Bracketology

[DavidBenAkiva]

I think you might be understating the current difference between Kansas and Duke right now. They are 10-4 in Q1 games, whereas we are 3-4. As you note, we will have potentially 4 more such games, and 1 of our Q2 games could bump up to a Q1 game (N-Texas). We might also see St. John's drop out of a Q1, but moving that loss to a Q2 may be worse for our overall profile, so I'm not sure that helps the cause ultimately. As you note, the Jayhawks have 2 Q2 games and a Q1 road game (against a top-20 team) left on their regular season schedule. Even if they lose out in their Q1 games (i.e., go 0-2), they'll be sitting at 10-6 in Q1 and let's say 9-1 in Q2 games. Our best case is a 7-4 mark, but that basically requires winning out.

In other words, it would largely take us winning out and Kansas losing their remaining Q1 games for us to clearly sit ahead of them in the brackets based on the quadrant system. Not saying that can't happen, but it's going to take a lot of things breaking correctly for us to pass Kansas.

Xavier would in theory be a little easier to catch, as they are out of Q1 games until their tournament. But again, they are sitting pretty at 7-4 in Q-1 games and 8-0 in Q2 games compared to our 3-4 and 8-1. So we'd likely need to win out to pass them by as well.

Basically, we are really being hurt right now by not doing better in Q1 games. The composite of the metrics-based systems have us as comfortably the #3 team in the country right now. But our W/L results just haven't quite worked out. Too many close losses. If even two of our 5-point-or-less Q1 losses are flipped, we're probably sitting on the one line. So we are probably 4-5 possessions from being on the 1 line (of course, we're also probably 4-5 possessions from being on the 3 line too). Such is the challenge of using the quadrant system. It really values winning the games over beating the spread. And right now, we just don't have the wins.

This is all true, but I also think that the rankings themselves are a factor in the seeding. It's not like the committee is only going to look at Q1, Q2, Q3 record, etc. It's one of several factors. And the fact that Duke is ranked higher than Kansas, MSU, etc. means something. My point is that I think a comparable record like 7-5 or 8-4 in Q1 in combination with our excellent rankings could push us into a 1 seed.

[CDu]

This is all true, but I also think that the rankings themselves are a factor in the seeding. It's not like the committee is only going to look at Q1, Q2, Q3 record, etc. It's one of several factors. And the fact that Duke is ranked higher than Kansas, MSU, etc. means something. My point is that I think a comparable record like 7-5 or 8-4 in Q1 in combination with our excellent rankings could push us into a 1 seed.

I don't think they are going to look ONLY at the quadrant results. But their first pass sure looks like they are giving it the most weight. And the "team sheets" that will be used (where the metric-based stats are squeezed to the very top edge of the page while the quadrant stuff fills the page) sure look like they will continue to weight that most heavily.

[dragoneye776]

And again, I will note that we have never won a tournament game in the Pacific time zone. Yes, that is only a four-game sample, but I believe the numbers for teams crossing 3 time zones is not good. If I have time, I will try to run those numbers. But I don’t think it is as simple as saying “we want the 1 seed, regardless of location.”

And then there is the point that 1 seeds are historically better than two seeds, so that isn’t an apples to apples application to this question of “holding the talent of the team constant, which seed do we want?”

We almost assuredly want the 1 seed in the Midwest (only 1 time zone difference) over the top 2 seed. But I don’t think the answer is so clearly the same when that 1 seed is moved out West.

Yes, good points. The 1 seed in the Midwest is definitely the best option. But for all the reasons you and others mentioned in which we may not get it (it instead going to Kansas, MSU, or even Xavier), do we want a closer 2 seed or the 1 seed out West?

[CDu]

Yes, good points. The 1 seed in the Midwest is definitely the best option. But for all the reasons you and others mentioned in which we may not get it (it instead going to Kansas, MSU, or even Xavier), do we want a closer 2 seed or the 1 seed out West?

I'm on board as being concerned about being a 1 seed in the West. We've never won a tourney game in the Pacific time zone. That in and of itself isn't a deal breaker, but it gives me great pause in saying which I prefer. Teams generally play worse when asked to play across 3 time zones.

[devilseven]

I'm on board as being concerned about being a 1 seed in the West. We've never won a tourney game in the Pacific time zone. That in and of itself isn't a deal breaker, but it gives me great pause in saying which I prefer. Teams generally play worse when asked to play across 3 time zones.

More recently, we did pretty well in the Pacific time zone during the PK80 tournament. We have also done very well in the Maui tournament which is even further away/

[CDu]

More recently, we did pretty well in the Pacific time zone during the PK80 tournament. We have also done very well in the Maui tournament which is even further away/

Yes, we did well in those tournaments. Perhaps I need to clarify what I meant: teams playing at a 3-time-zone deficit tend to play poorly compared to their opposition. The Maui invitational is a challenge for all of the teams because everyone is at a huge time-zone deficit and everyone is playing a ton of games in a short time. And in Portland, we played Florida and Texas, neither of whom had a time zone advantage.

If we would be paired with exclusively ET and CT opponents in the West, I'd be fine. But if we get, say, Arizona? Suddenly we are playing at that deficit, and likely to be playing close to a true road game as well.

So to me it is still very much an unknown as to which I'd prefer. And the answer would largely come down to what the bracket itself looks like.

[TexHawk]

In other words, it would largely take us winning out and Kansas losing their remaining Q1 games for us to clearly sit ahead of them in the brackets based on the quadrant system. Not saying that can't happen, but it's going to take a lot of things breaking correctly for us to pass Kansas.

Also worth nothing that two KU opponents (West Virginia and Oklahoma) are sitting on the fringe of the top 30. WVU is #31, close to giving KU another Q1 win. OU is #34, but are heading in the wrong direction, so I'm not particularly hopeful. Kansas State and Texas are also fringe top 50, and if they could jump a couple spots, that would make 7 of the 10 teams in the B12 conference tournament "Q1 win or loss eligible", if that makes sense.


Tangent re: MSU and Purdue. The Big10 tournament is being held a week early this year, given scheduling issues with MSG. Does anyone think that potentially hurts the winner in seeding discussions? The previous weekend is usually devoted to smaller conferences, and when the big boys start playing, the power tournaments tend to block out the sun, so to speak. As it is, the Big10 semis will be held the same day as Duke/UNC and other big games.

I don't *really* think it will hurt the winner, given the data the committee seems to be weighing, but taking the week off certainly won't help. In the old days, I bet it would have. Imagine Duke and MSU fighting for the last #1, and the night before the bracket announcement, Duke beats UVA handily in prime time, while MSU was last on TV a week earlier.

(10 days off before the big tournament begins is another concern altogether.)

[Duke07]

I actually think it’ll help the B10 winner if anything- adds another quality win. Our head to head win in the ACC tourney last year against UNC failed to make a difference- we still ended up on the 2 line w UNC on the 1. I think the Saturday/Sunday tourney championship games often have less of an impact than we think bc the committee is having to also fill out the rest of the bracket including the bubble teams. Most of the time it seems that they’ve already decided who the top 8 seeds are and where they want to place them.

Also worth nothing that two KU opponents (West Virginia and Oklahoma) are sitting on the fringe of the top 30. WVU is #31, close to giving KU another Q1 win. OU is #34, but are heading in the wrong direction, so I'm not particularly hopeful. Kansas State and Texas are also fringe top 50, and if they could jump a couple spots, that would make 7 of the 10 teams in the B12 conference tournament "Q1 win or loss eligible", if that makes sense.


Tangent re: MSU and Purdue. The Big10 tournament is being held a week early this year, given scheduling issues with MSG. Does anyone think that potentially hurts the winner in seeding discussions? The previous weekend is usually devoted to smaller conferences, and when the big boys start playing, the power tournaments tend to block out the sun, so to speak. As it is, the Big10 semis will be held the same day as Duke/UNC and other big games.

I don't *really* think it will hurt the winner, given the data the committee seems to be weighing, but taking the week off certainly won't help. In the old days, I bet it would have. Imagine Duke and MSU fighting for the last #1, and the night before the bracket announcement, Duke beats UVA handily in prime time, while MSU was last on TV a week earlier.

(10 days off before the big tournament begins is another concern altogether.)

[atoomer0881]

Also worth nothing that two KU opponents (West Virginia and Oklahoma) are sitting on the fringe of the top 30. WVU is #31, close to giving KU another Q1 win. OU is #34, but are heading in the wrong direction, so I'm not particularly hopeful. Kansas State and Texas are also fringe top 50, and if they could jump a couple spots, that would make 7 of the 10 teams in the B12 conference tournament "Q1 win or loss eligible", if that makes sense.


Tangent re: MSU and Purdue. The Big10 tournament is being held a week early this year, given scheduling issues with MSG. Does anyone think that potentially hurts the winner in seeding discussions? The previous weekend is usually devoted to smaller conferences, and when the big boys start playing, the power tournaments tend to block out the sun, so to speak. As it is, the Big10 semis will be held the same day as Duke/UNC and other big games.

I don't *really* think it will hurt the winner, given the data the committee seems to be weighing, but taking the week off certainly won't help. In the old days, I bet it would have. Imagine Duke and MSU fighting for the last #1, and the night before the bracket announcement, Duke beats UVA handily in prime time, while MSU was last on TV a week earlier.

(10 days off before the big tournament begins is another concern altogether.)

That's actually my biggest "concern" for B1G teams. For any team that doesn't make it to the championship game, and say, loses in the semis, that would mean their last game is March 3. If they play a Friday game in the First Round of the Tournament, that's March 16. That's almost two full weeks off between games. On the one hand, if you are nursing injuries, this could be valuable time off. On the other hand, if you are riding momentum, this lull could be devastating. I am very curious to see how B1G teams fare this year in the Tournament. I really wonder if moving up their conference tournament in order to play at the Garden, was really the best decision for these teams.

[Kedsy]

And again, I will note that we have never won a tournament game in the Pacific time zone. Yes, that is only a four-game sample, but I believe the numbers for teams crossing 3 time zones is not good. If I have time, I will try to run those numbers. But I don’t think it is as simple as saying “we want the 1 seed, regardless of location.”

And then there is the point that 1 seeds are historically better than two seeds, so that isn’t an apples to apples application to this question of “holding the talent of the team constant, which seed do we want?”

We almost assuredly want the 1 seed in the Midwest (only 1 time zone difference) over the top 2 seed. But I don’t think the answer is so clearly the same when that 1 seed is moved out West.

WEST REGION, LAST 20 YEARS
------------------------------

Year	West Winner	West Runner Up	West #1 Seed
----	------------	---------------	--------------
2017	Gonzaga		Xavier		Gonzaga
2016	Oklahoma	Oregon		Oregon
2015	Wisconsin	Arizona		Wisconsin
2014	Connecticut	Arizona		Arizona
2013	Wichita State	Ohio State	Gonzaga
2012	Louisville	Florida		Michigan State
2011	Connecticut	Arizona		Duke
2010	Butler		Kansas State	Syracuse
2009	Connecticut	Missouri	Connecticut
2008	UCLA		Xavier		UCLA
2007	UCLA		Kansas   	Kansas
2006	UCLA		Memphis		Memphis
2005	Louisville	West Virginia	Washington
2004	Connecticut	Alabama		Stanford
2003	Kansas		Arizona		Arizona
2002	Oklahoma	Missouri	Cincinnati
2001	Maryland	Stanford	Stanford
2000	Wisconsin	Purdue		Arizona
1999	Connecticut	Gonzaga		Connecticut
1998	Utah		Arizona		Arizona

Nine of the 20 West Region champions have come from the Eastern Time Zone, including Connecticut five times (and in three of those games, UConn beat a Western team in the regional final). I don't have time to run a full winning percentage analysis of Eastern teams in the West region, but I think perhaps we're making too much of Duke's performance in the Pacific time zone. Of the four losses, in two of them we lost to better-seeded teams, and in one as a #3 we lost to a #6 by two points, in our opponents' home state, 34 years ago. So, seems to me the relevant sample isn't even four games, it's one game.

[OldPhiKap]

Didn’t Ryan Kelly have an injury in 2011 late, too? Plus trying to bring Kyrie back?

[CDu]

Nine of the 20 West Region champions have come from the Eastern Time Zone, including Connecticut five times (and in three of those games, UConn beat a Western team in the regional final). I don't have time to run a full winning percentage analysis of Eastern teams in the West region, but I think perhaps we're making too much of Duke's performance in the Pacific time zone. Of the four losses, in two of them we lost to better-seeded teams, and in one as a #3 we lost to a #6 by two points, in our opponents' home state, 34 years ago. So, seems to me the relevant sample isn't even four games, it's one game.

And we might well be. But I think it's reason enough to be a question mark. Especially because I don't think we really have any evidence proving the opposite side of the arguement. Comparing the historical 2 seeds to 1 seeds omits the important variable that 1 seeds are historically also just better teams than 2 seeds (especially at the extremes, the top 2 1 seeds compared with the bottom 2 2 seeds). The best we could do would be to compare the results of the 4th best 1 seed to the best 2 seed. That is at least close to the scenario in question. But even then, that's a little bit of a leaned analysis, as it is still in theory not apples to apples (the #4 team should do better than the #5 team).

What we're really talking about here is whether the 4th best team is better off as the 4th 1 seed in a very nonpreferred geographic region or a 2 seed in a more preferable geographic region. And I don't think we have the data to answer that question. I certainly haven't seen the data. And to be honest, we'd largely be talking about sample sizes similar to 2 (I wouldn't discount the loss to Washington in Washington as that is a roughly comparable scenario with Arizona in LA) or 4 games if we do have the data.

[Kedsy]

Didn’t Ryan Kelly have an injury in 2011 late, too? Plus trying to bring Kyrie back?

Pretty sure that was 2012.

[El_Diablo]

Didn’t Ryan Kelly have an injury in 2011 late, too? Plus trying to bring Kyrie back?

Kelly was injured at the end of the 2012 season (he missed the ACC tournament and NCAA tournament).

[Kedsy]

And we might well be. But I think it's reason enough to be a question mark. Especially because I don't think we really have any evidence proving the opposite side of the arguement. Comparing the historical 2 seeds to 1 seeds omits the important variable that 1 seeds are historically also just better teams than 2 seeds (especially at the extremes, the top 2 1 seeds compared with the bottom 2 2 seeds). The best we could do would be to compare the results of the 4th best 1 seed to the best 2 seed. That is at least close to the scenario in question. But even then, that's a little bit of a leaned analysis, as it is still in theory not apples to apples (the #4 team should do better than the #5 team).

What we're really talking about here is whether the 4th best team is better off as the 4th 1 seed in a very nonpreferred geographic region or a 2 seed in a more preferable geographic region. And I don't think we have the data to answer that question. I certainly haven't seen the data. And to be honest, we'd largely be talking about sample sizes similar to 2 (I wouldn't discount the loss to Washington in Washington as that is a roughly comparable scenario with Arizona in LA) or 4 games if we do have the data.

If you're going to talk about tiny sample sizes, how do you explain Connecticut's success out West? For example, in 2009, UConn had the resume of a #2, got the #1 out West and won the region, one of five times they've made the Final Four from the West in the past 20 years. Connecticut's further East than Durham. No reason it should be our kryptonite but make them stronger, right?

[Duke07]

Scenarios I’d like to avoid- playing Zona out west (likely as a 1 v 4), playing Nova in the east (as a 2) or having MSU as our 2 in a region. Ideal scenario would be Xavier as our 1.

[CDu]

If you're going to talk about tiny sample sizes, how do you explain Connecticut's success out West? For example, in 2009, UConn had the resume of a #2, got the #1 out West and won the region, one of five times they've made the Final Four from the West in the past 20 years. Connecticut's further East than Durham. No reason it should be our kryptonite but make them stronger, right?

UConn is a crazy outlier. I mean, two of those runs they were fairly bad teams. And I'm not saying that it is impossible to succeed out West. Just that I would be interested to look at the evidence a bit closer for borderline 1/2 seeds playing 3 time zones away.

Scenarios I’d like to avoid- playing Zona out west (likely as a 1 v 4), playing Nova in the east (as a 2) or having MSU as our 2 in a region. Ideal scenario would be Xavier as our 1.

The ideal scenario would probably be Xavier as our #2 .

[Kedsy]

UConn is a crazy outlier. I mean, two of those runs they were fairly bad teams. And I'm not saying that it is impossible to succeed out West. Just that I would be interested to look at the evidence a bit closer for borderline 1/2 seeds playing 3 time zones away.

OK, are they more of an outlier than 1984 Duke/Washington (played before Coach K had ever won an NCAA game) or 2011 Duke/Arizona (when Derek Williams had a confirmed out-of-body experience)? If you're going to rely on 2 (or even 4) Duke datapoints, I think 5 opposing UConn datapoints may be legitimately used in response. Especially when one of the UConn datapoints is exactly the scenario (borderline 1/2 seed getting the #1 in the West) that you say you're interested in.

But even if you go back 20 (or 35) years and can successfully identify borderline 1/2s from the East that got a #1 in the West vs. borderline 1/2s from the East that got a #2 somewhere else, your sample size is going to be very, very small.

[CDu]

OK, are they more of an outlier than 1984 Duke/Washington (played before Coach K had ever won an NCAA game) or 2011 Duke/Arizona (when Derek Williams had a confirmed out-of-body experience)? If you're going to rely on 2 (or even 4) Duke datapoints, I think 5 opposing UConn datapoints may be legitimately used in response. Especially when one of the UConn datapoints is exactly the scenario (borderline 1/2 seed getting the #1 in the West) that you say you're interested in.

But even if you go back 20 (or 35) years and can successfully identify borderline 1/2s from the East that got a #1 in the West vs. borderline 1/2s from the East that got a #2 somewhere else, your sample size is going to be very, very small.

I agree. And it may well be that being a #1 is still preferred over all else. I'm just noting that I still have some uncertainty about which I prefer. I'm not leaning strongly in one direction or the other, in large part because of the small sample size of data.

[hallcity]

In case you're interested in ACC Tournament bracketology, here's a website that shows you the brackets if you enter your choices for the winner in the remaining conference games.