KenPom 2023-2024 KenPom Rankings Thread

[big red22]

2 hours ago, millerhusker said:

SOR is now #27, which would put us as a 7 seed. 

Michigan State is #46 and SOR seeding has them as the first team out. 

Seems more accurate than what the bracketologists have. 

SOR - 27

KPI - 28

Kenpom - 34

 

Those are 6-8 Seed numbers and most still have us as a 10... a couple had us still as a last four in on the 23rd.  I think if Selection Sunday happened today Nebraska and Northwestern would be 8 Seeds and Michigan State would get a 10

[Cookie Miller Wasn't Dirty]

2 hours ago, big red22 said:

SOR - 27

KPI - 28

Kenpom - 34

 

Those are 6-8 Seed numbers and most still have us as a 10... a couple had us still as a last four in on the 23rd.  I think if Selection Sunday happened today Nebraska and Northwestern would be 8 Seeds and Michigan State would get a 10

From what I’ve read, the results-based metrics (SOR, KPI) are more of a focus for selecting the 36 teams, but the predictive metrics (NET, KenPom, BPI) are used more for seeding. Not sure if there’s truth to that.

[hskr4life]

1 hour ago, Cookie Miller Wasn't Dirty said:

From what I’ve read, the results-based metrics (SOR, KPI) are more of a focus for selecting the 36 teams, but the predictive metrics (NET, KenPom, BPI) are used more for seeding. Not sure if there’s truth to that.

I’ve heard this as well.  It makes sense.  Let’s select teams who have done the best with the schedule they’ve been given and seed them based on how they should do moving forward.

[Cookie Miller Wasn't Dirty]

Clemson’s coach says exactly what I’ve been saying all year. I’ve pointed out BYU’s heavily flawed NET ranking. Same issue for KenPom and BPI. I didn’t realize it applied across the Big XII.

 

College basketball needs to fix this. Teams should not be rewarded for extending a 30 point win to 50. It appears Fred has become aware of this because he hasn’t put in the scrubs at the end of our recent blowouts. Unfortunately, until this changes, he should never put them in. Which is a shame for those walk-ons.
 

https://www.on3.com/news/clemson-coach-brad-brownell-frustrated-big-12-has-manipulate-net-rankings/

[Navin R. Johnson]

12 minutes ago, Cookie Miller Wasn't Dirty said:

Clemson’s coach says exactly what I’ve been saying all year. I’ve pointed out BYU’s heavily flawed NET ranking. Same issue for KenPom and BPI. I didn’t realize it applied across the Big XII.

 

College basketball needs to fix this. Teams should not be rewarded for extending a 30 point win to 50. It appears Fred has become aware of this because he hasn’t put in the scrubs at the end of our recent blowouts. Unfortunately, until this changes, he should never put them in. Which is a shame for those walk-ons.
 

https://www.on3.com/news/clemson-coach-brad-brownell-frustrated-big-12-has-manipulate-net-rankings/

 

Exactly.  It's not just playing time for the walk ons but also the younger kids you want to give playing time to but feel you can't risk it.  If it doesn't change teams might as well poor it on at the end and try to score every possession and not dribble the game out with seconds left.  It has always been considered good sportsmanship to hold the ball at the end but not anymore.  I really hate that metric.  I would max it out at 10 points at call it good.  If every coach knew that was the break point you could play to score up until that point.  Nobody really throws up a white flag if the game is still in single digits. 

[Cookie Miller Wasn't Dirty]

8 minutes ago, Navin R. Johnson said:

 

Exactly.  It's not just playing time for the walk ons but also the younger kids you want to give playing time to but feel you can't risk it.  If it doesn't change teams might as well poor it on at the end and try to score every possession and not dribble the game out with seconds left.  It has always been considered good sportsmanship to hold the ball at the end but not anymore.  I really hate that metric.  I would max it out at 10 points at call it good.  If every coach knew that was the break point you could play to score up until that point.  Nobody really throws up a white flag if the game is still in single digits. 

NET actually does cap the win margin at 10 points. So that’s good.

 

The problem is with efficiency metrics. It would be difficult systematically to parse out at what point in a game should efficiency matter, and at what point should it not. Because I’m sure these metrics just pull stats from the box score.

 

Offensive efficiency for example is calculated as Points divided by Possessions, where Possessions = FGA - OREB + TO + 0.475 * FTA.

 

It would be extremely difficult to go through every game and alter the final stats by removing any data that occurred after a certain point in the game.

[royalfan]

Not sure why efficiency metrics need to be included in calculations that are used anyway.  Seems the score paints a pretty good picture.  

[Cookie Miller Wasn't Dirty]

10 minutes ago, royalfan said:

Not sure why efficiency metrics need to be included in calculations that are used anyway.  Seems the score paints a pretty good picture.  

Math people trying to be too cute. I say that as a math person myself. The more complex the algorithm, the more distinguished the creator of it feels. Sometimes simplicity is the answer.

[Norm Peterson]

1 hour ago, Cookie Miller Wasn't Dirty said:

Clemson’s coach says exactly what I’ve been saying all year. I’ve pointed out BYU’s heavily flawed NET ranking. Same issue for KenPom and BPI. I didn’t realize it applied across the Big XII.

 

College basketball needs to fix this. Teams should not be rewarded for extending a 30 point win to 50. It appears Fred has become aware of this because he hasn’t put in the scrubs at the end of our recent blowouts. Unfortunately, until this changes, he should never put them in. Which is a shame for those walk-ons.
 

https://www.on3.com/news/clemson-coach-brad-brownell-frustrated-big-12-has-manipulate-net-rankings/

 

Well, there is some subjectivity as far as how heavily the committee weighs each factor. One year, they'll decide your road record matters more or your strength of schedule. Teams do what the committee's decision-making seems to encourage. We didn't get credit for playing a tough schedule in the past. So, now we schedule lightweights and try to beat them badly.

 

But if the Big 12 is as bad as he says, it should show up in the NCAA tourney when they make a bunch of 1st and 2nd-round exits.

[Cookie Miller Wasn't Dirty]

15 minutes ago, Norm Peterson said:

 

Well, there is some subjectivity as far as how heavily the committee weighs each factor. One year, they'll decide your road record matters more or your strength of schedule. Teams do what the committee's decision-making seems to encourage. We didn't get credit for playing a tough schedule in the past. So, now we schedule lightweights and try to beat them badly.

 

But if the Big 12 is as bad as he says, it should show up in the NCAA tourney when they make a bunch of 1st and 2nd-round exits.

I agree the predictive metrics are just one piece of the puzzle. Still, the goal of the metrics is to best quantify who should be in and where they should be seeded. If there’s a clear and obvious flaw, like rewarding 50 point wins over 30 point wins, something in the algorithm should be tweaked to mitigate that flaw.

[AuroranHusker]

1 hour ago, Cookie Miller Wasn't Dirty said:

I agree the predictive metrics are just one piece of the puzzle. Still, the goal of the metrics is to best quantify who should be in and where they should be seeded. If there’s a clear and obvious flaw, like rewarding 50 point wins over 30 point wins, something in the algorithm should be tweaked to mitigate that flaw.

 

 

Well, my 12 y.o. daughter's team played in a recent tournament, and the parameters capped the highest "margin" at 20 points when tiebreakers are the margin of victory ("point differential") over the course of the 3-game tourney, their other factor was points allowed as the next tiebreaker, so the defensive prowess was highlighted. I know this is simple stuff, but it was accounted for even for 3rd through 8th grade boys & girls. So, if they can figure it out for youth sports, you'd think they could simplify it & make a reasonable solution to the issue for NCAA men's & women's basketball. Sheesh.

 

Edited February 27 by AuroranHusker

[Sgt Lincoln Osiris]

1 hour ago, royalfan said:

Not sure why efficiency metrics need to be included in calculations that are used anyway.  Seems the score paints a pretty good picture.  

 

Kind of.  2018-2019 Virginia (National Champs) were the #2 most efficient offense according to KenPom, while being 210th in points per game.  They were 353/353 in tempo that year and had 6 future NBA players.  They were an elite offense, but you wouldn't know that looking just at their scores.  They were going to keep the game to 60 possessions, work for great shots, and score on a lot more of them than their opponent.  A high tempo team that is equally efficient with their possessions would beat teams by a larger amount.

 

The downside of using efficiency metrics is that they look at every possession, and I think everyone here seems to be in agreement there are unintended consequences. It seems like there could be a way to not count any possession when a team was up or down by some amount after a certain point in the game, but who is to say what those are?

[royalfan]

16 minutes ago, Sgt Lincoln Osiris said:

 

Kind of.  2018-2019 Virginia (National Champs) were the #2 most efficient offense according to KenPom, while being 210th in points per game.  They were 353/353 in tempo that year and had 6 future NBA players.  They were an elite offense, but you wouldn't know that looking just at their scores.  They were going to keep the game to 60 possessions, work for great shots, and score on a lot more of them than their opponent.  A high tempo team that is equally efficient with their possessions would beat teams by a larger amount.

 

The downside of using efficiency metrics is that they look at every possession, and I think everyone here seems to be in agreement there are unintended consequences. It seems like there could be a way to not count any possession when a team was up or down by some amount after a certain point in the game, but who is to say what those are?

Good point on the pace.  I think one could argue that an elite efficiency team playing at a drastically slow pace gives them a much larger chance of losing than one that plays at a pace with many more possessions.  Therefore, I think one could also argue that score is enough to he used in calculations that matter.  I love looking at the efficiency numbers etc.  just not sure they should be used in the key calculations that are being used, due to the curb stomping effect in games that is not good for college basketball.  

[Sgt Lincoln Osiris]

11 minutes ago, royalfan said:

Good point on the pace.  I think one could argue that an elite efficiency team playing at a drastically slow pace gives them a much larger chance of losing than one that plays at a pace with many more possessions.  Therefore, I think one could also argue that score is enough to he used in calculations that matter.  I love looking at the efficiency numbers etc.  just not sure they should be used in the key calculations that are being used, due to the curb stomping effect in games that is not good for college basketball.  

 

I also have to imagine the gaming of the system is going to decrease the effectiveness of the predictive model used by KenPom, Torvik, etc.  

 

[Cookie Miller Wasn't Dirty]

1 hour ago, Sgt Lincoln Osiris said:

It seems like there could be a way to not count any possession when a team was up or down by some amount after a certain point in the game, but who is to say what those are?

I’m sure these metrics all take data from box scores to calculate efficiency. It would be an administrative nightmare to parse out which possessions count and which ones don’t for every single game. Not saying it can’t be done, but I doubt the NCAA views the juice to be worth the squeeze.

 

You make a good point about Virginia. I think the best solution is to reduce the weight being given to efficiency.

 

Edit: Though as @royalfan points out, the margin for error is smaller for a slow paced team. So I think I’m back to believing efficiency should just be removed altogether.

Edited February 27 by Cookie Miller Wasn't Dirty

[hhcmatt]

18 minutes ago, Cookie Miller Wasn't Dirty said:

I’m sure these metrics all take data from box scores to calculate efficiency. It would be an administrative nightmare to parse out which possessions count and which ones don’t for every single game. Not saying it can’t be done, but I doubt the NCAA views the juice to be worth the squeeze.

 

Bartorvik does this

 

https://adamcwisports.blogspot.com/2018/09/t-rank-methodology-update.html

[Cookie Miller Wasn't Dirty]

32 minutes ago, hhcmatt said:

 

Bartorvik does this

 

https://adamcwisports.blogspot.com/2018/09/t-rank-methodology-update.html

Oh nice. I stand corrected.

 

Still some flaws there but it seems better. In the example he uses I don’t think it’s reasonable to assume a 29 point win by Michigan. It makes some sense to extrapolate the results if they’re playing a Q4 game, but being up 20 at half on Wisconsin shouldn’t lead to an assumed 29 point win.

 

Looks like he has BYU rated a little lower than NET, so that’s good. Still had them #6 after December which seems too high.

[hskr4life]

Seems like all these other teams lose and maybe drop a spot or two.  We lose and fricking drop to 39 in Pom. 

[Norm Peterson]

11 minutes ago, hskr4life said:

Seems like all these other teams lose and maybe drop a spot or two.  We lose and fricking drop to 39 in Pom. 

 

Thought the same thing. Road game, no less.

[hhcmatt]

24 minutes ago, hskr4life said:

Seems like all these other teams lose and maybe drop a spot or two.  We lose and fricking drop to 39 in Pom. 

 

we didn't really lose much in total efficiency margin it's just that teams were clumped enough

[Chuck Taylor]

On 2/27/2024 at 3:25 PM, Sgt Lincoln Osiris said:

 

I also have to imagine the gaming of the system is going to decrease the effectiveness of the predictive model used by KenPom, Torvik, etc.  

 

I don't know if that's true or not, but it's an interesting take. The metrics ratings is that they are designed to predict the outcome of games, so if beating a bad team by lot really means nothing, then KenPom etc. will change the formula.

[hhcmatt]

Lunardi had a good thread about the idea that the Big 12 is gaming the system

 

 

[KarlHessEatMy....]

Still a #9 seed in Lunardi bracket this morning!! so thats a good thing

[AuroranHusker]

It was still a Q1 opportunity last night, so a 9-pt. loss isn't going to change much unless it had been a 29-pt. loss.

 

The last Q1 opportunity during the regular season. Lost opportunity, but certainly not a bad loss, overall. Oh well, time to get 2 more B1G dubs.

[49r]

We drop 4 spots in KP to #38, but we are only 0.20 points away from being #35.  A big win on Sunday should get us right back to around where we were before last night.  Rutgers swaps spots back with Penn State.

 

Ken's prediction goes back to 21-10 (11-9).  Same for Torvik.  KP's odds of winning the last two games are:

• Rutgers - 77%
• Michigan - 67%

Here we go.

 

---

 

 

KenPom rankings as of 3-01-24

=======================

 

B1G (10-8):
2. Purdue - W
13. Illinois - L
21. Michigan State - W
23. Wisconsin - L, W
38. Nebraska

42. Northwestern - W, L

48. Maryland - L
50. Iowa - L
57. Ohio State - W, L
66. Minnesota - L, W

88. Rutgers - L

92. Penn State - W

101. Indiana - W, W

119. Michigan - W

 

Non-Conference (10-1):
357. Lindenwood - W

342. Florida A&M - W

---Cornhusker Classic---
244. Rider - W
210. Stony Brook - W

---Sanford Pentagon---
154. Oregon State - W

 

---Cornhusker Classic---

109. Duquesne - W

 

231. Cal State Fullerton - W

12. Creighton - L
71. @Kansas State - W
249. North Dakota - W

299. South Carolina State - W

Edited March 1 by 49r