ADVERTISEMENT

KenPom... just wow

BeerPoisoning

BeerPoisoning

All American
Feb 17, 2019
1,175
883
113
32
Earth
I know KenPom is considered the golden word of wisdom when it comes to college hoops analaytics, but it’s pure brainwashing this year. I can hardly find a damn article from someone announcing their own opinion. I know the NCAA thinks he understands quantum physics. I think he sounds like he ate batteries for breakfast.

Just gonna lay some stuff out.
35m4do9.jpg

Seeding doesn’t effect odds to win title.
I lost brain cells reading that article, apparently you can “significantly” lower a team’s seed without effecting win probability because the best teams will be in the final four. I guess Cinderellas also deserve to play teams they can beat. Expert logic, why not go all out? Let’s just cluster all the seeds together. #1s together, #2s together repeat through #9s on one side. Repeat same process #8s through #16s on the other.
Injuries give teams a free roll.
SeeDiNg DoEsNt MaTTeR... sit down
Objective Selection: No human logic.
Dating back to 2002 (KenPom’s appearance) his secret formula predictions hold a 16% accuracy rate at predicting the 4 best teams at 12/72 correct teams. NCAA is over 50% better with 25/72 correct #1 seeds.
Gonzaga’s offense ranks among best ever in KenPom’s system.
I can’t believe people PayPal this guy $20 annually to subscribe. What’s to gain? The ability to correctly predict a 16% chance of guessing a #1 seed in the final four? I’ve never heard of the “best offensive team” having a SOS of 84. Duke would drop 100/game on all those schools of the blind.

SportingNews just released an article a couple hours ago ago (which finally prompted this tirade from me) saying that 7 of the last 11 NCAA champions were KenPom’s #1 rated team. Are people aware that he updates final rankings after the tournament? Pre-tournament tells his real predictions. Yesterday a different site noted that Virginia wouldn’t win because no team in the bottom half of KenPom’s adjusted tempo metric has ever won. Wtf is adjusted tempo? According to KenPom it’s a team’s total possessions assuming they are playing an average tempo team. So we just assume that all opponents are running 64 possessions? Wow lol... Possessions aren’t even a real stat in college anyway, it’s an estimation from box score. But if we check out this metric: 07 Fl, 10 Duke, 11 UConn, 12 UK, 14 UConn all negate that argument.

I noticed he has a “Luck” rating. Describes it as winning or losing close games determines whether you are lucky or unlucky. On his website he has a tab labeled philosophies, one reads “Why I don’t believe in clutchness.” — Now just hold on. Isn’t clutch considered hitting a big shot with the game in question(close)? So he believes that teams get lucky/unlucky winning or losing these situations but doesn’t think that players have the intangibles to step up in big moments?

Am I badly mistaken or is this nuts?
 
  • Like
Reactions: jojozbludevilz
I know this is kinda deep but I just keep seeing people say “according to KenPom you should do this in your bracket” and half of the time it was something that made 0 sense.

So I spent an hour analyzing this guy... :confused:
 
BeerPoisoning said:
I know this is kinda deep but I just keep seeing people say “according to KenPom you should do this in your bracket” and half of the time it was something that made 0 sense.

So I spent an hour analyzing this guy... :confused:
Click to expand...
It only took an hour for you to do that? I’m impressed. It took me an hour just to read it lol.
 
  • Like
Reactions: hart2chesson, SoCal_Dukie3, Canadian Dukie and 3 others
pisgah101 said:
Don’t pay attention to GAAP he probably blocked you nownfor having a different opinion
Click to expand...

I’m cool with advanced team statistics that aren’t weighed on a formula of “assumed numbers” and predictions calculated in. That’s virtually all of metrics though, I’ve just never seen anyone go to the ridiculous level of this dude.

I can’t make sense of looking at made up numbers, especially when the results are estimating 16% of final four teams correctly. Different strokes for different folks I guess.

QC Dukie said:
It only took an hour for you to do that? I’m impressed. It took me an hour just to read it lol.
Click to expand...

Lmao it’s not that much research really, I’m just a substance person. I’ll add as much as possible. Could of been over an hour though idk I didn’t time myself
 
  • Like
Reactions: QC Dukie
BeerPoisoning said:
I know KenPom is considered the golden word of wisdom when it comes to college hoops analaytics, but it’s pure brainwashing this year. I can hardly find a damn article from someone announcing their own opinion. I know the NCAA thinks he understands quantum physics. I think he sounds like he ate batteries for breakfast.
Click to expand...
First off, I understand quantum physics. But it doesn't take that (or any advanced understanding of anything beyond multiplication) to understand KenPom or most of the other metrics out there.
The NCAA likely uses KenPom as one of their references. They use "various computer metrics" in addition to several things you would consider "eye tests."
https://www.ncaa.com/news/basketball-men/article/2018-10-19/how-field-68-teams-picked-march-madness

BeerPoisoning said:
Just gonna lay some stuff out.
35m4do9.jpg

Seeding doesn’t effect odds to win title.
I lost brain cells reading that article, apparently you can “significantly” lower a team’s seed without effecting win probability because the best teams will be in the final four. I guess Cinderellas also deserve to play teams they can beat. Expert logic, why not go all out? Let’s just cluster all the seeds together. #1s together, #2s together repeat through #9s on one side. Repeat same process #8s through #16s on the other.
Injuries give teams a free roll.
SeeDiNg DoEsNt MaTTeR... sit down
Click to expand...
You're missing the point. He's saying there isn't much difference playing a 15 vs playing a 16. Or a 6 vs 7. Take the UNC/Kentucky bracket. Most of the time, those two teams are going to get to the E8, whether or not you swapped their position in the bracket. So getting a 1 seed over a 2 isn't that much of an advantage mod geographical considerations.
BeerPoisoning said:
Objective Selection: No human logic.
Dating back to 2002 (KenPom’s appearance) his secret formula predictions hold a 16% accuracy rate at predicting the 4 best teams at 12/72 correct teams. NCAA is over 50% better with 25/72 correct #1 seeds.
Click to expand...
What? Are you saying that KenPom's best 4 actually make the F4 only 16% of the time whereas the teams seeded as 1s make it 35% of the time? Well there is some selection bias there because the committee sets the brackets, so 2 of KenPom's top 4 could be in the same bracket. Please explain if I'm misunderstanding.
BeerPoisoning said:
Gonzaga’s offense ranks among best ever in KenPom’s system.
I can’t believe people PayPal this guy $20 annually to subscribe. What’s to gain? The ability to correctly predict a 16% chance of guessing a #1 seed in the final four? I’ve never heard of the “best offensive team” having a SOS of 84. Duke would drop 100/game on all those schools of the blind.
Click to expand...
I hate to point this out, because you are going to get hung up on one data point, but Duke gave up 89 points to Gonzaga - the second most points they've given up all year. They also put up 90 on UNC, the third most they've given up all year. Good offenses do that. You're ignoring facts because of your bias (which is why KenPom uses numbers in the first place).
 
  • Like
Reactions: dukedevilz
BeerPoisoning said:
SportingNews just released an article a couple hours ago ago (which finally prompted this tirade from me) saying that 7 of the last 11 NCAA champions were KenPom’s #1 rated team. Are people aware that he updates final rankings after the tournament? Pre-tournament tells his real predictions. Yesterday a different site noted that Virginia wouldn’t win because no team in the bottom half of KenPom’s adjusted tempo metric has ever won. Wtf is adjusted tempo? According to KenPom it’s a team’s total possessions assuming they are playing an average tempo team. So we just assume that all opponents are running 64 possessions? Wow lol... Possessions aren’t even a real stat in college anyway, it’s an estimation from box score. But if we check out this metric: 07 Fl, 10 Duke, 11 UConn, 12 UK, 14 UConn all negate that argument.
Click to expand...
What do you mean possessions aren't a real stat? They are. If you have the ball for more possessions, you are likely going to score more points (all things equal). So if you play fast, and your opponent plays fast, then both teams will score more points than they usually do. Adjusted tempo accounts for this. It isn't made up.
Team A scores 100 points and Team B scores 100 points. Both of these scores were against UNC. Are they equally impressive?
What if Team A shot the ball with 26 seconds left on the shot clock every time and Team B shot with 2 seconds left on the clock every time. Which team do you think is more likely to win? (the answer is team B, because Team A likely had more possessions, so UNC had more possessions, so more chances to score). If you don't adjust for pace of play, you have no way to come Team As 100 points against UNC to Team Bs 100.
BeerPoisoning said:
I noticed he has a “Luck” rating. Describes it as winning or losing close games determines whether you are lucky or unlucky. On his website he has a tab labeled philosophies, one reads “Why I don’t believe in clutchness.” — Now just hold on. Isn’t clutch considered hittinng a big shot with the game in question(close)? So he believes that teams get lucky/unlucky winning or losing these situations but doesn’t think that players have the intangibles to step up in big moments?
Click to expand...
Vegas considers luck as well. If a team in the NFL (where turnover isn't as great as in college hoops and where teams are more evenly matched) scores the same number of points in a season that they give up, they should go somewhere around 8-8. If they go 12-4, what do you think Vegas thinks they will go the next season? It's closer to 8-8 than 12-4, because they should have won about half their games and probably got lucky in a few and won more close games than they lost. Is Vegas wrong too?
BeerPoisoning said:
Am I badly mistaken or is this nuts?
Click to expand...
You aren't badly mistaken. You don't like numbers - you've made that clear. But you don't like them because you probably don't understand. And you do sound like the old football coach that won't pass the ball. Everyone is using numbers because they ignore bias.
 
  • Like
Reactions: dukedevilz
BeerPoisoning said:
I’m cool with advanced team statistics that aren’t weighed on a formula of “assumed numbers” and predictions calculated in. That’s virtually all of metrics though, I’ve just never seen anyone go to the ridiculous level of this dude.
Click to expand...
What do you mean "assumed numbers"? I can guarantee you that KenPom isn't using assumed numbers. He is using data to adjust stats. Like what he does for pace (or opponent strength).

BeerPoisoning said:
I can’t make sense of looking at made up numbers, especially when the results are estimating 16% of final four teams correctly. Different strokes for different folks I guess.

Lmao it’s not that much research really, I’m just a substance person. I’ll add as much as possible. Could of been over an hour though idk I didn’t time myself
Click to expand...
He isn't making ANYTHING up. That's the point. When you said Duke would score 100 against the sisters of the blind, that is making things up because you don't have any objective way to make that claim. He uses data that says based on this team's results against similarly ranked teams, Duke should score X against this team. He isn't making anything up.
 
BeerPoisoning said:
Am I badly mistaken or is this nuts?
Click to expand...

You clearly don't understand the metrics, where they come from, or what they mean. Also, since you seem to be going so hard after KenPom, you should also realize his metrics are intended as a measure of the season the team has had, not a predictor of future success.
 
  • Like
Reactions: dukephysics
A team in the bottom half for adjusted tempo can't win the title? Really? Michigan's tempo was 324th last year. Loyola Chicago was 307th. Villanova won the title with the 274th fastest tempo in 2016. I know why people are discouraged from Virginia. One, their history. And two, less possessions means there's a higher probability of defeat if the other team catches fire. Virginia's defense is still top-notch - and this is the best shooting team Bennett has ever had. Can Virginia lose before the Final Four? Sure, but so can anyone.
 
  • Like
Reactions: BeerPoisoning
dukephysics said:
First off, I understand quantum physics. But it doesn't take that (or any advanced understanding of anything beyond multiplication) to understand KenPom or most of the other metrics out there.

The NCAA likely uses KenPom as one of their references. They use "various computer metrics" in addition to several things you would consider "eye tests."

https://www.ncaa.com/news/basketball-men/article/2018-10-19/how-field-68-teams-picked-march-madness
Click to expand...
I was taught that there’s areas of quantum physics experimented that can’t be predicted 100% because there’s probability involved in the outcome. KenPom uses probability in his metrics (there’s probability involved involved in like over 90% of all metrics) - I was making a joke. Just like I don’t literally think he eats batteries. If I’m wrong about the physics statement, you can take the W on that debate. I’m just making a joke, you’re the one with it in your username so I’ll assume you know best...... Ya I know metrics involve formulas of long drawn out multiplication and division, log5 or whatever it’s called. The NCAA does use KenPom, Sagarin and other tools. I am aware.

dukephysics said:
You're missing the point. He's saying there isn't much difference playing a 15 vs playing a 16. Or a 6 vs 7. Take the UNC/Kentucky bracket. Most of the time, those two teams are going to get to the E8, whether or not you swapped their position in the bracket. So getting a 1 seed over a 2 isn't that much of an advantage mod geographical considerations.
Click to expand...
I was being a smartass by suggesting clustering the #1 seeds and so on... I know he wasn’t being that dramatic. Btw - the article references moving a 1 seed to a 3 seed, not 2. Moving 1 to 3 will place teams into a tougher situation. It doesn’t matter if they *should* still win, the fact at hand is that now being a #3 seed they will meet the #2 seed in the S16 instead of E8. Idk how you can argue against that.

dukephysics said:
What? Are you saying that KenPom's best 4 actually make the F4 only 16% of the time whereas the teams seeded as 1s make it 35% of the time? Well there is some selection bias there because the committee sets the brackets, so 2 of KenPom's top 4 could be in the same bracket. Please explain if I'm misunderstanding.
Click to expand...
Yes, I am saying that. KenPom does probability and forecast models annually. I compared his teams with highest probability to win per region. You can go calculate it all if you’d like. I don’t think I made an error but I’m human.

dukephysics said:
hate to point this out, because you are going to get hung up on one data point, but Duke gave up 89 points to Gonzaga - the second most points they've given up all year. They also put up 90 on UNC, the third most they've given up all year. Good offenses do that. You're ignoring facts because of your bias (which is why KenPom uses numbers in the first place).
Click to expand...
Do you watch UNC and Duke games? If so and you’re a level-headed fan you would make note that they both prefer to run in transition. Majority of the ACC does not prefer to run like that, so the pace of play will dictate scoring. Gonzaga as well enjoys a fast-paced affair. That’s a much different scenario than running up the score on a school astronomical levels below your talent, like Gonzaga does in the WCC.

dukephysics said:
If you have the ball for more possessions, you are likely going to score more points (all things equal). So if you play fast, and your opponent plays fast, then both teams will score more points than they usually do. Adjusted tempo accounts for this. It isn't made up.

Team A scores 100 points and Team B scores 100 points. Both of these scores were against UNC. Are they equally impressive?

What if Team A shot the ball with 26 seconds left on the shot clock every time and Team B shot with 2 seconds left on the clock every time. Which team do you think is more likely to win? (the answer is team B, because Team A likely had more possessions, so UNC had more possessions, so more chances to score). If you don't adjust for pace of play, you have no way to come Team As 100 points against UNC to Team Bs 100.
Click to expand...
Your noting of fast pace = more points is exactly my argument in my answer above. You’re walking yourself in circles lol..... If slow-paced teams have to adjust to a faster pace of play to win then why did Virginia beat Carolina? It wasn’t dumb-luck.

We won the title in 2010, his metric system said that we were 198 in tempo. That’s in the bottom half of 353 teams, so the tempo method is now debunked by solid facts.

dukephysics said:
Vegas considers luck as well. If a team in the NFL (where turnover isn't as great as in college hoops and where teams are more evenly matched) scores the same number of points in a season that they give up, they should go somewhere around 8-8. If they go 12-4, what do you think Vegas thinks they will go the next season? It's closer to 8-8 than 12-4, because they should have won about half their games and probably got lucky in a few and won more close games than they lost. Is Vegas wrong too?
Click to expand...

Lots of things happen in the offseason, so the 8-8 record of the previous year will weight differently..... I don’t know what goes into making the initial line bet but Casinos are profitable because they force customers to gamble luck. If they gambled luck, it wouldn’t be a successful business.

Has a NCAA team won by a struck of luck? Yep. It happens. To every team every year handfuls of times? No, it doesn’t. There is an advanced stat called win % in close games. The percentage is found by the W/L ratio in a game that’s decided by less than 4 points. That’s how you measure close games.

KenPom calculates an adjusted offensive and defensive efficiency and from that formula decides whether the the win or loss was “lucky” or “unlucky” — If you think that’s better explained than W/L in close situations, you can have your opinion and I will have mine.

dukephysics said:
You aren't badly mistaken. You don't like numbers - you've made that clear. But you don't like them because you probably don't understand. And you do sound like the old football coach that won't pass the ball. Everyone is using numbers because they ignore bias.
Click to expand...
I thoroughly understand metrics, I actually look at them frequently. It’s interesting to look at and sometimes it really tells a true story. But I don’t consider metrics valuable information in most scenarios because probability is assumed numbers. This information does not include injuries, foul trouble or whatever special circumstance. Actually, KenPom himself said that it’s head-scratching to weigh injuries because players tend to play conservatively and coaches have different philosophies effecting play time.

Anyway... Box scores and a portion of advanced stats are pure information from 100% game happenings. I like those numbers.

I don’t coach football but I let it fly 99% of the time in Madden.

dukephysics said:
What do you mean possessions aren’t a real stat?
Click to expand...
In the NCAA, nobody keeps track of possessions. It’s clear that you believe what KenPom says so I’ll gladly screenshot and circle his words from his website.
2ppiwc5.jpg


dukephysics said:
What do you mean "assumed numbers"? I can guarantee you that KenPom isn't using assumed numbers. He is using data to adjust stats. Like what he does for pace (or opponent strength).
Click to expand...
I wrote assumed numbers on the screenshot above so I don’t have to answer this myself. Additionally, the part in square is KenPom telling you that rebounding is a skill and not a possession. I know you’ll automatically agree with him (although majority of the world doesn’t) but the NBA recognizes offensive rebounds as a new possession and they actually keep track of poss. So, I’m gonna stick with big brother.

dukephysics said:
He isn't making ANYTHING up. That's the point. When you said Duke would score 100 against the sisters of the blind, that is making things up because you don't have any objective way to make that claim. He uses data that says based on this team's results against similarly ranked teams, Duke should score X against this team. He isn't making anything up.
Click to expand...
You don’t read sarcasm well..... Anyway, I don’t think Gonzaga is a garbage team, I just don’t personally believe they have the best offense in the country. You can believe whatever you want, I ain’t stopping ya
 
Last edited:
pisgah101 said:
Yates in disguise?
Click to expand...
I think he might have more “time” than Yates. Haha. That’s something (else) I never got about him. Yates doesn’t have the “time” to watch the games blah blah, yet used to post stuff that Id only have the “time” to read on Taco Bell Tuesday’s on the Sh!tter.
 
  • Like
Reactions: Hurley22 and retiredsoldier
I work in genetics and love a good debate (especially clinically), but I come here for simplicity, good people, and some laughs. That said:

You can’t triple stamp a double stamp.

Debate that.
 
@BeerPoisoning I'm going to take my buddy's advice and not argue with you. You're welcome to your opinion on the issue. Personally, I like KenPom's approach because it is easily understandable (for most people) and gives reasonable results most of the time. By no means is it gospel. I'm pretty sure he would beat all of our eye tests in Vegas over the long term.
 
lyonhawk said:
You clearly don't understand the metrics, where they come from, or what they mean. Also, since you seem to be going so hard after KenPom, you should also realize his metrics are intended as a measure of the season the team has had, not a predictor of future success.
Click to expand...

This is directly from his website. Quoting the exact opposite of what you just said.
29ff69f.jpg


@dukephysics I’m assuming that you agreed with him since you liked his post so you can take a look at that if you’d like.

I read an article earlier today comparing KenPom, Sagarin and Sports betting. It said that KenPom and Sagarin both use sports betting point spreads in their predictions. Then the bettors use the sites to predict their gamble. — I have no idea how valid the article was though, I just remember seeing it. You could be right. I think the fair assessment is that they are all very close in comparison.
 
  • Like
Reactions: dukephysics
DukefanKY said:
Leave it to you to be different!!!
Click to expand...
Haha I sold my 300...I couldn’t get it to shoot right. I tried everything possible powder wise, bullets, case length, everything. The 08 shoots a dime at 100 and damn good at 300. Love it.
 
  • Like
Reactions: DukefanKY
You must log in or register to reply here.

Similar threads

pisgah101

Official games day thread- Duke vs Vermont- NCAA tournament

Replies
294
Views
7K
Duke Hoops Open Forum - Free
QC Dukie
Q
Dad2ze

What has High School Athletics turned into???

Replies
26
Views
1K
Duke Hoops Open Forum - Free
sheyduke
sheyduke
ConorONeill

Three upcoming road games ...

Replies
15
Views
630
Devil Insider
ConorONeill
ConorONeill
BluePhil

Duke vs. Pitt 2 - No win no Gameball

Replies
3
Views
467
Duke Hoops Open Forum - Free
BluePhil
BluePhil
ConorONeill

NCAAT - JMU-Duke | 5:15 p.m., 3/24

Replies
13
Views
534
Devil Insider
jeffrey236
jeffrey236
ADVERTISEMENT

Latest posts

ADVERTISEMENT
Top Bottom
Back