2012 NCAA Tournament: Division I Bracketology

[northhockey]

Re: 2012 NCAA Tournament: Division I Bracketology

Actually, I stand corrected.

A nightmare scenario for the committee is if N. Dakota loses to Minnesota in the WCHA semifinals and then Harvard loses in the ECAC final, and then someone outside the top 8 like BU takes an autobid. Let's also assume Northeastern & BC & Mercyhurst do well enough to avoid falling behind N. Dakota.

I believe under that scenario, Harvard would have a slightly better RPI and record vs. TUC than N. Dakota, while North Dakota would have a better record against common opponents Minnesota & BU (against whom N. Dakota is .500 and Harvard is winless). That would be a hugely controversial decision, whichever way it goes.

Yeah, that would be. Well hopefully the Sioux just win the next 2 games and get the autobid

[OnMAA]

Re: 2012 NCAA Tournament: Division I Bracketology

Yeah, that would be. Well hopefully the Sioux just win the next 2 games and get the autobid

As I said before. If you are in the post season, you have your destiny in your own hands. Just keep on winning. If you win your last game you are Golden.

If I got my math correct, Win your next five games, and you get to hoist the trophy.

[dave1381]

Re: 2012 NCAA Tournament: Division I Bracketology

My guess is the committee still takes UND over Harvard in that scenario. It's similar to UND vs. Dartmouth 2011 in some ways, but in that case, the committee simply didn't have much room for discretion based on the criteria, since UND didn't actually beat Dartmouth in any of the conventional selection criteria (though surely UND's 5-3 record vs. Minnesota & UMD was more impressive than any over Dartmouth's accomplishments).

In this Harvard vs. UND scenario, UND does have a transparent edge here: UND got the job done against Minnesota & BU, going 3-3, while Harvard did not, while Harvard would have the slight edge in the other categories. Taking UND over Harvard would be similar to taking Dartmouth over Clarkson in 2008 (where the Dartmouth 1-0-1 H2H over Clarkson ruled over all else). The case for UND would be stronger than the case for 2008 Dartmouth in my opinion.

But yes, surely both teams don't want to leave it any one else's hands...

[ARM]

Re: 2012 NCAA Tournament: Division I Bracketology

BTW, I have no doubt that BU's present comparison win over N. Dakota is an aberration. The committee would surely give that to N. Dakota.

N. Dakota has a .57 to .54 edge over RPI (effectively three games). The common opponents are a wash, both teams swept Vermont and struggled against Wisconsin (N. Dakota getting a draw). Note women didn't officially switch their common opponent criteria like I guess the men did, but the committee has always had the discretion to do so. BU has a tiny, tiny advantage in the record vs. TUC.

So this is a clear N. Dakota comparison win as far as I can tell. N. Dakota is 7th, Harvard is 8th, SLU is 9th, BU is 10th.

I think that the BU win over UMD is an aberration as well. UMD has more games (and thus losses) from UW, while BU has more with UConn. BU was plus one versus BC, but UMD was plus won versus UND. So I'd say that COP is a wash. UMD has a better RPI, so right now, they should win that comparison.

Still a joke as to how UND can be all the way down at 7, but its fine they will send us to minneapolis either way.

I think that is true. One question is what would the committee do with UND should you fall to 8th but still make the field? Can they send you to Madison if Wisconsin is the top seed, or would that mean that they'd have to fly you there? Even a flight to Madison wouldn't be out of the question, because that would still leave the committee with (presumably) four teams from the ECAC and HEA combined, plus Mercyhurst, so they could get by perhaps with but one more flight to send someone to Minneapolis (unless Mercyhurst needs a flight as well.) And of course, UMD is still alive as well.

[pakidnyc]

If it stays the way the current 8 are in the pariwise ranking this is what most likely would happen....

Harvard to Wisco
NEU to Cornell
UND to UM
MC vs BC

Now if SLU beats Harvard next week, or if BU wins HE, then things can change in a hurry.

The ECAC could be won by any of the 4 teams (assuming Clarkson takes care of business today) so if Clarkson or SLU win it, Cornell (and Harvard?) would surely still get a bid, so then who above is out?

[pakidnyc]

I think that the BU win over UMD is an aberration as well. UMD has more games (and thus losses) from UW, while BU has more with UConn. BU was plus one versus BC, but UMD was plus won versus UND. So I'd say that COP is a wash. UMD has a better RPI, so right now, they should win that comparison.

Okay, but let's call spade a spade. UND has 4 games vs St Cloud St. and 1 vs Lindenwood. You can look at every team and say they have weak opponents.

[ARM]

Re: 2012 NCAA Tournament: Division I Bracketology

Okay, but let's call spade a spade. UND has 4 games vs St Cloud St. and 1 vs Lindenwood. You can look at every team and say they have weak opponents.

That's not what is being done. Dave and I are discussing the Common OPponents category of the PWR. BU takes the comparisons with both UMD and UND as shown on USCHO, based in part by doing better against common opponents on a percentage basis. We are just commenting that both UMD and UND look worse compared to BU because of more games played -- and lost -- versus Wisconsin. The Terriers didn't have success against UW either; they just played the Badgers fewer times. Conversely, BU played Vermont and Connecticut more than UND and UMD respectively. Neither SCSU nor Lindenwood is an opponent of BU, so they are not part of this criterion. So you can call a spade a spade, but in this hand, clubs are trump.

[dave1381]

Re: 2012 NCAA Tournament: Division I Bracketology

Also, this way of doing the common opponents comparison isn't just some idealistical academic exercise that's presently irrelevant for the selections (like, say, looking at KRACH), but it's actually exactly what the men's tournament formally instituted this year, and it's how the men's PWR work. My guess is the women have already been doing the same, even though I don't believe they formalized it, because they already had more discretion than the men's committee.

USCHO is still going with the old common opponent calculation for the women's PWR at the moment, but it may be true that the new method is a better model of what the committee would decide.

[OnMAA]

Re: 2012 NCAA Tournament: Division I Bracketology

The ECAC could be won by any of the 4 teams (assuming Clarkson takes care of business today) so if Clarkson or SLU win it, Cornell (and Harvard?) would surely still get a bid, so then who above is out?

IF SLU wins ECAC, or even if they just beat Harvard in the Semi, it could spell the end for Harvard, as SLU has a good record vs other teams under consideration compared to the Crimson. From what I can tell, Clarkson pretty much has to win the ECAC to make it in. Either way if a lower seed team wins ECAC, it will probably mean the end off Harvard's season.

[ARM]

Re: 2012 NCAA Tournament: Division I Bracketology

My guess is the women have already been doing the same, even though I don't believe they formalized it.

I believe that the committee must have done something such as that with COP last year to place Minnesota above Mercyhurst, because the Lakers took the USCHO version of COP due to the advantage of 4-0 over 1-0 versus Wayne State, coupled with MC being 0-1 playing Wisconsin while Minnesota was 1-3-1. So in those games alone, MC was 4-1 while UM was 2-3-1, but it wasn't apples to apples. Because Mercyhurst also played SCSU and BSU and went 4-1 (compared to Minnesota's 7-1), the total number came out to 8-2 versus 9-4-1.

[dave1381]

Re: 2012 NCAA Tournament: Division I Bracketology

If the last at-large bid came down to Harvard-SLU (meaning neither wins ECACs), I suspect Harvard would still get the bid, even if SLU draws and beats Harvard. Harvard would still have the H2H edge, the common opponents edge, and a slight RPI edge. SLU would have the record vs. teams under consideration edge, but I don't think it would be large enough to nullify Harvard's advantages -- you also have to take into account that SLU would suffer from an ECAC final loss.

But yes, Harvard is the last team in at the moment (i.e. who would be bumped if BU/Clarkson/SLU/UMD win an autobid). But it's possible Northeastern could be the team that gets bumped if it loses in the Hockey East semifinals, and Harvard reaches the ECAC final. There's also a case for Harvard over North Dakota if the Sioux lose to Minnesota, and Harvard loses in the ECAC final, as I discussed earlier this thread -- though my best guess is the committee goes with N. Dakota there.

[dave1381]

Re: 2012 NCAA Tournament: Division I Bracketology

I believe that the committee must have done something such as that with COP last year to place Minnesota above Mercyhurst.

Yeah, it's hard to tell for sure because that was a three-team comparison with BC. I agree they used some discretion there. I think they more or less ended up ranking the three teams based on record vs. TUC since that's where the biggest differences were.

[joehockey]

Re: 2012 NCAA Tournament: Division I Bracketology

What happens if PC should win the HE autobid in playoff?

[dave1381]

Re: 2012 NCAA Tournament: Division I Bracketology

PC surely gets shipped to Wisconsin. Let's also assume BC-Mercyhurst stand pat.

If UND & Northeastern make it, it's probably UND-Minnesota, Northeastern-Cornell.
If UND & an ECAC team make it, it's probably UND-Minnesota, ECAC team vs. Cornell.
If an ECAC team & Northeastern make it, it might be Northeastern-Cornell, ECAC team at Minnesota. But maybe Northeastern is far enough ranked above the ECAC team they don't sent Northeastern to Cornell though.

[ARM]

Re: 2012 NCAA Tournament: Division I Bracketology

PC surely gets shipped to Wisconsin.

Isn't the Cornell vs UW comparison close enough that the Big Red can still reverse it if Wisconsin loses, particularly to UMD?