MBB Dork Polls 2020-21

[DavidBenAkiva]

At a most interesting time, KenPom dropped his initial 2021 season rankings and Duke looks very good at #3.

Duke is 4th in AdjO and 6th at AdjD with a pretty fast tempo, ranked 38th.

Overall, Duke is a hair behind Baylor and Gonzaga and the only ACC team in the top 10.


https://www.kenpom.com

[Troublemaker]

I can't link to it because it's behind a paywall. But someone I trust has told me that KenPom came out with his first estimate for homecourt advantage this season, and it's 2.25 points. (Obviously different places/arenas will have slightly different advantages; 2.25 is just the average).

So, yeah, if in non-Covid seasons, the homecourt advantage is about 3.5 (the most popular estimate, although sometimes I see it as low as 3.1), then so far, HCA is still about 60-65% as strong as it was before, i.e. the impact of fans (and their impact on refs) is less than half of HCA. Pretty much what I expected in the preseason. Now that's a decent amount to slice off, but things like sight lines, no travel, etc. still represent the majority of HCA.

[House P]

I can't link to it because it's behind a paywall. But someone I trust has told me that KenPom came out with his first estimate for homecourt advantage this season, and it's 2.25 points. (Obviously different places/arenas will have slightly different advantages; 2.25 is just the average).

So, yeah, if in non-Covid seasons, the homecourt advantage is about 3.5 (the most popular estimate, although sometimes I see it as low as 3.1), then so far, HCA is still about 60-65% as strong as it was before, i.e. the impact of fans (and their impact on refs) is less than half of HCA. Pretty much what I expected in the preseason. Now that's a decent amount to slice off, but things like sight lines, no travel, etc. still represent the majority of HCA.

Very interesting. I suspect we will have an even better idea of HCA without fans once conference play picks up.

In fact, some conferences have dramatically altered their schedules in a way which should facilitate excellent estimates of HCA without fans. Consider the Patriot league which has decided to significantly adapt conference play from the usual round robin. Here is Boston University's entire schedule for the 2021 season. BU plays 2 games vs Lafayette, 4 games vs Army, 4 games vs Colgate, and 6 games vs Holy Cross.



Date
Opponent
Location


Sat Jan 2
Holy Cross
Away


Sun Jan 3
Holy Cross
Home


Sat Jan 9
Colgate
Home


Sun Jan 10
Colgate
Home


Sat Jan 16
Army
Home


Sun Jan 17
Army
Home


Sat Jan 23
Lafayette
Away


Sun Jan 24
Lafayette
Away


Sat Jan 30
Holy Cross
Away


Sun Jan 31
Holy Cross
Home


Sat Feb 6
Colgate
Away


Sun Feb 7
Colgate
Away


Sat Feb 13
Army
Away


Sun Feb 14
Army
Away


Sat Feb 20
Holy Cross
Home


Sun Feb 21
Holy Cross
Away

[Bluedog]

I can't link to it because it's behind a paywall. But someone I trust has told me that KenPom came out with his first estimate for homecourt advantage this season, and it's 2.25 points. (Obviously different places/arenas will have slightly different advantages; 2.25 is just the average).

So, yeah, if in non-Covid seasons, the homecourt advantage is about 3.5 (the most popular estimate, although sometimes I see it as low as 3.1), then so far, HCA is still about 60-65% as strong as it was before, i.e. the impact of fans (and their impact on refs) is less than half of HCA. Pretty much what I expected in the preseason. Now that's a decent amount to slice off, but things like sight lines, no travel, etc. still represent the majority of HCA.

Interesting indeed. Will await more data though, particularly in conference. But also there are SOME places where a certain number of students or family members of the players ARE allowed, I believe. So, it's not 100% fanless although it's mostly that way.

[BlueDevil2K]

Through last night's games...

kenpom: 29 (offense 16, defense 84)
sagarin: 30
t-rank: 39
net: 61 (up from 67, but still...)

A win over UVa would certainly help these numbers...

[DavidBenAkiva]

Through last night's games...

kenpom: 29 (offense 16, defense 84)
sagarin: 30
t-rank: 39
net: 61 (up from 67, but still...)

A win over UVa would certainly help these numbers...

Death, taxes, and Duke having a top 20 offense on KenPom. Remember when this team couldn't score?

Coincidently, if you filter T-Rank for games played since January 24th, Duke is 13th overall with the 6th highest AdjOE (but 113th AdjDE).

In the T-Ranketology, Duke is 9th in the "First Teams Out" list, just behind Syracuse and ahead of Georgia Tech. UNC is in the Last Four group. While Duke has played its way out of the NCAA Tournament, there's a real chance to play back into the field over the final weeks. At worst, Duke can play spoiler.

[-jk]

Huh. Kenpom just updated and we didn't move. I was hoping for a small bump.

-jk

[CDu]

Huh. Kenpom just updated and we didn't move. I was hoping for a small bump.

-jk

He didn’t update yet. His software updates the record before it updates the efficiency stats. We will move up a little, though probably not a ton.

[-jk]

He didn’t update yet. His software updates the record before it updates the efficiency stats. We will move up a little, though probably not a ton.

Thanks. I saw the record change and thought he had it in.

-jk

[House P]

He didn’t update yet. His software updates the record before it updates the efficiency stats. We will move up a little, though probably not a ton.

KenPom had Duke as a 1 point underdog going into the game, so a 1 point win probably won’t change things that much.

A 1 point win feels way better than a 1 point loss from a fans perspective, but the Dork Polls don’t care much about our feelings. :)

[BlueDevil2K]

NET to 55? That surprises me a bit...a big win over Wake and a close win over UVa are each worth six spots?

[ice-9]

Could be other teams lost or didn't do as well as expected?

[CDu]

NET to 55? That surprises me a bit...a big win over Wake and a close win over UVa are each worth six spots?

The win was close enough to the expected margin that we didn’t move much. In Torvik, we moved from 39 to 36. In KrnPom, our defense jumped from 85 to 79, but our offensive efficiency barely moved (we performed really close to our expected efficiency. But UNC played a depleted Louisville team and performed better than “expected”, and passed us. So our ranking didn’t shift favorably because we weren’t close enough to the teams above us.

It was an important win, but it was very close to what we were expected to do.

[Bluedog]

It was an important win, but it was very close to what we were expected to do.

Yes, if you look at efficiency ratings. However, I would venture a bet that 95% of non-tournament teams would NOT be a 1-point underdog vs UVa so from that perspective, it wasn't close to expected and propels us mightily there, while barely budging our efficiency ranking.

I guess that shows that the dork polls like Duke a lot more than humans do. (Although NET still doesn't love us, perhaps it's because NET weights wins more and also had a 10-point cap.)

[uh_no]

Yes, if you look at efficiency ratings. However, I would venture a bet that 95% of non-tournament teams would NOT be a 1-point underdog vs UVa so from that perspective, it wasn't close to expected and propels us mightily there, while barely budging our efficiency ranking.

I guess that shows that the dork polls like Duke a lot more than humans do. (Although NET still doesn't love us, perhaps it's because NET weights wins more and also had a 10-point cap.)

this is exactly it. dork polls like close losses just as much as close wins.... or almost.

what changed the most is likely our luck went up.

[DavidBenAkiva]

After the Syracuse victory, Duke is up to #28 in T-Rank, #28 on Haslemetrics, #28 on Sagarin, and #30 on KenPom.

Duke is up to 11th in AdjOE on KenPom. Duke has been top 10 in AdjOE every year since the 2009 season on KenPom. This team has figured out how to score. This is not a great defensive team, though they are forcing turnovers and blocking shots.

[JasonEvans]

After the Syracuse victory, Duke is up to #28 in T-Rank, #28 on Haslemetrics, #28 on Sagarin, and #30 on KenPom.

Yeah, but we continue to lag in the NET, which is among the NCAA's most valued rankings. We are now 47th in the NET and need to get that into the top 40s to feel good about getting a tourney bid.

We are 2-3 in Quad 1 games. Both the @GaT and @UNC games will be Quad 1 opportunities. Hosting Louisville, #53 in the NET, is a Quad 2 game (we are 4-3 in Quad 2 games).

[BlueDevil2K]

Yeah, but we continue to lag in the NET, which is among the NCAA's most valued rankings. We are now 47th in the NET and need to get that into the top 40s to feel good about getting a tourney bid.

We are 2-3 in Quad 1 games. Both the @GaT and @UNC games will be Quad 1 opportunities. Hosting Louisville, #53 in the NET, is a Quad 2 game (we are 4-3 in Quad 2 games).

47th, but up 20 spots in 3 games. Given the opponents and the fact that two of the games are on the road, 3 more wins would really help...

[JasonEvans]

47th, but up 20 spots in 3 games. Given the opponents and the fact that two of the games are on the road, 3 more wins would really help...

3 more wins and we are in, not even a question. At that point we'd be like a #8 seed with potential to move higher with a strong ACC tourney run.

To me the question is what happens if we go 2-1 or even 1-2. There are possibilities, depending on the ACC tourney, where we still go dancing in both those scenarios.

[CDu]

Lunardi (again, grain of salt here) now has us as one of the last 4 teams out. I'd imagine that with a win over Louisville we'd move into his field.

[flyingdutchdevil]

Lunardi (again, grain of salt here) now has us as one of the last 4 teams out. I'd imagine that with a win over Louisville we'd move into his field.

Especially given that Lunardi has Louisville as a "Last Four Byes" team.

I cannot believe Duke is here. I will admit that I wrote them off right after that horrid ND game.

[Mtn.Devil.91.92.01.10.15]

I swear I won't mention this again this season,but the fact that people put any credence in Lunardi makes me just a little bit insane.

[throatybeard]

I think I'd actually rather be a 10 or 11 seed than an 8 or 9. Baylor/Gonzaga avoision is devoutly to be wished.

[flyingdutchdevil]

I think I'd actually rather be a 10 or 11 seed than an 8 or 9. Baylor/Gonzaga avoision is devoutly to be wished.

Agreed. But for ratings, I wouldn't be surprised to see us as a 8/9 seed in the Zags bracket. NCAA loves their money!

[DavidBenAkiva]

I think I'd actually rather be a 10 or 11 seed than an 8 or 9. Baylor/Gonzaga avoision is devoutly to be wished.

I would love for Duke to be an 11 seed in this tournament. Avoiding Baylor/Gonzaga as long as possible would be ideal. Michigan is looking scary lately, too. One of those other B1G teams between Ohio State and Illinois looks like they will get a 1 seed, too. The one that gets left out, which I am guessing will be Ohio State, will end up as a 2 seed with either Baylor or Gonzaga. Avoid that region as best you can.

My ideal situation for Duke would be that they get an 11 seed with a team like Villanova or Iowa as the 2 seed and then hope the 1 seed gets upset in the Sweet 16.

[brevity]

I swear I won't mention this again this season,but the fact that people put any credence in Lunardi makes me just a little bit insane.

This season, at least right now, Joe Lunardi's chronicling of Duke's potential rise works to our advantage. Should Duke make the field as an at-large, the casual basketball fan's response goes from "What are they doing there?" to "At least we were warned about this."

Even in past years, when I paid closer attention to college basketball in general and followed most multi-bid conferences throughout the season, there would always be a handful of at-large teams in the field (like an 11-7 SEC team) that were so aggressively average that they still escaped my radar. In some ways, to a non-ACC fan, that's Duke this year.

[budwom]

Since this year has shown (if nothing else) the allure of money with regard to televised sports, I'm sure CBS would be delighted to have Duke in the field...

[CDu]

This season, at least right now, Joe Lunardi's chronicling of Duke's potential rise works to our advantage. Should Duke make the field as an at-large, the casual basketball fan's response goes from "What are they doing there?" to "At least we were warned about this."

Even in past years, when I paid closer attention to college basketball in general and followed most multi-bid conferences throughout the season, there would always be a handful of at-large teams in the field (like an 11-7 SEC team) that were so aggressively average that they still escaped my radar. In some ways, to a non-ACC fan, that's Duke this year.

Yeah, that is a very good point.

To be clear, Lunardi is in no way impressive, especially in terms of aligning with seeds. His schtick was "I get the field almost exactly right", which is not terribly impressive but still serves as a useful barometer. It's less a matter of skill than simple tabulation. But he does at least get the "neighborhood" right.

So in addition to your point, the fact that Lunardi has us so close to the field means we probably really are so close to the field. Maybe we're actually "in" as of right now, maybe we're really in the "next 4 out" rather than "first 4 out", but we're probably not far from where Lunardi has us.

And, obviously, a lot of play left. If we can win at least 2 of the next 3, we're probably in good shape to make the tourney. Winning all 3 would put us in very good position to make the tourney.

[DukieInBrasil]

I think I'd actually rather be a 10 or 11 seed than an 8 or 9. Baylor/Gonzaga avoision is devoutly to be wished.
when said in a Curley (3 Stooges) accent that's actually probably the right word. Otherwise i guess this is a portmanteau of avoidance + aversion ?

[JasonEvans]

My ideal situation for Duke would be that they get an 11 seed with a team like Villanova or Iowa as the 2 seed and then hope the 1 seed gets upset in the Sweet 16.

The 11 plays the 6 in the R64 and then the 3 in the R32. You get the 2-seed in the Sweet 16. I see being a 10 or an 11 as pretty much the same as most of the #2s are not all that different from most of the #3s.

But, I am rooting for Duke to run the table and be a #5 or #6... a kid can dream, can't he?

[camion]

I think I'd actually rather be a 10 or 11 seed than an 8 or 9. Baylor/Gonzaga avoision is devoutly to be wished.

I don't think Gonzaga or Baylor would be thrilled to see an 8/9 Duke in the round of 32.

[throatybeard]

I don't think Gonzaga or Baylor would be thrilled to see an 8/9 Duke in the round of 32.

If I'm Mark Few, one trip to the championship game ever (a loss), and trying to build a legendary rep, heading one of the two clear best teams in the country, I'd love to beat [a somewhat marginal] Duke along the way. 1s often have to play someone who's matchup toxin, or underseeded, or or or. Weird stuff happens. You gotta play who's in front of you. I'd rather be able to say I beat Duke--Mike Krzyzewski--on the way to a title.

[CDu]

If I'm Mark Few, one trip to the championship game ever (a loss), and trying to build a legendary rep, heading one of the two clear best teams in the country, I'd love to beat [a somewhat marginal] Duke along the way. 1s often have to play someone who's matchup toxin, or underseeded, or or or. Weird stuff happens. You gotta play who's in front of you. I'd rather be able to say I beat Duke--Mike Krzyzewski--on the way to a title.

The counterargument is that wherever we are seeded is likely to be an underseeding. We are playing FAR better basketball right now than we have over the course of the season. Torvik says that over our last 10 games, we've played like a 3 or 4 seed. So basically unless we miss the tournament by losing 2+ of the remaining regular season games, we're going to be almost certainly seeded several spots below where we are playing. An 8/9 seeded Duke means we're probably playing like a 4 seed. If we win out including the ACC tournament, that would mean we're playing like a 1 or 2 seed but would likely only get a 5 seed or so. So in Gonzaga's case, while getting a chance to scalp Duke sounds nice, facing the equivalent of a 5 seed in the second round seems a really unfortunate circumstance for the likely #1 overall seed.

Yes, they would still be favored, but I doubt that is a matchup they want in round 2.

If we are playing like an 8/9 seed come tournament time, we probably don't make the tournament, or at best it is in a play-in game.

[Kedsy]

Big test for the Dork polls seems to be in the making. Loyola-Chicago is ranked #11 in the NET, #10 in KenPom, #14 in Torvik, #16 in BPI, #21 in the AP poll, and #22 in the Coaches poll. But the mocks like Lunardi and USA Today are giving them an 8-seed, presumably because of their scheduling/lack of quality wins (Lunardi puts them in Gonzaga's subregion; USA Today puts them with Baylor).

If they're really a top 10 or 15 team, how unfair is it to make them an 8-seed? For that matter, how fair is it to the 1-seed to have to play a top-tier team in the second round? If they're really 8-seed quality, what's the point of the NET?

[uh_no]

Big test for the Dork polls seems to be in the making. Loyola-Chicago is ranked #11 in the NET, #10 in KenPom, #14 in Torvik, #16 in BPI, #21 in the AP poll, and #22 in the Coaches poll. But the mocks like Lunardi and USA Today are giving them an 8-seed, presumably because of their scheduling/lack of quality wins (Lunardi puts them in Gonzaga's subregion; USA Today puts them with Baylor).

If they're really a top 10 or 15 team, how unfair is it to make them an 8-seed? For that matter, how fair is it to the 1-seed to have to play a top-tier team in the second round? If they're really 8-seed quality, what's the point of the NET?

There are two things that are sure in this world:

- whatever the most sensible choice is, the committee will do something else
- whatever the committee does do, lunardi will have likely gotten it wrong

[Nugget]

There are two things that are sure in this world:

- whatever the most sensible choice is, the committee will do something else
- whatever the committee does do, lunardi will have likely gotten it wrong

I've found historically that the AP poll does as good or better a job of predicting what the Committee will do re: seeding at the top of the bracket as any of the metrics. So, looked at that way, for Loyola to be ranked around 20 but generally mocked as an 8 seed (not just by Lunardi -- the Bracket Matrix consensus has them an 8 also) isn't too far off. Maybe Loyola is really a 6 seed rather than an 8. I don't think the NCAA has ever suggested they would strictly go by the NET rankings.

[brevity]

If I'm Mark Few, one trip to the championship game ever (a loss), and trying to build a legendary rep, heading one of the two clear best teams in the country, I'd love to beat [a somewhat marginal] Duke along the way. 1s often have to play someone who's matchup toxin, or underseeded, or or or. Weird stuff happens. You gotta play who's in front of you. I'd rather be able to say I beat Duke--Mike Krzyzewski--on the way to a title.

I'm sure the undefeated Wichita State team of 2014 had the same can-do attitude when they faced upstart 8-seed Kentucky.

[sagegrouse]

There are two things that are sure in this world:

- whatever the most sensible choice is, the committee will do something else
- whatever the committee does do, lunardi will have likely gotten it wrong

You left out:

The Tournament Selection Committee is on a fool's errand. This is a problem every year, but in the time of COVID, it is worse than ever. There were very few non-conference match-ups that had any meaning, so the TSC is trying to create seeds among teams in different conferences, when the only evidence comes from games between schools in the same conference.

[throatybeard]

I'm sure the undefeated Wichita State team of 2014 had the same can-do attitude when they faced upstart 8-seed Kentucky.

Not even a remotely comparable situation.

Kentucky 2014 spent the entire regular season in the top 25, falling out at the very end, dropping from 14th to unranked in a 3-4 stretch that included two losses to third-ranked and then top-ranked Florida. They won two in the SECT and then ran afoul of Florida a third time in the final. That final stretch wasn't really that bad (a couple non-Florida conference losses though), but it took a season-long top 25 team down to the tier of the just unranked--which of course is where sevens and eights live. So maybe they should have been a seven. They were never even remotely in danger of missing the NCAAT.

However this Duke team finishes, the Miami loss was a really bad one, and Duke has spent the entire month of February in severe NCAAT jeopardy, with a possible happy outcome starting to come into view.

Wichita State lost a one possession game to Kentucky. WSU was pretty good but widely regarded as the weakest 1--both got sent to the West, usually a punishment for eastern and Midwestern teams who stumbled late. People were definitely not writing metrics-based articles on 538 about WSU as one of a top two, down from which there's one of the biggest gaps between teams two and three in the last two decades. And that's where Gonzaga is.

Eights do upset ones, and late February Duke is better than Groundhog Day Duke. But this comparison is way off.

Also, you should have to play someone good in the second round. Eights and nines are supposed to have roughly the 29th-36th best resumes in the whole country, surrounding the 90th percentile. I'm pretty sure Mark Few knows that, and if things break bad, he's not going to be whining about how the nefarious CBS-genunflecting NCAA seeding committee conspired to screw him with Duke.

[JasonEvans]

This will knock Baylor down quite a bit on the KenPoms of the world...should be a Baylor 25+ point win.

Baylor had an efficiency margin of +34.84 coming into the ISU game. Today, their EM is down to +33.00. That is a fairly large move but still gives Baylor a pretty big leg up on the #3 team, Michigan, which has an EM +30.59. Gonzaga is in another stratosphere at +38.30.

To put that +38.30 in perspective, here are the annual EM leaders over the past decade. These are all post-tournament numbers, so not exactly a good match to pre-tournament figures but still...

2020- Kansas +30.23, Gonzaga +26.95 (last year was a pretty weak year at the top in college basketball, would have made for a wild tourney... oh well)
2019 - Virginia +34.22, Gonzaga +32.85, Mich St +30.81, Duke +30.62
2018 - Villanova +33.76, Virginia +29.53, Duke +28.86
2017 - Gonzaga +32.05, Villanova +29.88
2016 - Villanova +32.01, UNC +29.82
2015 - Kentucky +36.91, Wisconsin +33.72, Duke +32.48
2014 - Louisville +30.41, Arizona +30.11
2013 - Louisville +32.92, Florida +31.18
2012 - Kentucky +32.59, Ohio St +30.07
2011 - Ohio St +33.47, Duke +28.42 (what about only games in which Kyrie played?)
2010 - Duke +33.29, Kansas +31.85

As you can see, Baylor would be the top team most years in the past, though not every time. This year's Gonzaga though has a higher EM than any team over the past decade. In fact, if you go back even further in Ken's database (back to 2002), you cannot find any year where Gonzaga's +38EM would not be easily the highest in the land (wish I could find numbers on 1999 Duke!!). We will see if it holds up for Gonzaga into the post-season, but it would appear the Zags have a real shot at being a truly legendary team this year and going down as one of the greats in the history of the sport.

-Jason "and yes, I will be cross posting this into the dork poll thread seeing as it is about 98% dork poll content" Evans

[House P]

Baylor had an efficiency margin of +34.84 coming into the ISU game. Today, their EM is down to +33.00. That is a fairly large move but still gives Baylor a pretty big leg up on the #3 team, Michigan, which has an EM +30.59. Gonzaga is in another stratosphere at +38.30.

To put that +38.30 in perspective, here are the annual EM leaders over the past decade. These are all post-tournament numbers, so not exactly a good match to pre-tournament figures but still...

2020- Kansas +30.23, Gonzaga +26.95 (last year was a pretty weak year at the top in college basketball, would have made for a wild tourney... oh well)
2019 - Virginia +34.22, Gonzaga +32.85, Mich St +30.81, Duke +30.62
2018 - Villanova +33.76, Virginia +29.53, Duke +28.86
2017 - Gonzaga +32.05, Villanova +29.88
2016 - Villanova +32.01, UNC +29.82
2015 - Kentucky +36.91, Wisconsin +33.72, Duke +32.48
2014 - Louisville +30.41, Arizona +30.11
2013 - Louisville +32.92, Florida +31.18
2012 - Kentucky +32.59, Ohio St +30.07
2011 - Ohio St +33.47, Duke +28.42 (what about only games in which Kyrie played?)
2010 - Duke +33.29, Kansas +31.85

As you can see, Baylor would be the top team most years in the past, though not every time. This year's Gonzaga though has a higher EM than any team over the past decade. In fact, if you go back even further in Ken's database (back to 2002), you cannot find any year where Gonzaga's +38EM would not be easily the highest in the land (wish I could find numbers on 1999 Duke!!). We will see if it holds up for Gonzaga into the post-season, but it would appear the Zags have a real shot at being a truly legendary team this year and going down as one of the greats in the history of the sport.

-Jason "and yes, I will be cross posting this into the dork poll thread seeing as it is about 98% dork poll content" Evans

[Response copied from other thread because this is a more appropriate thread.]

According to KenPom, the 99 Duke team had a post-tourney EM of +42.7!

The 2001 Duke team had a post-tourney EM of +37.3.

As others have pointed out, it is difficult to compare EM's across seasons, but one way to think about it is that the 99 Duke team was 30.8 points per game better than the average D1 team that season. (42.7 points per 100 possessions * 72.3 possessions per game = 30.8 points per game).

[Kedsy]

Baylor had an efficiency margin of +34.84 coming into the ISU game. Today, their EM is down to +33.00. That is a fairly large move but still gives Baylor a pretty big leg up on the #3 team, Michigan, which has an EM +30.59. Gonzaga is in another stratosphere at +38.30.

To put that +38.30 in perspective, here are the annual EM leaders over the past decade. These are all post-tournament numbers, so not exactly a good match to pre-tournament figures but still...

2020- Kansas +30.23, Gonzaga +26.95 (last year was a pretty weak year at the top in college basketball, would have made for a wild tourney... oh well)
2019 - Virginia +34.22, Gonzaga +32.85, Mich St +30.81, Duke +30.62
2018 - Villanova +33.76, Virginia +29.53, Duke +28.86
2017 - Gonzaga +32.05, Villanova +29.88
2016 - Villanova +32.01, UNC +29.82
2015 - Kentucky +36.91, Wisconsin +33.72, Duke +32.48
2014 - Louisville +30.41, Arizona +30.11
2013 - Louisville +32.92, Florida +31.18
2012 - Kentucky +32.59, Ohio St +30.07
2011 - Ohio St +33.47, Duke +28.42 (what about only games in which Kyrie played?)
2010 - Duke +33.29, Kansas +31.85

As you can see, Baylor would be the top team most years in the past, though not every time. This year's Gonzaga though has a higher EM than any team over the past decade. In fact, if you go back even further in Ken's database (back to 2002), you cannot find any year where Gonzaga's +38EM would not be easily the highest in the land (wish I could find numbers on 1999 Duke!!). We will see if it holds up for Gonzaga into the post-season, but it would appear the Zags have a real shot at being a truly legendary team this year and going down as one of the greats in the history of the sport.


Here are pre-Tournament numbers:

2020- Kansas +30.23, Gonzaga +26.95
2019 - Virginia +35.66, Gonzaga +32.79, Duke +31.99, Mich St +31.36,
2018 - Virginia +32.15, Villanova +31.41, Duke +29.13
2017 - Gonzaga +33.05, Villanova +30.77
2016 - Kansas +29.80, Mich St +29.09
2015 - Kentucky +37.43, Wisconsin +33.53, Arizona +32.31, Villanova +30.96, Virginia +30.89, Duke +29.31
2014 - Louisville +30.57, Arizona +30.40
2013 - Florida +31.98, Louisville +31.14
2012 - Kentucky +31.73, Ohio St +29.90
2011 - Ohio St +32.23, Duke +29.55*
2010 - Kansas +32.51, Duke +31.55

From 2002 to 2009, three teams had pre-Tourney EM above Baylor's 33: 2008 Kansas (33.96), 2005 Illinois (33.31), and 2002 Duke (34.02).

2015 Kentucky's 37.43 is in the same ballpark as 2021 Gonzaga's 38.30, and their lead over #2 (3.9) isn't too far off Gonzaga's lead of 5.3. And we all know what happened to UK in 2015.

Since 2002, the pre-tournament #1 team in KenPom has won the national championship three (3) times (Kansas in '08; Kentucky in '12; Virginia in '19). In fact, the pre-tournament #1 KenPom team has only made the Final Four in seven (7) of the last 18 (since 2002) tournaments. So I wouldn't go ordering my Gonzaga national champion t-shirt just yet.


* what about only games in which Kyrie played? I can't find KenPom's rankings the day after Kyrie got hurt, but I know Duke was #2 on 12/31 (behind Ohio State), and I also know that our EM went down during the last three games Kyrie played, so I doubt our EM only in games Kyrie played was earth-shatteringly special.


.

[House P]

what about only games in which Kyrie played? I can't find KenPom's rankings the day after Kyrie got hurt, but I know Duke was #2 on 12/31 (behind Ohio State), and I also know that our EM went down during the last three games Kyrie played, so I doubt our EM only in games Kyrie played was earth-shatteringly special.
.

Here are the adjusted EMs for Duke's 2010-11 games with Kyrie.




Adj Offense
Adj Defense
AdjEM


Princeton
130.3
75.7
54.5


Miami OH
110.3
65.4
44.9


Colgate
125.6
79.4
46.2


Marquette
110.5
85.9
24.5


Kansas St.
119.8
80.9
38.9


Oregon
121.5
79.6
41.9


Michigan St.
120.5
101.5
18.9


Butler
121.4
88.3
33.2


AVG-1st 8 games
120.0
82.1
37.9


Hampton
133.1
77.3
55.8


Michigan
126.4
108.3
18.1


Arizona
115.9
119.2
-3.2


AVG-NCAA Tourney
125.1
101.6
23.5


AVG- ALL 11 Games
121.4
87.4
34.0



Clearly the team was playing fantastic basketball during the first 8 games with Kyrie.

However, I should point out the an average EM of 37.9 over 8 games does not necessarily translate to a KenPom rating of 37.9. Ken would "de-value" the guady AdjEM numbers put up against overmatched opponents like Miami-OH and Colgate. So Duke's actual KenPom rating after 8 games would be a bit lower than 37.9.

On the other hand, Princeton ended the season as KenPom's #79 team. So, KenPom would consider Duke's 37 point win vs the Tigers to be as impressive as any game Duke played that season.

[Kedsy]

Here are the adjusted EMs for Duke's 2010-11 games with Kyrie.




Adj Offense
Adj Defense
AdjEM


Princeton
130.3
75.7
54.5


Miami OH
110.3
65.4
44.9


Colgate
125.6
79.4
46.2


Marquette
110.5
85.9
24.5


Kansas St.
119.8
80.9
38.9


Oregon
121.5
79.6
41.9


Michigan St.
120.5
101.5
18.9


Butler
121.4
88.3
33.2


AVG-1st 8 games
120.0
82.1
37.9


Hampton
133.1
77.3
55.8


Michigan
126.4
108.3
18.1


Arizona
115.9
119.2
-3.2


AVG-NCAA Tourney
125.1
101.6
23.5


AVG- ALL 11 Games
121.4
87.4
34.0



Clearly the team was playing fantastic basketball during the first 8 games with Kyrie.

However, I should point out the an average EM of 37.9 over 8 games does not necessarily translate to a KenPom rating of 37.9. Ken would "de-value" the guady AdjEM numbers put up against overmatched opponents like Miami-OH and Colgate. So Duke's actual KenPom rating after 8 games would be a bit lower than 37.9.

On the other hand, Princeton ended the season as KenPom's #79 team. So, KenPom would consider Duke's 37 point win vs the Tigers to be as impressive as any game Duke played that season.

Our adj EM for the entire season (including Kyrie's games) was 29.6, so if KenPom's average was a bit lower in the early Kyrie games and thus the total for 11 games was also a bit lower (say, 32 or 33), that would match the "eye test" that the team with Kyrie was better but not otherworldly better than the team for the entire season.

[proelitedota]

(wish I could find numbers on 1999 Duke!!).

Kenpom did historical data for fun a few years back. 1999 Duke was 40+.

[DavidBenAkiva]

Kenpom did historical data for fun a few years back. 1999 Duke was 40+.

Both the 99 and 01 Duke teams crested +40 AdjEM within their seasons.


2 teams have breached +40: 99 Duke and 01 Duke
2 more have breached +38: 15 Kentucky and 13 Florida
14 more got to +36: 19 Virginia, 19 Duke, 15 Virginia, 11 Ohio St, 10 Duke, 10 Kansas, 08 Kansas, 06 Duke, 01 Stanford, 00 Cincinnati, 98 Duke, 98 UNC, 97 Kentucky, 97 Wake Forest

source (https://twitter.com/kenpomeroy/status/1357116430093717505)

I find this list fascinating for its lack of National Championships. Duke won in '01 as did '99 Virginia, '10 Duke, and '08 Kansas. The other 14 didn't and many fell short of the Final Four.

Always take the field.

[CDu]

Both the 99 and 01 Duke teams crested +40 AdjEM within their seasons.



source (https://twitter.com/kenpomeroy/status/1357116430093717505)

I find this list fascinating for its lack of National Championships. Duke won in '01 as did '99 Virginia, '10 Duke, and '08 Kansas. The other 14 didn't and many fell short of the Final Four.

Always take the field.

Yes, it's simple math. Even a team that was a 90% favorite in every game (which would be absurd) would only be a coin flip (53%) chance of winning the title.

If we assume that the best team is a 99% chance in round 1, 90% chance in round 2, 80% chance in round 3, 75% chance in round 4 and 5, and 70% chance in the final (which would be a fairly heavy favorite), that's only a 28% chance of winning the title.

The VERY best teams are typically at most like a 40% chance of winning. That doesn't mean the top 1 seed in a given year. That means "team on the short list of best ever". Typical top 1 seeds are probably more like a 20-30% chance.

So yes, always take the field. Six-game single-elimination tournaments mean the season's best team wins fairly rarely.

[tommy]

Yes, it's simple math. Even a team that was a 90% favorite in every game (which would be absurd) would only be a coin flip (53%) chance of winning the title.

If we assume that the best team is a 99% chance in round 1, 90% chance in round 2, 80% chance in round 3, 75% chance in round 4 and 5, and 70% chance in the final (which would be a fairly heavy favorite), that's only a 28% chance of winning the title.

The VERY best teams are typically at most like a 40% chance of winning. That doesn't mean the top 1 seed in a given year. That means "team on the short list of best ever". Typical top 1 seeds are probably more like a 20-30% chance.

So yes, always take the field. Six-game single-elimination tournaments mean the season's best team wins fairly rarely.

Yup, according to these sites, a given #1 seeded team has only about a 15% chance to win the championship.

https://www.betfirm.com/seeds-national-championship-odds/

https://www.boydsbets.com/bracket-tips-by-seed/

[Kedsy]

I posted this in another thread, but since it's full-on dork poll material, I figured I should cross-post it here.

I've been playing around with Torvik game scores. He has data back to 2008.

The first two columns are regular season avg and std dev; next two columns are avg and SD for post-season; last column is average for first two games of the postseason.



Year Avg StDev postAvg postSD 1st 2 ps avg
2008 93.0 10.2 80.8 16.4 92.0
2009 90.5 16.7 81.5 27.1 88.0
2010 93.6 10.5 97.0 2.3 94.0
2011 92.7 11.1 94.2 9.1 97.0 w/ Kyrie: 95.3
2012 87.4 14.3 76.3 15.5 85.0
2013 90.4 14.1 84.6 19.7 73.0 w/ Ryan Kelly: 93.3
2014 91.4 10.8 82.3 14.9 92.5
2015 92.3 11.5 94.8 12.0 82.5
2016 88.6 11.9 81.4 12.5 80.5
2017 89.6 12.7 94.5 3.8 95.0
2018 91.7 9.7 91.7 9.2 86.0
2019 92.3 12.8 94.9 2.0 96.0 w/ Zion: 95.3
2020 86.7 17.1 n/a n/a n/a
2021 73.2 22.0 98.0 1.4 98.0 w/o Jalen: 77.8


It wasn't just wins and losses; this was by far the worst Duke team, definitely since 2008 and almost certainly since 1995. And we weren't that much better without Jalen than we were with him. But, FWIW, our two post-season games were the best first two post-season games since at least 2008, so we got that going for us.

[proelitedota]

Duke is the highest team on Kenpom and Bartorvick list to be left out of the field.

We need to play angry next season.

[uh_no]

Duke is the highest team on Kenpom and Bartorvick list to be left out of the field.

We need to play angry next season.

we need to win games.

[Mtn.Devil.91.92.01.10.15]

we need to win games.

We need to play games.

[throatybeard]

Apparently we're a bubbling cauldron of need.