUAD 4 (6-1)
H #276 Binghamton - W
H #295 Northern Illinois - W
H #280 Jackson State - W
H #292 Chicago State - L
H #308 Western Michigan - W
A #317 DePaul - W
H #359 Detroit Mercy - W

This week features a Quad 3 home game against Michigan.

 

[RoseBowl2022]

And if Minnesota keeps winning (up to #77 after last night's win), this week's game could be our last Quad 3 game of the regular season.

 

[phatcat]

Chicago State AND Detroit Mercy won!?!

Lunardi still 9 seed https://www.espn.com/espn/feature/s...logy-2024-march-madness-men-field-predictions

 

[curdog]

Chicago State AND Detroit Mercy won!?!

Lunardi still 9 seed https://www.espn.com/espn/feature/s...logy-2024-march-madness-men-field-predictions

I think I would peefer a 10 seed so a win avoids a top seed. This bracket has cats against Houston in the 2nd round. That would be a really bad matchup

 

[GrandRapidsCatFan]

What would our NET ranking have to be to feel comfortable about receiving a bid? Because right now, I have a terrible feeling about our position

 

[NUAlum2010]

What would our NET ranking have to be to feel comfortable about receiving a bid? Because right now, I have a terrible feeling about our position

NUs NET rating kind of is what it is at this point. Unlikely to get appreciably better given the remaining schedule (short of winning by 20+ at MSU).

While it’s objectively bad for an at large tournament team based on recent seasons, they will have more than made up for it with the rest of the resume if they get to 12-8 or 13-7 in conference.

Winning at least 3 more games - even if by 1 point each - is really the only thing that matters at this point.

Could NU be left out at 21-11 (12-8) with a first game BTT exit? I suppose. But that would be one hell of a snub - regardless of NET - given the current bubble and NUs record against quad 1/2. Without looking at an all time list of snubs, I’m going to guess that resume would compete for the top spot on a snub list.

 

[RoseBowl2022]

What would our NET ranking have to be to feel comfortable about receiving a bid? Because right now, I have a terrible feeling about our position

The good news is the non-NET rankings are better -- especially Strength of Record (#25) and SOR seems to be a major factor for the Selection Committee.

 

[phatcat]

What would our NET ranking have to be to feel comfortable about receiving a bid? Because right now, I have a terrible feeling about our position

I feel like we are in at 12-8, presuming our usual BTT pants pooping. 11-9, we might need an opening round win in the dreaded BTT.

 

[SmellyCat]

I feel like we are in at 12-8, presuming our usual BTT pants pooping. 11-9, we might need an opening round win in the dreaded BTT.

The only thing good thing about 11-9 is that NU still probably still gets a bye and would presumably play a worse opponent who played the day before. Not saying NU couldn't poop their pants in that game, but that is a nice setup for getting that one last win.

But I'd rather get 13 wins and the second seed just because it would be awesome.

 

[PurpleWhiteBoy]

Any rating that has us #56 is somewhat silly.
If the algorithm looked at the loss of Ty Berry and attempted to measure the impact of that, then I'd cut it some slack.
But the NET doesn't do that - its all about the final score of the games that have been played and knows nothing about roster personnel for any given game.

I have NU at #28 based on wins and losses and #34 based on final point margins. At some point I will weight recent games more heavily - just haven't gotten around to it.

 

[UpsetAlert]

Chicago State AND Detroit Mercy won!?!

Lunardi still 9 seed https://www.espn.com/espn/feature/s...logy-2024-march-madness-men-field-predictions

Brooklyn please!

 

[SmellyCat]

I'm rooting for Indianapolis, but that can't happen if NU is in the 8/9 game, because you have to assume Purdue is getting the 1 there.

I like how Lunardi's projection today has Dayton/Illinois as a potential second-round game in which the winner would play NU fresh off its upset of Houston.

 

[curdog]

If it makes you feel better

What would our NET ranking have to be to feel comfortable about receiving a bid? Because right now, I have a terrible feeling about our position

If it makes you feel better, the team with the highest Net ranking to make the tourney is Rutgers at 77 (can’t remember the year). Also, the highest Net ranking to not make the tourney is 37 (I forget the team).

 

[ricko654321]

I'm rooting for Indianapolis, but that can't happen if NU is in the 8/9 game, because you have to assume Purdue is getting the 1 there.

I like how Lunardi's projection today has Dayton/Illinois as a potential second-round game in which the winner would play NU fresh off its upset of Houston.

Marquette is also probably going to Indy and could likely be a 2 seed, which brings in a possible 7/10 matchup to Indy.

Agree the NCAA is most likely not putting us in Purdue’s pod if we are an 8/9 seed. If we are then the likely destinations are Brooklyn (UConn pod), Memphis (Houston pod), or SLC (Arizona pod). If you look further down the 1’s and 2’s you tend see more Memphis / Charlotte / Omaha candidates - teams like Tennessee (Memphis), UNC and Duke (Charlotte), Iowa St and Kansas (Omaha) etc.

 

[phatcat]

Marquette is also probably going to Indy and could likely be a 2 seed, which brings in a possible 7/10 matchup to Indy.

Agree the NCAA is most likely not putting us in Purdue’s pod if we are an 8/9 seed. If we are then the likely destinations are Brooklyn (UConn pod), Memphis (Houston pod), or SLC (Arizona pod). If you look further down the 1’s and 2’s you tend see more Memphis / Charlotte / Omaha candidates - teams like Tennessee (Memphis), UNC and Duke (Charlotte), Iowa St and Kansas (Omaha) etc.

PU fans are sick of Boo & Company, it would be interesting to play them twice more (BTT/NCAA)

 

[Secho99]

PU fans are sick of Boo & Company, it would be interesting to play them twice more (BTT/NCAA)

Teams from the same conference can’t face each other until the Sweet 16 if they played twice during the regular season. If you played three times (including conference tourney) you aren’t allowed to face each other until the Elite 8.

 

[ricko654321]

What would our NET ranking have to be to feel comfortable about receiving a bid? Because right now, I have a terrible feeling about our position

The NCAA committee always says that the NET is not viewed as an actual ranking, it’s more for identifying and counting categories of wins (Quad 1-4). I think in terms of where you get seeded, the results based metrics and quality of wins / losses (where our 4-5 in Quad 1 looks quite good) matters more. Though the predictive metrics aren’t totally disregarded. That’s why Lunardi, the Athletic, etc have had us at an 8/9 seed despite a NET in the mid 50’s. And why they have Auburn as a 4 seed at the weekend release despite being 6th in NET, Bama as a 3 seed despite being 5 in the NET - both had only 2 Q1 wins which is weighing against their strong predictive metrics.

 

[NUCat320]

I’m no expert, but NU would need a colossal collapse to miss the tournament. Say, lose the two road games, lose in the first round of the BTT, and win only one of the three remaining home games.

This team is too good for a collapse. Boo is too much of a talent and he and the Cats are headed for some well-earned rest.

Currently in 96 of 98 bracket matrix fields, aggregated as the last 9 seed. Like everyone, I’d prefer a 10 or 11 to a 9, but I’d *really* prefer a BTT win and a 6.