NET and KenPom Tracker 23-24 (SU = 84 3/9/24)

[FireballPhil]

Wasn't this supposed to be a super conference with the best of the old ACC and the best of the old Big East. Maybe we need to look at why that didn't happen.

Or maybe it did but the powers that be have another narrative to play.

 

[STEVEHOLT]

We're 10-3 having played a tougher schedule than 9-4 Pitt, and we beat them head-to-head; however, the difference is they lost close games, while we lost by 17, 19, and 22. The teams that beat us are better than the teams that beat Pitt, though. It's hard to climb the NET, especially when the ACC as a league isn't as strong as some other conferences and we're only playing each other from here forward. Grabbing a good win early in Maui would have really helped.

Getting blown out against good teams should count against you more than beating okay teams by a few works for you.

 

[CardiacCuse]

I’m struggling to reconcile some of the computer models this year. NET and KP don’t seem to like SU much.

me too!

by NET:
UC Irvine (9W-5L), Liberty (8-4), Louisiana Tech (7-5), Xavier (7-6) Boise st, Oregon, st marys (8-6), Mcneese (What)

just some of the squads ranked higher than the Orange...

for one, I think the formula waaay overvalues margin of victory (so tell the players that even if they lose they have to fight for 40)

for two...NCAABB got way too cute and forgot to just look at what is the most important stat there is: W-L

by KP:
north texas, seton hall, uc irvine, liberty, USC, washington, michigan, iowa, xavier, st mary's all with worse records than the orange...and ranked higher by the algorithm...

these ranking systems are broken.

 

[CardiacCuse]

Getting blown out against good teams should count against you more than beating okay teams by a few works for you.

strongly disagree.

margin of victory shouldnt matter that much...they can use it...but they overvalue it a lot.

it only matters for bettors and vegas and shouldnt be that important for ranking teams

these systems shouldnt be predictive...they should be reflective

for instance Mcneese state - g look at their resume -

trying to figure how they are 43rd in NET and Su is 78 ...

only thing i can see is Mcneese has a lot of blowout wins (vs creampuffs)

from now, coaches need to schedule creampuffs and run up the score IMO

 

[STEVEHOLT]

strongly disagree.

margin of victory shouldnt matter that much...they can use it...but they overvalue it a lot.

it only matters for bettors and vegas and shouldnt be that important for ranking teams

these systems shouldnt be predictive...they should be reflective

for instance Mcneese state - g look at their resume -

trying to figure how they are 43rd in NET and Su is 78 ...

only thing i can see is Mcneese has a lot of blowout wins (vs creampuffs)

from now, coaches need to schedule creampuffs and run up the score IMO

so margin of victory isn’t important but record is because you can falsely elevate your ranking by hammering cream puffs? That would also elevate your record. I’d say the rankings are quite accurate for teams that play a typical schedule. I’d bet McNeese would give us everything we could handle on a neutral site.

 

[Fly Rodder]

Some Syracusefan commentors and ranking algorithms.

They don't respect us!
*SU gets demolished by higher ranked teams*
We should be higher because we played stronger teams
*SU beats some cream puffs*
I don't get it, but what about the MARGIN OF VICTORY!
SU get demolished by UVA
We should be higher, they're against us.
*two decent victories later*
Wait, how about now? ITS BROKEN!

People would be less concerned if they went back and read jncuse 's post 3-4 times more.

 

[SWC75]

There's always "Against Ranked Teams".

 

[CardiacCuse]

Some Syracusefan commentors and ranking algorithms.

People would be less concerned if they went back and read jncuse 's post 3-4 times more.

well apparently i've triggered you...I just dont feel that the ranking systems are fair at the moment...especially NET

 

[Cuse99&01]

strongly disagree.

margin of victory shouldnt matter that much...they can use it...but they overvalue it a lot.

it only matters for bettors and vegas and shouldnt be that important for ranking teams

these systems shouldnt be predictive...they should be reflective

for instance Mcneese state - g look at their resume -

trying to figure how they are 43rd in NET and Su is 78 ...

only thing i can see is Mcneese has a lot of blowout wins (vs creampuffs)

from now, coaches need to schedule creampuffs and run up the score IMO

You have used McNeese St as a comparison and how it doesn’t make sense how they are 43 while SU is 78. A more egregious comparison is Iowa St. They are currently 6. Overall record is 11-2, which is inflated by going 8-0 vs quad 4 teams. Cuse is 1-0 against quad 4 teams. Neither ISU or SU have a quad 1 win (both 0-1) and both are 2-1 against quad 2 teams. ISU has played one true road game (@ DePaul). Makes no sense how ISU’s resume equates to a NET of 6 while Cuse sits at 78.

 

[UnderHisOwnPower]

You have used McNeese St as a comparison and how it doesn’t make sense how they are 43 while SU is 78. A more egregious comparison is Iowa St. They are currently 6. Overall record is 11-2, which is inflated by going 8-0 vs quad 4 teams. Cuse is 1-0 against quad 4 teams. Neither ISU or SU have a quad 1 win (both 0-1) and both are 2-1 against quad 2 teams. ISU has played one true road game (@ DePaul). Makes no sense how ISU’s resume equates to a NET of 6 while Cuse sits at 78.

Not Even Trying

 

[atlanticuse]

I still find RealtimeRPI the most reliable. RPI or NET or whatever, it’s still good.
We’re currently 14 but projected to be 34 end of year. Our NET equivalent would probably be high-30s / low-40s.
We play a lot of top-100 teams the rest of the way, and as others have said, we can’t lose to FSU, Wake, GT, Louisville, and ND. And even UNC and Duke don’t look that strong yet.

If we exceed expectations, we go dancing. If not, we don’t.
(Something is wrong with the LSU game result)

RealTimeRPI.com Syracuse Orange Men's College Basketball Rating Percentage Index (RPI) Ratings - A leading sports ratings and resources community on the Internet

RealTimeRPI.com: RPI, RPI Rankings, college basketball rpi, Real Time NCAA College Basketball and Sports Ratings - the most accurate independent analysis of the NCAA college basketball Rating Percentage Index (RPI)

www.realtimerpi.com

 

[OrlandoCuse]

I still find RealtimeRPI the most reliable. RPI or NET or whatever, it’s still good.
We’re currently 14 but projected to be 34 end of year. Our NET equivalent would probably be high-30s / low-40s.
We play a lot of top-100 teams the rest of the way, and as others have said, we can’t lose to FSU, Wake, GT, Louisville, and ND. And even UNC and Duke don’t look that strong yet.

If we exceed expectations, we go dancing. If not, we don’t.
(Something is wrong with the LSU game result)

RealTimeRPI.com Syracuse Orange Men's College Basketball Rating Percentage Index (RPI) Ratings - A leading sports ratings and resources community on the Internet

RealTimeRPI.com: RPI, RPI Rankings, college basketball rpi, Real Time NCAA College Basketball and Sports Ratings - the most accurate independent analysis of the NCAA college basketball Rating Percentage Index (RPI)

www.realtimerpi.com

That site’s not working correctly. Lots of incorrect data. Shows NC State as undefeated. My guess is it’s not being actively maintained.

 

[STEVEHOLT]

That site’s not working correctly. Lots of incorrect data. Shows NC State as undefeated. My guess is it’s not being actively maintained.

Aside from the incorrect data, it’s the best thing going

 

[atlanticuse]

In the past, it has been reliable and tracked our chances at the tournament better than a lot of the “official” trash - we can debate KenPom and NET all day.

But as RPI is no more, they’ve probably given up housekeeping.

As noted in my post, there are some data issues (our game with LSU and NC State among others).

 

[Fly Rodder]

well apparently i've triggered you...I just dont feel that the ranking systems are fair at the moment...especially NET

I get triggered by people complaining about things that they don't understand and refuse to take the time to understand them. It's just using numbers to try and answer two questions:

• How do you make a prediction for a game?
• Do matchups matter in predicting a game?

1) Don't think of the rankings as 61 is better than 62 or even 80. Think of them as tiers or groups of teams separated by minor fluctuations, the rank is just whether 9.64 is greater than 9.62, but it's not by itself very meaningful.​

2) There's no one twiddling the numbers behind the scenes to make SU look bad. That's silly. Ken Pomeroy doesn't care who ranks where in the system.​

It's based on 100 possessions. Not scoring say 75 points and holding the opposition to 65 points. Getting crushed by UVA in a low possession game was not good and it was recent enough that it is still heavily affecting any metrics. And these differences are mostly minuscule: For example, Michigan has a very efficient offense, they score a lot of points per 100 possessions, but the difference between their 24th offense (115.3 AdjO) and the 104th ranked offense (K-State at 107.5 AdjO) is only 7.8 points per 100 possessions. And very few, if any, teams average even close to a 100 possessions per game, so it's really closer to 6 actual points - two 3 point shots during an entire game.

 

[CardiacCuse]

You have used McNeese St as a comparison and how it doesn’t make sense how they are 43 while SU is 78. A more egregious comparison is Iowa St. They are currently 6. Overall record is 11-2, which is inflated by going 8-0 vs quad 4 teams. Cuse is 1-0 against quad 4 teams. Neither ISU or SU have a quad 1 win (both 0-1) and both are 2-1 against quad 2 teams. ISU has played one true road game (@ DePaul). Makes no sense how ISU’s resume equates to a NET of 6 while Cuse sits at 78.

theres not just one example there's 20

 

[CardiacCuse]

I still find RealtimeRPI the most reliable. RPI or NET or whatever, it’s still good.
We’re currently 14 but projected to be 34 end of year. Our NET equivalent would probably be high-30s / low-40s.
We play a lot of top-100 teams the rest of the way, and as others have said, we can’t lose to FSU, Wake, GT, Louisville, and ND. And even UNC and Duke don’t look that strong yet.

If we exceed expectations, we go dancing. If not, we don’t.
(Something is wrong with the LSU game result)

RealTimeRPI.com Syracuse Orange Men's College Basketball Rating Percentage Index (RPI) Ratings - A leading sports ratings and resources community on the Internet

RealTimeRPI.com: RPI, RPI Rankings, college basketball rpi, Real Time NCAA College Basketball and Sports Ratings - the most accurate independent analysis of the NCAA college basketball Rating Percentage Index (RPI)

www.realtimerpi.com

https://www.cbssports.com/college-basketball/rankings/rpi/

i think CBS is maintaining an RPI site

funny in this AI era of computers doing everything for us...no one can come up with a good way to rank bball teams

 

[CardiacCuse]

I get triggered by people complaining about things that they don't understand and refuse to take the time to understand them. It's just using numbers to try and answer two questions:

• How do you make a prediction for a game?
• Do matchups matter in predicting a game?

1) Don't think of the rankings as 61 is better than 62 or even 80. Think of them as tiers or groups of teams separated by minor fluctuations, the rank is just whether 9.64 is greater than 9.62, but it's not by itself very meaningful.​

2) There's no one twiddling the numbers behind the scenes to make SU look bad. That's silly. Ken Pomeroy doesn't care who ranks where in the system.​

It's based on 100 possessions. Not scoring say 75 points and holding the opposition to 65 points. Getting crushed by UVA in a low possession game was not good and it was recent enough that it is still heavily affecting any metrics. And these differences are mostly minuscule: For example, Michigan has a very efficient offense, they score a lot of points per 100 possessions, but the difference between their 24th offense (115.3 AdjO) and the 104th ranked offense (K-State at 107.5 AdjO) is only 7.8 points per 100 possessions. And very few, if any, teams average even close to a 100 possessions per game, so it's really closer to 6 actual points - two 3 point shots during an entire game.

i said especially NET

predictive systems and rankings systems are different animals...and should be treated as such, imo

my problem isnt with the "predictiveness" of kenpom...its with the btsht crazy rankings of NET

 

[upperdeck]

theres not just one example there's 20

But Net uses miuch more than just quad type wins..

does anyplace show how they rank in the efficiency variables?

NET seems to be a black box.

 

[CardiacCuse]

But Net uses miuch more than just quad type wins..

does anyplace show how they rank in the efficiency variables?

NET seems to be a black box.

it uses quads...which creates artificial boundaries between very similar teams

as if the the 29th and 31st or 74th and 76th ranked teams are worlds apart...

its a huge flaw

and it way overvalues early season games...they did I think bc it used to be that teams didnt really start playing for real until conference time...tried to make non-con more meaningful

but regardless,,,,,the quads are based on RPI anyway

 

[upperdeck]

it uses quads...which creates artificial boundaries between very similar teams

as if the the 29th and 31st or 74th and 76th ranked teams are worlds apart...

its a huge flaw

and it way overvalues early season games...they did I think bc it used to be that teams didnt really start playing for real until conference time...tried to make non-con more meaningful

but regardless,,,,,the quads are based on RPI anyway

i thought it creates quads using the formula and they quads are used to decide how many of each you get and thats used to help them make decisons at the end of the yr.

but none of that explains how they get the rankings

even starting at 1-2-3

byu 1-1 1-0 3-0 7-0
Purdue 4-1 3-0 2-0 3-0

what logic would put BYU ahead of Purdue?

more games, more better games. Purdue by a ton.

something else is in here.

 

[CardiacCuse]

i thought it creates quads using the formula and they quads are used to decide how many of each you get and thats used to help them make decisons at the end of the yr.

but none of that explains how they get the rankings

even starting at 1-2-3

byu 1-1 1-0 3-0 7-0
Purdue 4-1 3-0 2-0 3-0

what logic would put BYU ahead of Purdue?

more games, more better games. Purdue by a ton.

something else is in here.

the qauds change depending on home road neutral...its all bs

 

[atlanticuse]

https://www.cbssports.com/college-basketball/rankings/rpi/

i think CBS is maintaining an RPI site

funny in this AI era of computers doing everything for us...no one can come up with a good way to rank bball teams

and we're in a similar spot (17 = CBS, 14 = RealTimeRPI)

Never thought I'd say it, but I miss RPI.

We definitely could have used a "big" win in Maui (Oahu). That would have elevated us higher in the computer and human rankings. But we weren't ready for that level of difficulty yet (team wasn't gelling).

 

[littlereg]

Eye test never fails.

 

[Blood Orange]

I still find RealtimeRPI the most reliable. RPI or NET or whatever, it’s still good.
We’re currently 14 but projected to be 34 end of year. Our NET equivalent would probably be high-30s / low-40s.
We play a lot of top-100 teams the rest of the way, and as others have said, we can’t lose to FSU, Wake, GT, Louisville, and ND. And even UNC and Duke don’t look that strong yet.

If we exceed expectations, we go dancing. If not, we don’t.
(Something is wrong with the LSU game result)

RealTimeRPI.com Syracuse Orange Men's College Basketball Rating Percentage Index (RPI) Ratings - A leading sports ratings and resources community on the Internet

RealTimeRPI.com: RPI, RPI Rankings, college basketball rpi, Real Time NCAA College Basketball and Sports Ratings - the most accurate independent analysis of the NCAA college basketball Rating Percentage Index (RPI)

www.realtimerpi.com

What was the specific reason they stopped using RPI as a measure? I know some people had issues with it, but it still got the job done. The NET seems massively flawed.