2012 NCAA Tournament: Bracketology

[Priceless]

Re: 2012 NCAA Tournament: Bracketology

That doesn't account for ties in consolation games.

I wondered about that...

[JimDahl]

Re: 2012 NCAA Tournament: Bracketology

Would someone with access to the Pairwise database be able to post something like:

19 games remaining
1,238,432 possible outcomes (I did not do the actual math...)

Outcomes where team finishes 13th or above (or whatever cutoff we think is necessary based on likely auto-qualifiers)

BC: 1,238,432
Michigan: 1,238,432
UMD: 1,238,432
Miami: 1,100,234
etc.

The problem with choosing an arbitrary cutoff is there's no way to know what ranking is safe until the conference tournaments end. So, I posted the possible ranges in this thread on the previous page, and the likelihoods as percentages of the 1,179,648 remaining possible outcomes:

PWR possibilities

I'm still doing some error-checking, so let me know if anything looks amiss.

[JF_Gophers]

Re: 2012 NCAA Tournament: Bracketology

W T F, Harvard has 11 ties (12 wins)!?

[Numbers]

Re: 2012 NCAA Tournament: Bracketology

Concerning having your fate tied to others: It is true that happens all the time. However, it happens because the other result to which your fate is tied also has something to do with the fate of the teams involved in that result.

This is not the case with the PWR. (Disclaimer - any who watches my posts here knows I am addicted to numbers. PWR offers lots of ways to experiment with numbers, so I like it. Really I do. And, certainly I think a 'simple math (copyright, ScoobyDoo)' system is better than 5 guys thinking, "Well, Lowell just looked better than Minnesota this year....")

With PWR, lots of your fate is tied to the TUC cliff. Who is TUC and who is not is not something which you directly affect. Also, the key thing is "Record against TUCs." That means that a game against UNH is the same as a game against BC. So, my problem with the system here is:

#1 - TUC cliff is always arbitrary. And, I am not sure there is a way to fix that. And,
#2 - If NCAA wants a TUC component to the PWR, it should find a way to evaluate it at least slightly based on SOS within the TUC group.

But, in short, hypothetically, not talking about Lowell here, but hypothetically, the last team in the tourney could come down to :
Merrimack or Northern Michigan.

Hhhhmmmmm, let's seeee..... Well, In October, NMU played RIT and beat them. Today, RIT beat Air Force in the AHA to just squeak into the TUC group. So, NMU's win against RIT in October now counts huge for them, and they get in over Merrimack. Or, AFA beat RIT, so RIT is not TUC, so there is no extra credit to NMU for beating them, so Merrimack gets in.

The thing is, neither Merrimack or NMU (hypothetically, again) have anything to do the the RIT/AFA game. And, RIT and AFA are playing, and their game decides who gets in, and they don't give a rip about that.

It just seems awkward to me...

But, like I say, it sure makes this week fun...

[JF_Gophers]

Re: 2012 NCAA Tournament: Bracketology

Concerning having your fate tied to others: It is true that happens all the time. However, it happens because the other result to which your fate is tied also has something to do with the fate of the teams involved in that result.

This is not the case with the PWR. (Disclaimer - any who watches my posts here knows I am addicted to numbers. PWR offers lots of ways to experiment with numbers, so I like it. Really I do. And, certainly I think a 'simple math (copyright, ScoobyDoo)' system is better than 5 guys thinking, "Well, Lowell just looked better than Minnesota this year....")

With PWR, lots of your fate is tied to the TUC cliff. Who is TUC and who is not is not something which you directly affect. Also, the key thing is "Record against TUCs." That means that a game against UNH is the same as a game against BC. So, my problem with the system here is:

#1 - TUC cliff is always arbitrary. And, I am not sure there is a way to fix that. And,
#2 - If NCAA wants a TUC component to the PWR, it should find a way to evaluate it at least slightly based on SOS within the TUC group.

But, in short, hypothetically, not talking about Lowell here, but hypothetically, the last team in the tourney could come down to :
Merrimack or Northern Michigan.

Hhhhmmmmm, let's seeee..... Well, In October, NMU played RIT and beat them. Today, RIT beat Air Force in the AHA to just squeak into the TUC group. So, NMU's win against RIT in October now counts huge for them, and they get in over Merrimack. Or, AFA beat RIT, so RIT is not TUC, so there is no extra credit to NMU for beating them, so Merrimack gets in.

The thing is, neither Merrimack or NMU (hypothetically, again) have anything to do the the RIT/AFA game. And, RIT and AFA are playing, and their game decides who gets in, and they don't give a rip about that.

It just seems awkward to me...

But, like I say, it sure makes this week fun...

But again, it took a season full of other results in which those teams did directly impact to make that RIT/AFA game important to those teams now.

[Patman]

Re: 2012 NCAA Tournament: Bracketology

But again, it took a season full of other results in which those teams did directly impact to make that RIT/AFA game important to those teams now.

Honestly, it just shows how difficult it is to seperate teams with "winning percentages" between 60 and 65%. That's where most of these teams sit.

[NMU8405]

Re: 2012 NCAA Tournament: Bracketology

But, in short, hypothetically, not talking about Lowell here, but hypothetically, the last team in the tourney could come down to :
Merrimack or Northern Michigan.

Hhhhmmmmm, let's seeee..... Well, In October, NMU played RIT and beat them. Today, RIT beat Air Force in the AHA to just squeak into the TUC group. So, NMU's win against RIT in October now counts huge for them, and they get in over Merrimack. Or, AFA beat RIT, so RIT is not TUC, so there is no extra credit to NMU for beating them, so Merrimack gets in.

The thing is, neither Merrimack or NMU (hypothetically, again) have anything to do the the RIT/AFA game. And, RIT and AFA are playing, and their game decides who gets in, and they don't give a rip about that.

When did NMU play RIT? I have no idea what you are talking about here but NMU has wins over Wisconsin, St Cloud, and Michigan Tech. Also 3 wins over Miami which could help boost them if Miami wins the CCHA. Frankly it comes down to either Western or Cornell losing two games this weekend. Should either split this weekend, then it might come down to Ws over the teams mentioned and how deep they go, but falling short of winning their conf.

[gscott13]

Re: 2012 NCAA Tournament: Bracketology

W T F, Harvard has 11 ties (12 wins)!?

a buddy of mine's father went to harvard and i was telling him this week about their record and how when you dig deeper, 22 out of 32 games ended tied or were decided by one goal.... and 27 of 32 were 2 goal games (one of which included an empty netter)

[aparch]

Re: 2012 NCAA Tournament: Bracketology

Yeah, not sure where he's getting the NMU/RIT connection. The closest those two teams have ever been was the Catamount Cup where RIT played Ferris and LSSU, and later when RIT traveled to Wisconsin.

In October, RIT played a fairly insulated schedule against: Niagara, St. Lawrence, Canisius, Union, and Mercyhurst.


Although, RIT performing well *might* help thanks to comparisons for NMU vs FSU/LSSU/Wisco vs RIT.

[dggoddard]

Re: 2012 NCAA Tournament: Bracketology

W T F, Harvard has 11 ties (12 wins)!?

Wisconsin would love to have 11 ties.

"The Badgers are now 0-11-18 in 29 games that went overtime in WCHA action since 2007-08."

[NMU8405]

Re: 2012 NCAA Tournament: Bracketology

Yeah, not sure where he's getting the NMU/RIT connection. The closest those two teams have ever been was the Catamount Cup where RIT played Ferris and LSSU, and later when RIT traveled to Wisconsin.

In October, RIT played a fairly insulated schedule against: Niagara, St. Lawrence, Canisius, Union, and Mercyhurst.


Although, RIT performing well *might* help thanks to comparisons for NMU vs FSU/LSSU/Wisco vs RIT.

NMU currently has a 44.1% chance of making the NCAAs. NMU fans should be hoping for a Union win in the ECAC, Duluth in the WCHA (though a deep run by Tech and SCSU would help) and Miami/Michigan in the CCHA

[MichPieGuy]

Re: 2012 NCAA Tournament: Bracketology

Wisconsin would love to have 11 ties.

"The Badgers are now 0-11-18 in 29 games that went overtime in WCHA action since 2007-08."

That's almost an unfathomable record! Very hard to reason that, because overall I've been impressed with the job Eaves has done as coach at Sconnie. (This weekend's 3-game marathon a good case in point...I "watched" the games on the game tracker in lieu of available TV.) Am hoping the Pio's do well this weekend and that they somehow land in Green Bay, in which case I'll make the swing around the north half of the lake to take in the Regional.

[Wol4ine]

Re: 2012 NCAA Tournament: Bracketology

The 2 teams I see having the biggest influence on NMU are WMU and Cornell. If WMU wins at least one more, you'll need Cornell to lose twice more.

[Fighting Sioux 23]

Re: 2012 NCAA Tournament: Bracketology

Looking at Jim's numbers, the following 11 teams can get a #1 seed if they win their conference tournament: (% if they win conference tourney/overall %)

Boston College (100%/100%)
Michigan (100%/98.7%)
Duluth (100%/41.8%)
Miami (97.4%/38.5%)
Boston University (93.3%/29.9%)
North Dakota (85.2%/10.6%)
Minnesota (76.1%/20.9%)
Union (73.1%/19.2%)
Denver (31.9%/4%)
Maine (30%/7.5%)
Cornell (1.2%/0.3%)

Ferris State still has a 28.7% chance of landing a #1 seed despite not being able to win their tournament.

[Numbers]

Re: 2012 NCAA Tournament: Bracketology

Hey,

The deal with NMU and RIT was a totally made up scenario to illustrate my point. Which mostly is that the TUC business in the PWR is a rather arbitrary thing.

To try again to make my point clear, let's just look at UNH, which right now is right on the very edge of having an RPI of .5000. Now, the deal is, if they end up with a .5000, all the results against them count in the TUC category, but if they end up at .4999, their results don't count. And, those results can make the difference between some team being in, or out. I fail to see the logic in that.

But, that's just the way I think about it.