2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

[leswp1]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

Welcome everyone to the 2020 NCAA Ice Hockey Tournament - Simulated Selection Show! A special thanks to ticapnews, and the USCHO guys Jim Connelly and Jayson Moy for their bracketologies. Now, let's get started!

The four #1 seeds are North Dakota, Cornell, Minnesota State, and Minnesota Duluth! North Dakota earns the top overall seed, and let's see where they get slotted:

Overall top seed North Dakota gets slotted in the Albany Regional. There, the NCHC Champion Fighting Hawks will take on the Atlantic Hockey Champions, the Sacred Heart Pioneers, who will make their first ever NCAA Tournament Appearance! North Dakota and Sacred Heart will play the first game of the East Regional, on Saturday, March 28, 2020. Joining the Fighting Hawks and the Pioneers in Albany will be the #2 Seed Clarkson Golden Knights, who will square off with the #3 Seeded Massachusetts Minutemen in the evening game on the 28th. The East Regional Final will take place on the evening of Sunday, March 29, 2020, with a spot in the Frozen Four at stake!

The Cornell Big Red earned the second overall #1 seed, and have been placed in the Worcester Regional. Cornell, the ECAC Champions, will take on perennial powerhouse, but the last at-large team in the field, Michigan Wolverines out of the Big Ten. Cornell and Michigan will play the first game of the Northeast Regional, on Friday, March 27, 2020. The winner of that game will take on the winner of the evening game on the 27th, which will feature #2 Seed Boston College Eagles matching up with the #3 Seeded Bemidji State Beavers. The Northeast Regional Final will take place on the afternoon of Saturday, March 28, 2020, with the winner going to Detroit two weeks later.

The third overall #1 seed, and WCHA Regular Season Champion, Minnesota State Mankato will travel to Allentown, Pennsylvania for the Midwest Regional. The Mavericks will play the #4 Seeded Western Michigan Broncos in the afternoon matchup on Saturday, March 28, 2020. Minnesota State and Western Michigan currently have the honor of having played the most NCAA Tournament games without ever earning a victory, so that streak will snap for one of those two programs. Joining the Mavericks and the Broncos in the Midwest Regional are Regional host, and Big Ten Champion, the Penn State Nittany Lions, the #2 seed in the regional. Penn State will square off against the champions out of Hockey East, the #3 Seed Massachusetts Lowell RiverHawks. The winners of these matchups will play in the Midwest Regional final on the afternoon of Sunday, March 29, 2020 to earn a trip to the Frozen Four and a game against the Northeast Regional Champion.

Last, but certainly not least, and two-time defending National Champion Minnesota Duluth Bulldogs will be the #1 seed in the Loveland Regional. The Bulldogs will play the #4 Seeded Ohio State Buckeyes on Friday, March 27, 2020 in a rematch of their 2018 Frozen Four Semifinal, as well as their 2017 West Regional Semifinal, both won by UMD in tight one-goal games. Joining Duluth and Ohio State in Loveland is the Regional host, and #2 seeded Denver Pioneers, who will matchup against the #3 seed, out of Hockey East, the Maine Black Bears. The winners of the semifinals will square off on Saturday, March 28, 2020, for a spot in the Frozen Four against the East Regional Champion.

So there we have it, regionals are set as follows:

Albany (March 28-29, 2020)
1 North Dakota
8 Clarkson
9 Massachusetts
16 Sacred Heart

Worcester (March 27-28, 2020)
2 Cornell
6 Boston College
12 Bemidji State
15 Michigan

Allentown (March 28-29, 2020)
3 Minnesota State
7 Penn State
10 Mass.-Lowell
14 Western Michigan

Loveland (March 27-28, 2020)
4 Minnesota-Duluth
5 Denver
11 Maine
13 Ohio State

Boooo! Who wants to go to Allentown and be anywhere near Ped State except Billy Joel???

[Sol Diablo]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

No wins in the Frozen Faceoff and UMD is still a one seed!

No losses in the tournaments & ASU drops to a five seed.......

[WeAreNDHockey]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

We should revise the dear departed Posters game via a simulator. Some of us haven't skated in a long, long time!

[leswp1]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

We should revise the dear departed Posters game via a simulator. Some of us haven't skated in a long, long time!

Hahahaha! this is awesome! I was never able to get there in real life but virtually..... I can do a mean single axel and scratch spin to confuse the rest of play

[Ralph Baer]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

Congrats to Cornell on retaining the Belt.

                   Current Holder         Next Defense
The Belt           Cornell                Boston College 03/27/2020
The Belt Alt 3x3   Penn State             Minnesota State 03/28/2020
The Belt Alt SO    Massachusetts Lowell   Western Michigan 03/28/2020
AHC Belt           Sacred Heart           next season
B1G Belt           Penn State             04/10/2020 at earliest
ECAC Belt          Cornell                04/12/2020 at earliest
HEA Belt           Massachusetts Lowell   04/10/2020 at earliest
NCHC Belt          North Dakota           04/10/2020 at earliest
WCHA Belt          Bemidji State          04/10/2020 at earliest
Ivy Belt           Cornell                next season
New England Belt   Massachusetts Lowell   04/10/2020 at earliest
Connecticut Belt   Sacred Heart           next season
Massachusetts Belt Boston College         04/10/2020 at earliest
Michigan Belt      Michigan Tech          next season
Minnesota Belt     Minnesota              next season
New York Belt      Cornell                04/12/2020 at earliest
D II Play up Belt  Bemidji State          04/10/2020 at earliest
D II P u Alt SO    American International next season
D III Play up Belt Rensselaer             next season
Play up Belt       Rensselaer             next season

Corrections are always appreciated.

[Ralph Baer]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

I agree, that is how it should be done. Perhaps you can get Jayson Moy to do the analysis.

I was just joking when I wrote this, (Note the smiley.) However it turned out that USCHO actually wrote several articles about this thread including the seeding. Great work FS23!

[Sol Diablo]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

The final pairwise.

1 North Dakota
2 Cornell
3 MSU-Mankato
4 Minnesota-Duluth
5 Denver
6 Boston College
7 Penn State
8 Clarkson
9 Massachusetts
10 Mass.-Lowell
11 Maine
12 Bemidji State
13 Ohio State
14 Western Michigan
15 Michigan
AQ Sacred Heart

In an unexpected twist the bubble burst for Arizona State.

Now to see where the committee sends them...

Maybe someone can help explain this to me.
Referencing the last PWR probability matrix that was issued, utilizing the KRACH Monte Carlo simulator, it shows ASU having a 96% probability of getting into the regionals.

https://www.collegehockeynews.com/ra...lityMatrix.php

It shows UML, Michigan, BSU, Western Michigan, Maine having significantly lower probabilities of getting in. There was only 1 auto bid that got in that was ranked lower than ASU.
I am struggling that these teams got in with these simulated results without winning their conference while having a low probability prior to the couple extra games. A rolled up probability of all 5 teams getting in is less than 4%. Just trying to understand the algorithms. Can anyone share the model or point me to the site that has it? Thanks!

[Fighting Sioux 23]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

Maybe someone can help explain this to me.
Referencing the last PWR probability matrix that was issued, utilizing the KRACH Monte Carlo simulator, it shows ASU having a 96% probability of getting into the regionals.

https://www.collegehockeynews.com/ra...lityMatrix.php

It shows UML, Michigan, BSU, Western Michigan, Maine having significantly lower probabilities of getting in. There was only 1 auto bid that got in that was ranked lower than ASU.
I am struggling that these teams got in with these simulated results without winning their conference while having a low probability prior to the couple extra games. A rolled up probability of all 5 teams getting in is less than 4%. Just trying to understand the algorithms. Can anyone share the model or point me to the site that has it? Thanks!

ticapnews can likely provide a better explanation for you, but collegehockeyranked.com allows you to add the simulated games to those that were actually played to determine the pairwise results. It is my understanding that this is what ticapnews did to come up with the final pairwise. As to your five teams that you had questions about, Lowell and Bemidji State won their conference tournaments in the simulations, so they got auto bids (as well as jumped ahead of ASU). My guess is Michigan (beating OSU on the road, before losing to Penn State on the road), Western Michigan (beating SCSU 2 out of three at home, and then losing to North Dakota and beating Minnesota Duluth at a neutral venue), and Maine (sweeping UCONN at home, beating BC at a neutral venue, losing to Lowell at a neutral venue) had their RPIs increase by at least 0.0035 to jump ahead of Arizona State.

The Monte Carlo simulator does a good job of giving an idea of who will likely make the tournament. However, it should not be seen as an absolute. For example, Team X could be listed as a 100% in based on the 20,000 simulations, but there could be 3% of actual scenarios where they do not get in, and those 3% of scenarios simply did not occur in the 20,000 simulations. A good example of this is to look at how Minnesota missed the tournament in 2018 for a good example. With one day of hockey left in 2018 before the NCAA Tournament, Minnesota had something like a 99% probability to get in based on the Monte Carlo simulator. It took 5 or 6 games going one specific way (including several upsets) for the Gophers to miss out.

[Sol Diablo]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

ticapnews can likely provide a better explanation for you, but collegehockeyranked.com allows you to add the simulated games to those that were actually played to determine the pairwise results. It is my understanding that this is what ticapnews did to come up with the final pairwise. As to your five teams that you had questions about, Lowell and Bemidji State won their conference tournaments in the simulations, so they got auto bids (as well as jumped ahead of ASU). My guess is Michigan (beating OSU on the road, before losing to Penn State on the road), Western Michigan (beating SCSU 2 out of three at home, and then losing to North Dakota and beating Minnesota Duluth at a neutral venue), and Maine (sweeping UCONN at home, beating BC at a neutral venue, losing to Lowell at a neutral venue) had their RPIs increase by at least 0.0035 to jump ahead of Arizona State.

The Monte Carlo simulator does a good job of giving an idea of who will likely make the tournament. However, it should not be seen as an absolute. For example, Team X could be listed as a 100% in based on the 20,000 simulations, but there could be 3% of actual scenarios where they do not get in, and those 3% of scenarios simply did not occur in the 20,000 simulations. A good example of this is to look at how Minnesota missed the tournament in 2018 for a good example. With one day of hockey left in 2018 before the NCAA Tournament, Minnesota had something like a 99% probability to get in based on the Monte Carlo simulator. It took 5 or 6 games going one specific way (including several upsets) for the Gophers to miss out.

Thanks. I recall the Gophers & Sioux did not get in because four teams that were outside the bubble won conference championships that year and jumped ahead. But UML and BSU this time were already ahead of ASU in pairwise, so they didn't jump. SC jumped in. But I still struggle with Western Michigan & Maine jumping ahead. But must be the RPI as you said. I will check out the sites you mentioned. I appreciate you replying and doing the simulation, very interesting.

[Fighting Sioux 23]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

Thanks. I recall the Gophers & Sioux did not get in because four teams that were outside the bubble won conference championships that year and jumped ahead. But UML and BSU this time were already ahead of ASU in pairwise, so they didn't jump. SC jumped in. But I still struggle with Western Michigan & Maine jumping ahead. But must be the RPI as you said. I will check out the sites you mentioned. I appreciate you replying and doing the simulation, very interesting.

Yeah, the Pairwise is a fickle mistress, particularly when your team is idle and can't really help itself out.

Western Michigan got the opportunity to take 2 out of 3 against a strong St. Cloud team (ranked top 25 PWR), and got an upset win over Duluth (top 4). Their only losses were to North Dakota (#1) and St. Cloud. That's going to raise your RPI a bit.

As for Maine, they went 3-1, taking both games against Connecticut (a middle of the road team), and then defeating Boston College (top 6), before losing to Lowell (top 10ish). That will also do wonders to your RPI.

Both of those teams only needed to gain 0.0035 on Arizona State, who was idle. Maine doesn't surprise me in the slightest, as they only lost 1 of 4 games (to a good team at a neutral venue), and upset a top 6 team. Western Michigan also doesn't surprise me. They played a hell of a schedule down the stretch, and had a winning record. The team that does surprise me a bit is Michigan. They went 1-1 and had to overcome a 0.0023 RPI difference. However, both of their games were on the road, and against top 10 competition. Apparently winning 1 of 2 in those circumstances allowed them to gain at least 0.0023 in the RPI...perhaps as a result of the quality win bonus for beating Ohio State on the road.

Again though, I didn't run the pairwise numbers. I left that to the real experts.

[Patman]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

ticapnews can likely provide a better explanation for you, but collegehockeyranked.com allows you to add the simulated games to those that were actually played to determine the pairwise results. It is my understanding that this is what ticapnews did to come up with the final pairwise. As to your five teams that you had questions about, Lowell and Bemidji State won their conference tournaments in the simulations, so they got auto bids (as well as jumped ahead of ASU). My guess is Michigan (beating OSU on the road, before losing to Penn State on the road), Western Michigan (beating SCSU 2 out of three at home, and then losing to North Dakota and beating Minnesota Duluth at a neutral venue), and Maine (sweeping UCONN at home, beating BC at a neutral venue, losing to Lowell at a neutral venue) had their RPIs increase by at least 0.0035 to jump ahead of Arizona State.

The Monte Carlo simulator does a good job of giving an idea of who will likely make the tournament. However, it should not be seen as an absolute. For example, Team X could be listed as a 100% in based on the 20,000 simulations, but there could be 3% of actual scenarios where they do not get in, and those 3% of scenarios simply did not occur in the 20,000 simulations. A good example of this is to look at how Minnesota missed the tournament in 2018 for a good example. With one day of hockey left in 2018 before the NCAA Tournament, Minnesota had something like a 99% probability to get in based on the Monte Carlo simulator. It took 5 or 6 games going one specific way (including several upsets) for the Gophers to miss out.

Except for a case or two the RPI does the real work these days (removal of other elements over the years has made a big difference) and since the standings are bunched together are they are in a sport like hockey you can get a lot of movement from little events. If the teams below you have a great tournament you're kinda fried. On the other hand, if you're in the main body you usually have a game or two lead to start with and unable to lose more than two games in a series format or a single game (excepting those tournaments who still haven't tossed the third place game). As such falling from grace is hard... AZ state needed to have the teams behind them meet their scheduled outs... in this simulation they didn't... and in real life this could have very well happened even if its unlikely.

[Ralph Baer]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

I can't believe that I am eagerly awaiting Thursday's simulations.

[leswp1]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

I can't believe that I am eagerly awaiting Thursday's simulations.

I can. I feel the same way.

I have chartered a virtual personalized jet (riverhawk on the nose of the plane with the plane wings being- wings!) to PA, stocked with chocolate milk, snacks and alcohol for those lucky enough to be able to drink it. It also has magical powers to deflect all infectious agents. I have one other passenger and someone from uhn on the manifest....

[Ralph Baer]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

I can. I feel the same way.

I have chartered a virtual personalized jet (riverhawk on the nose of the plane with the plane wings being- wings!) to PA, stocked with chocolate milk, snacks and alcohol for those lucky enough to be able to drink it. It also has magical powers to deflect all infectious agents. I have one other passenger and someone from uhn on the manifest....

But your team is there. Mine lost in Fake Placid.

[leswp1]

Re: 2020 NCAA Hockey Tournament - What If: A Simulation of the 2019-2020 Postseason

But your team is there. Mine lost in Fake Placid.

You are welcome to the band wagon!