Looking at KenPom, the Big10

[IL Wildcat]

Looks like some kind of super conference this season. All 12 of their teams are in the top 36 of KenPom. This is the genesis of Pomeroy's issues. What they really are is a bunch of mediocre teams, with not a single dominant team among them, beating up on each other all season. One of the biggest evidences of this is the fact that Purdue is 14-14 and 2 spots ahead of #27 Kentucky at #25. How in the world can that be? That is an embarrassment to KenPom in my opinion. There is no question that the Big10 metrics are all screwed up this year.

 

[Cats192]

Yeah, some real issues with the computer systems coming out this year

 

[Smashcat]

Lost faith in KenPom in 2015 when Duke didn’t qualify to win a championship per his system but their numbers suddenly got better as they advanced in the NCAAs to make them a KemPom qualifier.

 

[the nail at the foul line]

How many second weekend teams are in the Big 10?

Maryland most likely, then there is Penn State, if they get lucky and Ohio State, if they get lucky. Michigan State possibly. I assume their first weekend is going to be pretty bad.

 

[DerekMcPwn]

Looks like some kind of super conference this season. All 12 of their teams are in the top 36 of KenPom. This is the genesis of Pomeroy's issues. What they really are is a bunch of mediocre teams, with not a single dominant team among them, beating up on each other all season. One of the biggest evidences of this is the fact that Purdue is 14-14 and 2 spots ahead of #27 Kentucky at #25. How in the world can that be? That is an embarrassment to KenPom in my opinion. There is no question that the Big10 metrics are all screwed up this year.

You’re gonna have to explain that. If the conference is “just a bunch of mediocre teams beating up on each other,” then the end result should be bad overall efficiency scores, not good ones.

In a universe with only two teams and no games played, if Team A beats Team B 80-40 in a 60-possession game, Team A has excellent offensive and defensive efficiency (1.33 points scored per possession, .667 points per possession allowed). Team A would be far and away the highest efficiency team on KenPom with those scores.

If the tables turn in the next game, and Team B wins the same game by the same margin, then the scores for both teams average out to 1 point per possession scored, 1 point per possession allowed, which would be very poor efficiency numbers (dropping both teams to about 170 overall on KenPom).

If Team B instead wins the second game 60-58, then the resulting efficiencies are:

Team A: 1.15 points scored per possession, .833 points allowed per possession (would still rank first overall on KenPom this year)

Team B: .833 points scored per possession, 1.15 points allowed per possession (would still rank almost dead last on KenPom, despite having just defeated an incredibly powerful team).

A conference beating up on itself does not lead to excellent efficiency scores for all teams, unless that the teams in that conference torched all of their nonconference opponents.

So what is your exact theory, mathematically, about what is happening here?

 

[DerekMcPwn]

Let me hit also this descriptively instead of mathematically. What you’re saying doesn’t make sense because in a closed conference measurement, everything is zero sum. Any time one team scores a lot of points, another team pays a price it its defensive efficiency. Any time one team prevents another team from scoring, it hurts the other team’s offensive efficiency. Net efficiency, which determines KenPom ranking, is just the difference between the two.

In other words, a snake that eats its tail doesn’t get bigger.

 

[Wildcats1st]

Last year the big ten had 8 teams in. It’s not a matter of screwed up metrics it’s the fact that the big ten was strong out of conference and they rarely play a bad team in conference. Minnesota and northwestern may be it for bad teams. If LSU Florida MSU ark and others took care of business out of conference the sec would look better. Instead it’s a crap conference with best win probably being UK against UL or on the road vs TT.

 

[HenryMuto]

Or maybe the Big Ten is actually really good and people here can't accept that. KenPom has some flaws but the Big Ten is going to get 10 teams in the dance and the SEC is going to get maybe 5. When good teams play each other someone has to lose.

 

[DerekMcPwn]

Last year the big ten had 8 teams in. It’s not a matter of screwed up metrics it’s the fact that the big ten was strong out of conference and they rarely play a bad team in conference. Minnesota and northwestern may be it for bad teams. If LSU Florida MSU ark and others took care of business out of conference the sec would look better. Instead it’s a crap conference with best win probably being UK against UL or on the road vs TT.

Exactly. Other than Auburn, which played an atrociously weak schedule, the SEC absolutely **** the bed during nonconference, Kentucky included.

That may build a lie into the metrics, if the SEC’s losses and the B10’s wins were all flukes, but what’s more likely to be the falsehood - the net outcome of hundreds of games or a single fan’s eye test?

 

[HenryMuto]

Also what does it say about the SEC and Big Ten when the best team in the SEC can't beat the 9th place team in the Big Ten on a neutral court.

 

[sk73]

Last year the big ten had 8 teams in. It’s not a matter of screwed up metrics it’s the fact that the big ten was strong out of conference and they rarely play a bad team in conference. Minnesota and northwestern may be it for bad teams. If LSU Florida MSU ark and others took care of business out of conference the sec would look better. Instead it’s a crap conference with best win probably being UK against UL or on the road vs TT.

Toss in UK vs. Mich. St. #1 at the time

 

[Mike_Edwards]

Looks like some kind of super conference this season. All 12 of their teams are in the top 36 of KenPom. This is the genesis of Pomeroy's issues. What they really are is a bunch of mediocre teams, with not a single dominant team among them, beating up on each other all season. One of the biggest evidences of this is the fact that Purdue is 14-14 and 2 spots ahead of #27 Kentucky at #25. How in the world can that be? That is an embarrassment to KenPom in my opinion. There is no question that the Big10 metrics are all screwed up this year.

It’s absolute garbage and needs to go mainstream public before these yahoos give almost 10% of the tournament at large bids to that trash conference.

 

[irishcat1965]

What’s it say about the Big Ten that the overrated conference hasn’t won a NCAA title since 2000? It says the teams don’t get it done in March when it counts.

 

[Graced 1]

Hey, if Ohio St didnt beat UK you have about 12 other favorite teams that would have a chance.
Must be nice to like multiple teams-always have a team with a chance to win it.You probably pull for Louisville too

 

[~Keyser Soze~]

B1G has 14 teams and Northwestern and Nebraska can't possibly be anywhere near the top 36. They might be the 2 worst power 5 teams in the country.

 

[IL Wildcat]

You’re gonna have to explain that. If the conference is “just a bunch of mediocre teams beating up on each other,” then the end result should be bad overall efficiency scores, not good ones.

In a universe with only two teams and no games played, if Team A beats Team B 80-40 in a 60-possession game, Team A has excellent offensive and defensive efficiency (1.33 points scored per possession, .667 points per possession allowed). Team A would be far and away the highest efficiency team on KenPom with those scores.

If the tables turn in the next game, and Team B wins the same game by the same margin, then the scores for both teams average out to 1 point per possession scored, 1 point per possession allowed, which would be very poor efficiency numbers (dropping both teams to about 170 overall on KenPom).

If Team B instead wins the second game 60-58, then the resulting efficiencies are:

Team A: 1.15 points scored per possession, .833 points allowed per possession (would still rank first overall on KenPom this year)

Team B: .833 points scored per possession, 1.15 points allowed per possession (would still rank almost dead last on KenPom, despite having just defeated an incredibly powerful team).

A conference beating up on itself does not lead to excellent efficiency scores for all teams, unless that the teams in that conference torched all of their nonconference opponents.

So what is your exact theory, mathematically, about what is happening here?

Good points Derek. Way to pounce on me with both feet bro.

I never said that the "reason" the big ten is overrated is because they're mediocre and beating up on each other. Clearly, that is not WHY KenPom's numbers are skewed. I would have a couple theories regarding why their numbers are inflated.

#1 The initial ratings for some of their teams were skewed. Mich St is a likely culprit. They were highly rated from the beginning. They have remained highly rated despite their mounting losses. Ohio St is another example. (UK didn't help their over-rating, either)

#2 The ACC-BIG10 Challenge brought the ACC's early ratings into play as well. This was one of the few years the Big10 won that challenge. It was assumed at the time that the reason they won it was because of how dominant the B1G is this season. We've since learned that it was because the ACC is terrible this year.

Those are a couple theories. Attack away bro. Since the conference teams became overrated, they've since just beat up on each other.

I can't fully explain how the ratings got screwed up, obviously, but it's clear they're skewed. Wouldn't you agree? Or do you think all 12 of their teams are among the top 36 nationally?

 

[IL Wildcat]

B1G has 14 teams and Northwestern and Nebraska can't possibly be anywhere near the top 36. They might be the 2 worst power 5 teams in the country.

Oops, their top 12 teams are all top 36. I forgot their were 14 teams.

 

[Mike_Edwards]

One of my primary complaints in the recent Cal era is trading baskets with inferior opponents. That plays a large role in efficiency numbers. We need to start curb stomping the teams we should curb stomp.

 

[irishcat1965]

One of my primary complaints in the recent Cal era is trading baskets with inferior opponents. That plays a large role in efficiency numbers. We need to start curb stomping the teams we should curb stomp.

One of the reasons is Cal has more recruiting more athletes and less pure basketball players. I think that changes a bit with next year’s class.

 

[$Z71$]

Or maybe the Big Ten is actually really good and people here can't accept that. KenPom has some flaws but the Big Ten is going to get 10 teams in the dance and the SEC is going to get maybe 5. When good teams play each other someone has to lose.

If they get 10 teams then you believe a 12 loss Rutgers or Wisconsin deserve to be in? They are 10 loss teams who will have 12-13 by selection Sunday. 8 teams max should get in. The Big 10 is the top conference top to bottom I agree. But an 11-12 loss Rutgers with there resume shouldn’t be in the tournament. They lost to St Bonaventure and Pitt OOC and the best wins they have are Seton Hall OOC and Illinois and Penn State in conference.

 

[irishcat1965]

If they get 10 teams then you believe a 12 loss either or 12 loss Wisconsin deserve to be in? They are 10 loss teams who will have 12-13 by selection Sunday. 8 teams max should get in. The Big 10 is the top conference top to bottom I agree. But an 11-12 loss Rutgers with there resume shouldn’t be in the tournament. They lost to St Bonaventure and Pitt OOC and the best wins they have are Seton Hall OOC and Illinois and Penn State in conference.

Rutgers also can’t win a game away from home. They are a fraud.

 

[$Z71$]

Rutgers also can’t win a game away from home. They are a fraud.

Totally agree. Big 10 has zero teams who I’d consider a top 10 favorite to win it all. Their top teams are Maryland and Penn State. They have a ton of above average teams but no legit top teams. No one will convince me otherwise. They definitely have bottom feeders, so to say there are no easy wins in the big 10 there definitely are.

As I’ve stated I do feel like they are the best conference top to bottom. But they have little to no chance to have the champion or even a final four team from their conference. Big 12, Big East, SEC and ACC all have a much higher percentage chance of having the champion this year than the Big 10 does. I wouldn’t be surprised if they don’t have a team get past the sweet 16, but considering they are getting 8-10 teams in the will probably have 1 reach the Elite 8.

 

[KyCatFan]

I said it in another thread that I believe the BIG10 is full of mostly middle of the pack teams. I just don't see a legit title contender in the bunch and won't be surprised to see most of them out of the NCAA Tournament in the first weekend.

 

[DerekMcPwn]

Good points Derek. Way to pounce on me with both feet bro.

I never said that the "reason" the big ten is overrated is because they're mediocre and beating up on each other. Clearly, that is not WHY KenPom's numbers are skewed. I would have a couple theories regarding why their numbers are inflated.

#1 The initial ratings for some of their teams were skewed. Mich St is a likely culprit. They were highly rated from the beginning. They have remained highly rated despite their mounting losses. Ohio St is another example. (UK didn't help their over-rating, either)

#2 The ACC-BIG10 Challenge brought the ACC's early ratings into play as well. This was one of the few years the Big10 won that challenge. It was assumed at the time that the reason they won it was because of how dominant the B1G is this season. We've since learned that it was because the ACC is terrible this year.

Those are a couple theories. Attack away bro. Since the conference teams became overrated, they've since just beat up on each other.

I can't fully explain how the ratings got screwed up, obviously, but it's clear they're skewed. Wouldn't you agree? Or do you think all 12 of their teams are among the top 36 nationally?

1. The initial rankings aren’t relevant. KenPom’s preseason rankings apply certain assumptions based on a team’s stats from the previous season and its returning players, but by this time in the season, all of that guesswork has been cycled out and replaced with actual performance numbers from the current season.

2. The B10/ACC challenge undoubtedly helped the B10 if the B10 won, but it was a single game for each participating team, and only a half dozen games out of 140 total nonconference games, so it wouldn’t be very explanatory by itself.

I think the more plausible explanation, although it’s painful to admit, is that the B10 has consistently performed very well against other conferences.

 

[HenryMuto]

Kentucky moved up to #26 from #27 after Ohio State's win today.

 

[HenryMuto]

Kentucky moved up to #26 from #27 after Ohio State's win today.

Just kidding. I think it was because Butler got smoked and dropped below Kentucky.

 

[aikenite4uk]

These stat rankings start with a bias based on pre season rankings, and don't take into consideration improvement over the course of the season. Same with football's S&P rankings...that early season bias always makes UK football underrated. We're consistently ranked below teams we regularly roll...see Missouri, and until this season USC.

Net, I think the Big Ten is somewhat overrated, and the SEC is somewhat underrated. That early season bias continues into the conference season making it impossible for SEC teams to generate "quality wins" and excuses "poor losses" from the overrated conferences. The SEC has benefited from this bias in the past so I'm not really crying foul here. Check out the Vegas odds...we're consistently in the top 8-10 teams in terms of odds to win the championship. These systems that don't recognize late season team improvement will ALWAYS undervalue UK. Vegas knows the score.

 

[The_Answer1313]

But the numbers aren't skewed.

The Big Ten is just good this season.

Check the predictions of those Big Ten games on Kenpom.
Then check Vegas spreads.

Very close to one another.

So perhaps he has the teams accurately rated.

If you are ranking teams and predicting things along with Vegas spreads, you're doing something right.

 

[The_Answer1313]

This season


Win probability forecasts
Predicted
Range W-L Win% Exp
Win%
50-60% 611-479 56.1 54.8
60-70% 634-359 63.8 65.0
70-80% 703-254 73.5 74.7
80-90% 704-97 87.9 84.9
90-95% 357-27 93.0 92.5
95-98% 204-5 97.6 96.4
98-99% 98-4 96.1 98.5
99-100% 74-1 98.7 99.6

That is pretty damn accurate if you ask me.

 

[aikenite4uk]

But the numbers aren't skewed.

The Big Ten is just good this season.

Check the predictions of those Big Ten games on Kenpom.
Then check Vegas spreads.

Very close to one another.

So perhaps he has the teams accurately rated.

If you are ranking teams and predicting things along with Vegas spreads, you're doing something right.

Anybody that thinks that Purdue is better than UK is living in a dream world. We have beaten MSU (9 losses) and Uavel BOTH of whom are in the top 9 in KenPom. Ridiculous. Again, the fatal flaw in these rankings is they don't take into consideration improvement. We're way better than the fluky loss to Evansville and some of the other early season losses.
Latest Oddshark odds:
These are before last night's win against UF
UK 20-1
MSU 20-1
Purdue 75-1
Michigan 50-1
Iowa/ILL 75-1
Penn State 40-1

The teams listed below UK at higher odds are ALL ranked ahead of us in KenPom. So, clearly Vegas is rating us ahead of many teams vs. the KenPom system. I guess we'll find out who's right soon enough.

 

[DerekMcPwn]

These stat rankings start with a bias based on pre season rankings, and don't take into consideration improvement over the course of the season. Same with football's S&P rankings...that early season bias always makes UK football underrated. We're consistently ranked below teams we regularly roll...see Missouri, and until this season USC.

Net, I think the Big Ten is somewhat overrated, and the SEC is somewhat underrated. That early season bias continues into the conference season making it impossible for SEC teams to generate "quality wins" and excuses "poor losses" from the overrated conferences. The SEC has benefited from this bias in the past so I'm not really crying foul here. Check out the Vegas odds...we're consistently in the top 8-10 teams in terms of odds to win the championship. These systems that don't recognize late season team improvement will ALWAYS undervalue UK. Vegas knows the score.

Your points are false, and were already discussed.

1. Preseason rankings are ultimately irrelevant in KenPom because by mid-season they’ve completely cycled out and the efficiency scores are based entirely on current season performance. And even if that weren’t true (it definitely is), Kentucky started the season ranked way higher in KenPom than we are now, so we would benefit heavily from that preseason “bias.”

2. Of course KenPom factors improvement, but 1) only to the extent your performance numbers reflect it, and 2) it doesn’t erase early season bed-wetting. I agree that Kentucky in the Calipari era, on average, grows more inside of each season than any other team, but that doesn’t mean the entire first half of the season should be thrown out. The metrics are still more predictive with that data than without it.

 

[aikenite4uk]

Your points are false, and were already discussed.

1. Preseason rankings are ultimately irrelevant in KenPom because by mid-season they’ve completely cycled out and the efficiency scores are based entirely on current season performance. And even if that weren’t true (it definitely is), Kentucky started the season ranked way higher in KenPom than we are now, so we would benefit heavily from that preseason “bias.”

2. Of course KenPom factors improvement, but 1) only to the extent your performance numbers reflect it, and 2) it doesn’t erase early season bed-wetting. I agree that Kentucky in the Calipari era, on average, grows more inside of each season than any other team, but that doesn’t mean the entire first half of the season should be thrown out. The metrics are still more predictive with that data than without it.

Contradict yourself much? Wrong again as usual. The first half of the season is weighted equally to the second half in these systems.

 

[DerekMcPwn]

Contradict yourself much? Wrong again as usual. The first half of the season is weighted equally to the second half in these systems.

...what? That’s not an argument against anything I said.

 

[fatguy87]

Lost faith in KenPom in 2015 when Duke didn’t qualify to win a championship per his system but their numbers suddenly got better as they advanced in the NCAAs to make them a KemPom qualifier.

What disqualified them from winning the championship? There is no such thing in his system.

 

[The_Answer1313]

I'm starting to think the main issue people have with these metrics is the inability to predict everything correctly.

If it can't be 100%, why are we bothering lol

 

[IL Wildcat]

1. The initial rankings aren’t relevant. KenPom’s preseason rankings apply certain assumptions based on a team’s stats from the previous season and its returning players, but by this time in the season, all of that guesswork has been cycled out and replaced with actual performance numbers from the current season.

2. The B10/ACC challenge undoubtedly helped the B10 if the B10 won, but it was a single game for each participating team, and only a half dozen games out of 140 total nonconference games, so it wouldn’t be very explanatory by itself.

I think the more plausible explanation, although it’s painful to admit, is that the B10 has consistently performed very well against other conferences.

I'm not saying the Big10 isn't a good conference. Clearly, there are several good teams there. I'm just saying that the conference as a whole is too highly rated. I don't believe they have a legit title contender in the bunch. It seems to me their best team is probably Maryland, which just got beat by Ohio State today. That is not a top 10 team. I'd say the entire group should drop in most ratings by 10 spots, and then it might be close to accurate. I do think their top 14 are all decent teams. I just think KenPom's numbers are screwed up and they're all overrated. The voters in the major polls agree with me, by the way. The eyeball test definitely agrees with me, as does common sense.

If kenpom is right, UK would lose to 14-14 Purdue on a neutral floor. Do you really believe that?

 

[BigBlueFanGA]

Exactly. Other than Auburn, which played an atrociously weak schedule, the SEC absolutely **** the bed during nonconference, Kentucky included.

That may build a lie into the metrics, if the SEC’s losses and the B10’s wins were all flukes, but what’s more likely to be the falsehood - the net outcome of hundreds of games or a single fan’s eye test?

I'm sorry but when Michigan State is the top kenpom B10 team and Michigan is the 3rd, both of whom are rated way better than UK, something is broken. I dont care about the math. Michigan isnt good. We already beat Michigan State. At one point, Michigan went 4-8, two of those losses to teams we beat. Explain it any way you'd like, Kenpom is terribly flawed and far more than usual this year.

 

[kybassfan]

I think this has come down to scheduling as regards out of conference games. The scheduling decisions really have to start considering the ranking systems as some point. This is common place in football. Probably going to be needed in basketball too.

 

[SpartanJD]

How many second weekend teams are in the Big 10?

Maryland most likely, then there is Penn State, if they get lucky and Ohio State, if they get lucky. Michigan State possibly. I assume their first weekend is going to be pretty bad.

The big ten is going to have so many tough second round games because we won’t have more than 2-3 top 4 seeds when it’s all said and done. So expecting 5-11 seeds to make the sweat 16 is tough.

Still I will say 4 B10 teams and two SEC teams make the S16

 

[CapnBlubs]

The big ten is going to have so many tough second round games because we won’t have more than 2-3 top 4 seeds when it’s all said and done. So expecting 5-11 seeds to make the sweat 16 is tough.

Still I will say 4 B10 teams and two SEC teams make the S16

But once again, won't win a title. 1 in the last 30 seasons.

SEC = 6 in that span.