Why UConn and ND will NOT Meet in the Tournament

[BigBird]

It's simple. The two teams, even if they are the best on their respective sides of the bracket will play if one and only one set of events comes to pass. On the other hand, they won't meet if any one of many possible contrary events comes to pass. I am not thinking about basketball so much as probability theory. It's just math.

The current thinking is that both teams aren't just better than the field, but are so much better as to nearly guarantee a ND / UConn final. I am not as sure. Let's put it this way, If I had to place a sizable wager (and I wouldn't willingly do so), I would bet the math, not the assumptions of the mass media.

Yeah, I know. I feel the limb bending as I walk further out on it.

 

[cardfan]

I have UConn and TN in the final game. Agree with Geno, when everyone says it will be this team or that team for sure, it won't happen. See Baylor.

 

[Goodgood]

Yes, last year. Baylor vs ND, heavily favored to meet in the finals. Both didn't make it.

 

[Sonny44]

Depends on how valid the assumptions are. In this case there is overwhelming evidence in their support. UCONN vs. ND in the finals. Take it to the bank.

 

[jdubct]

Depends on how valid the assumptions are. In this case there is overwhelming evidence in their support. UCONN vs. ND in the finals. Take it to the bank.

I'd say that the chance of Uconn and ND meeting in the finals is 90%. Sure anything can happen, but the gap between the two teams and the rest of the field makes it highly unlikely that we'll see a Baylor-like loss (and a 1 point loss it was) by either team prior to the final game.

 

[Vowelguy]

It's simple. The two teams, even if they are the best on their respective sides of the bracket will play if one and only one set of events comes to pass. On the other hand, they won't meet if any one of many possible contrary events comes to pass. I am not thinking about basketball so much as probability theory. It's just math.

Actually Math says that the number of alternatives doesn't matter; the probability of those alternatives is what matters. Consider an extreme example: UConn, ND, and 2nd grade teams. It doesn't matter if there were 64 teams or 64000 teams (meaning billions of alternative outcomes), UConn-nd is highly likely to occur.

But, yes, there is a reason we play the games.

 

[diggerfoot]

While vowelguy is correct, not all probabilities are created equal, I do not think sonny is correct in assuming the probabilities overwhelmingly favor a UConn-ND match-up.

Quite frankly, I do think that UConn was given a very favorable path to the championship game. Multiplying all the independent probabilities may not be like the 2nd grade example vowelguy used, but still extremely high, including whatever match-up occurs in the semi-final game. Nebraska at their place and USC and/or Stanford make the overall probability less than 100%, but not much less than that.

ND's path is significantly more difficult in my mind. Their most impressive win is a highly inconsistent Tennessee team that lost three times at home. Count me as one of those who thinks the SEC is getting more credit than they deserve. While the ACC is the best conference top to bottom, ND often played the other top teams when those teams had key injuries and none are noted for their defense, not even ND this year, which is key at tournament time. If people think ND is like Baylor last year I would say they have not been paying attention to who ND was beating, when and how they were beating them.

Baylor has a coach who beat the odds at tournament time to win in 2005. The Griner era is held against her, but in the end this is still a guards game. They could beat the odds again. Louisville has a coach who frequently beats the odds at tournament time. Certainly, the gap between ND and 2014 Louisville is much less than Baylor and 2013 Louisville. They could beat the odds again. Let's say the odds of ND beating either is 75%. Others might legitimately think that higher, but that is my opinion. All together with the other match-ups that makes the odds for ND reaching the championship game a little over 50%. Let's say the odds for UConn are more like 90%, though others might legitimately think that higher as well. That makes the UConn-ND match-up close to a 50-50 proposition, not a sure thing, even though I personally would love to see it happen.

 

[cohenzone]

While vowelguy is correct, not all probabilities are created equal, I do not think sonny is correct in assuming the probabilities overwhelmingly favor a UConn-ND match-up.

Quite frankly, I do think that UConn was given a very favorable path to the championship game. Multiplying all the independent probabilities may not be like the 2nd grade example vowelguy used, but still extremely high, including whatever match-up occurs in the semi-final game. Nebraska at their place and USC and/or Stanford make the overall probability less than 100%, but not much less than that.

ND's path is significantly more difficult in my mind. Their most impressive win is a highly inconsistent Tennessee team that lost three times at home. Count me as one of those who thinks the SEC is getting more credit than they deserve. While the AAC is the best conference top to bottom, ND often played the other top teams when those teams had key injuries and none are noted for their defense, not even ND this year, which is key at tournament time. If people think ND is like Baylor last year I would say they have not been paying attention to who ND was beating, when and how they were beating them.

Baylor has a coach who beat the odds at tournament time to win in 2005. The Griner era is held against her, but in the end this is still a guards game. They could beat the odds again. Louisville has a coach who frequently beats the odds at tournament time. Certainly, the gap between ND and 2014 Louisville is much less than Baylor and 2013 Louisville. They could beat the odds again. Let's say the odds of ND beating either is 75%. Others might legitimately think that higher, but that is my opinion. All together with the other match-ups that makes the odds for ND reaching the championship game a little over 50%. Let's say the odds for UConn are more like 90%, though others might legitimately think that higher as well. That makes the UConn-ND match-up close to a 50-50 proposition, not a sure thing, even though I personally would love to see it happen.

This makes my head hurt.

 

[Icebear]

Start by assigning probability to the first 32 games using Sangarin or more appropriate source then borrow time on your local Super Computer.

 

[UcMiami]

In a normal year I do think the probability is 50/50 or worse, but with ND playing at home for the regional, what would normally be two tough games become 'easier' and so their path to the FF becomes easier. And as we all know anything can happen at the FF. The other side to that is their have been 7 undefeated teams in NCAA history and this year there likely will be an eighth, but the pressure of being undefeated at tournament time just adds an extra level of pressure on a team. Geno and the whole Uconn coaching staff has experience in handling those pressures. I do think that history helps Uconn and we will see how Muffet and ND handle it this year.

 

[DobbsRover2]

Although this year in a real sense is not affected by what has happened in the previous 32 tournaments, it can be helpful to look at the stats for the past years to see if a matchup between the top 2 rated teams is at all likely. If it's never happened before, then it's unlikely the odds are very high; while if it happens with some frequency, then the odds may seem higher. Of course, it's always possible that in a number of years the committee has lined up the top two teams to meet in the semis, so the numbers may not be totally clean. And of course there is a wrinkle this year in that the top two teams are undefeated, thus much stronger favorites.

Checking for meetings between the top 2 ranked teams (fairly likely to be the top 2 teams) in the NC game, it is clear that upsets happen frequently in the tourney, except when they don't. There have been 5 times in 32 when the top 2 teams faced off -- 1982, 1983, 1989, 2000, and 2010. So three times at the beginning back in the 1980s and twice more recently. But there have also been 3 times in 1986, 1995, and 2003 when USC or UTenn were officially ranked at #3 but I have a gut feeling they had a large number of votes to be in the NC game among the bracketers based on past tourney history.

8 of 32 times where possibly the top two teams made it to the NC game comes out at 25%, and when you add in the undefeated status of the two teams and certain favorable venues getting there this year, the odds are probably pretty good, as in many years the top two teams are of a much less clear status.

 

[ndmb]

Start by assigning probability to the first 32 games using Sangarin or more appropriate source then borrow time on your local Super Computer.

I did a simplified version of just that, actually, but with Massey. If UConn and Notre Dame play the highest seed possible in every game:

UConn:
Prairie View: ~100%
Georgia: ~100%
@Nebraska: 99%
Duke: 96%
South Carolina: 98%

Total Massey odds of getting to the final: At least 93.1%

Notre Dame:
Robert Morris: ~100%
Vanderbilt: 99%
Purdue (at ND): 98%
Baylor (at ND): 91%
Tennessee: 90%

Total Massey odds of getting to the final: At least 79.5%

Total Massey odds that the two teams meet: At least 74%

Btw, ND gets a pretty big bump from hosting the regionals in Massey's system (it would be 87% over Baylor at a neutral court) and UConn takes a small hit from potentially playing Nebraska at home. And if they do meet in the final:

Massey odds that UConn beats ND: 74%

 

[sparky]

I have such a headache with all these #'s. Plain and simple UConn all the way

 

[BigBird]

Yeah, this stuff gives me a headache too. So here's a snapshot, and the thought that got me started on this. If both UConn and ND are 90% likely to win their sides of the bracket, the probability is "only" 80% that they both come through to the final. But is the 90% guess too high or too low? The resultant joint probability swings drastically as you fiddle with the individual success estimates. Example: If Fresno upsets Nebraska in LA, one of the better teams that UConn could have played will never face them. In this case, 90% might be too low for the Huskies. This is what makes probability analysis so difficult here. The events are not independent of each other. The key point is that if you bet one team and give the field (of 32) and you do it twice, you are facing some odds that are longer than they might have first appeared.

 

[ndmb]

Yeah, this stuff gives me a headache too. So here's a snapshot, and the thought that got me started on this. If both UConn and ND are 90% likely to win their sides of the bracket, the probability is "only" 80% that they both come through to the final. But is the 90% guess too high or too low? The resultant joint probability swings drastically as you fiddle with the individual success estimates. Example: If Fresno upsets Nebraska in LA, one of the better teams that UConn could have played will never face them. In this case, 90% might be too low for the Huskies. This is what makes probability analysis so difficult here. The events are not independent of each other. The key point is that if you bet one team and give the field (of 32) and you do it twice, you are facing some odds that are longer than they might have first appeared.

I'm nitpicking here, but it'd actually be 81% in the scenario you described.

The only way to do it would be what Icebear suggested. You come up with the odds of winning every single game and extrapolate. For example, let's assume you have four teams A playing B, C playing D, and a winner game. The odds that A wins the first game are 100%, the odds of C beating D are 75%. The odds of A beating C are 50% and the odds that A beats D are 100%. The formula to figure out the odds that A gets through is:

Odds of C beating D (.75) * the odds of A beating C (.5) + the odds of D beating C (.25) * the odds of A beating D (1). Which would give you 63%. For a 64 team format, it gets pretty messy.

 

[BigBird]

Right on all accounts, NDMB. .9 x .9 = .81 every time.

(What? Notre Dame Marching Band?)...

 

[Zorro]

The first two rounds are walkovers. Then in the third round, ND plays Ok. St., or someone who has beaten Ok St. Maybe they have a five percent chance of losing that game. If they win, then they play Baylor, or someone who has beaten Baylor. Getting ready to duck, I would give Baylor a .25 chance. If they get by OkSt and Baylor (or) then they will play in the fifth game the survivor of Md., WVU, L'ville and the LV. Lets give them a 60% probability of winning that one. Multiplying the (my guestimated) probabilities of winning those three games; .95 x .75 x .6 gives them a .4275 chance of making the NC game. Or you can adjust their winning probabilities in those three games to whatever you consider more reasonable. I doubt that you will get much over .50.

 

[ndmb]

The first two rounds are walkovers. Then in the third round, ND plays Ok. St., or someone who has beaten Ok St. Maybe they have a five percent chance of losing that game. If they win, then they play Baylor, or someone who has beaten Baylor. Getting ready to duck, I would give Baylor a .25 chance. If they get by OkSt and Baylor (or) then they will play in the fifth game the survivor of Md., WVU, L'ville and the LV. Lets give them a 60% probability of winning that one. Multiplying the (my guestimated) probabilities of winning those three games; .95 x .75 x .6 gives them a .4275 chance of making the NC game. Or you can adjust their winning probabilities in those three games to whatever you consider more reasonable. I doubt that you will get much over .50.

I did get much over 50%, using Masseys, which I think is reasonable. I think that 90% odds over Baylor and 90% odds over Tenn/WVU/Louisville are more accurate than 75% and 60%.

 

[ndmb]

Right on all accounts, NDMB. .9 x .9 = .81 every time.

(What? Notre Dame Marching Band?)...

You got me. Muffet sent me to spy on The Boneyard to look for recruiting violations.

(they're my initials )