Bracketology for 2020 NCAAs

[adamw]

Well if Jim read my article (I'm sure he was one of the ones rolling his eyes) - he'd know that those brackets are neither "projections" nor "predictions" at all. It bothers the ever loving hell out of me when they're called that.

[Jeff Hopkins '82]

Well if Jim read my article (I'm sure he was one of the ones rolling his eyes) - he'd know that those brackets are neither "projections" nor "predictions" at all. It bothers the ever loving hell out of me when they're called that.

I read your article, Adam.  I didn't roll my eyes, but I did smirk a bit at the disclaimer. ::whistle::

But based on the article, I get these brackets:

Worcester, Mass. Regional
1. North Dakota (1)
2. Boston College (6)
3. Arizona State (10)
4. AIC (16)

Loveland, Colo. Regional
1. Minnesota Duluth (4)
2. Denver (5)
3. UMass Lowell (11) or Northeastern (12)
4. Ohio State (13)

Allentown, Pa. Regional
1. Minnesota State (2)
2. Penn State (
3. UMass Lowell (11) or Northeastern (12)
4. Providence (15)

Albany, N.Y. Regional
1. Cornell (3)
2. UMass (7)
3. Clarkson (9).
4. Maine (14)

[Trotsky]

Albany, N.Y. Regional
1. Cornell (3)
2. UMass (7)
3. Clarkson (9).
4. Maine (14)

That regional is fascinating.

[KenP]

Interesting... At this point it is down to a 59-team race.  Even though their PWR is higher than both Princeton and St. Lawrence... Vermont has 0.0% chance to make the NC$$ tournament.

[Jeff Hopkins '82]

Albany, N.Y. Regional
1. Cornell (3)
2. UMass (7)
3. Clarkson (9).
4. Maine (14)

That regional is fascinating.

It is, isn't it?  I did my own bracketology a few days ago, and put Az. State in Albany and Clarkson in Worcester, but that was simply because I didn't want to see Cornell and Clarkson match up a potential 4th time in the regionals.

[Trotsky]

Interesting... At this point it is down to a 59-team race.  Even though their PWR is higher than both Princeton and St. Lawrence... Vermont has 0.0% chance to make the NC$$ tournament.

Merrimack's really close, too.

According to this, 20 of those 59 can only make it now by winning their auto bid, including half the ECAC.

[Trotsky]

Albany, N.Y. Regional
1. Cornell (3)
2. UMass (7)
3. Clarkson (9).
4. Maine (14)

That regional is fascinating.

It is, isn't it?  I did my own bracketology a few days ago, and put Az. State in Albany and Clarkson in Worcester, but that was simply because I didn't want to see Cornell and Clarkson match up a potential 4th time in the regionals.

I was thinking it would be wild to see us play Clarkson 3 times in 30 days: Feb 29 in Ithaca for the ECAC RS title, March 21 in Lake Placid for the ECAC title, and March 29 in Albany for advance to the Frozen Four.

[Jeff Hopkins '82]

Albany, N.Y. Regional
1. Cornell (3)
2. UMass (7)
3. Clarkson (9).
4. Maine (14)

That regional is fascinating.

It is, isn't it?  I did my own bracketology a few days ago, and put Az. State in Albany and Clarkson in Worcester, but that was simply because I didn't want to see Cornell and Clarkson match up a potential 4th time in the regionals.

I was thinking it would be wild to see us play Clarkson 3 times in 30 days: Feb 29 in Ithaca for the ECAC RS title, March 21 in Lake Placid for the ECAC title, and March 29 in Albany for advance to the Frozen Four.

I'd rather see us meet up the fourth time in Detroit, not Albany.  And with Clarkson in Worcester with NoDak as #1, and us as #3, that meeting would be the finals.

[BearLover]

The Pairwise Probability Matrix, which I believe "accounts for" uncertainty the same way playoffstatus.com does, gives Cornell (currently in 3rd place in the Pairwise/KRACH) a 67% chance of finishing with exactly the 3-seed. It gives NoDak (currently first) an 87% likelihood of finishing with the 1-seed. It gives Minn State (currently second) 75% odds of finishing with the 2-seed. Each of these teams still has ~12 games left before seeding. Sounds legit?

[adamw]

The Pairwise Probability Matrix, which I believe "accounts for" uncertainty the same way playoffstatus.com does, gives Cornell (currently in 3rd place in the Pairwise/KRACH) a 67% chance of finishing with exactly the 3-seed. It gives NoDak (currently first) an 87% likelihood of finishing with the 1-seed. It gives Minn State (currently second) 75% odds of finishing with the 2-seed. Each of these teams still has ~12 games left before seeding. Sounds legit?

We're not accounting for any "uncertainty" at all - yet.

One thing that does it make it legit - yeah - is the gap in RPI between all of these teams. There is a larger gap between 1 and 2 - 2 to 3 - and, in particular 3 to 4 - than there is most of groups of 5 teams. In particular, there's a .0231 gap from 3 to 4. By comparison - No. 4 and 10 are closer than that gap.

[BearLover]

The Pairwise Probability Matrix, which I believe "accounts for" uncertainty the same way playoffstatus.com does, gives Cornell (currently in 3rd place in the Pairwise/KRACH) a 67% chance of finishing with exactly the 3-seed. It gives NoDak (currently first) an 87% likelihood of finishing with the 1-seed. It gives Minn State (currently second) 75% odds of finishing with the 2-seed. Each of these teams still has ~12 games left before seeding. Sounds legit?

We're not accounting for any "uncertainty" at all - yet.

One thing that does it make it legit - yeah - is the gap in RPI between all of these teams. There is a larger gap between 1 and 2 - 2 to 3 - and, in particular 3 to 4 - than there is most of groups of 5 teams. In particular, there's a .0231 gap from 3 to 4. By comparison - No. 4 and 10 are closer than that gap.

I may be using the wrong terminology, but what do you mean you are not accounting for uncertainty? And would you put your money where your mouth is and assert that if this season were to be replicated from this point forward one million times, in 2/3 of those scenarios Cornell would finish with exactly the 3-seed? That Minn St would finish 3/4 of the time with exactly the 2-seed? That NoDak would finish 7/8 of the time with the 1-seed? Do you also believe Cornell is almost three times as likely to win the ECAC championship as Clarkson, and more likely to win it than the rest of the ECAC combined? I believe these numbers are quite a bit off, and a few years ago someone on here ran a regression [correct terminology?] showing that the tails of these models are off--the chances of a top team beating a bottom team are overstated by the model, which over the course of a lengthy stretch (in this case, ~12 remaining games) leads to significantly underrating volatility.

[Tom Lento]

The Pairwise Probability Matrix, which I believe "accounts for" uncertainty the same way playoffstatus.com does, gives Cornell (currently in 3rd place in the Pairwise/KRACH) a 67% chance of finishing with exactly the 3-seed. It gives NoDak (currently first) an 87% likelihood of finishing with the 1-seed. It gives Minn State (currently second) 75% odds of finishing with the 2-seed. Each of these teams still has ~12 games left before seeding. Sounds legit?

I've long had the same complaints about these models that you routinely bring up, and I've found it helps me to think of this less as a prediction of real probabilities and more as a mechanism of demonstrating the likely finishing positions assuming the rest of the game results follow from the existing game results. Honestly, there's no real reason to suspect that the high percentages indicate anything wrong with the model, at least not based on what it's effectively doing. At the end of the day, these monte carlo simulations based on KRACH (which is itself based solely on record) are incredibly information-poor and we have to account for that in our interpretation.

I ended up subscribing to The Athletic last year and I found their NHL playoff possibility models very interesting because they have access to a lot more team and player-level data. Of course, those richer models didn't come out looking very good last season what with Tampa Bay getting bounced in the first round and St. Louis winning it all. Even so, if I end up becoming an unemployed hobbyist for a while I'd love to try to get something similar for NCAA hockey, although the much smaller sample of games will make it super noisy.

[BearLover]

I found the old thread, which includes jfeath's incredibly helpful analysis:

[Ugh sorry, having problems posting the link on my phone, but it's post #12 or so on the thread titled "2018 ECAC Permutations"]

[upprdeck]

all i know its a lot more fun being 6 games left wondering how Cornell can screw it up than it is 6 games left hoping other teams screw it up.

[Trotsky]

Amen, brother.