What's new
USCHO Fan Forum

This is a sample guest message. Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!

  • The USCHO Fan Forum has migrated to a new plaform, xenForo. Most of the function of the forum should work in familiar ways. Please note that you can switch between light and dark modes by clicking on the gear icon in the upper right of the main menu bar. We are hoping that this new platform will prove to be faster and more reliable. Please feel free to explore its features.

2012 NCAA Tournament: Division I Bracketology

I think that it is possible even w/o those big upsets. It's probably also possible that either ECAC or Hockey East only gets one with just the right set of events.

I think the only possible scenario for the Hockey East and ECAC to only get 1 is not realistic. Harvard has good standing and Dartmouth could go on a run. And if either of them or possibly Clarkson would pull out the ECAC championship, Cornell would get in. Same scenario for BC if NU or BU wins HE.

I guess the right set of events could be Cornell and BC both winning auto bid, but even in that scenario I think Harvard will end up getting in ahead of Mercyhrust.
 
Re: 2012 NCAA Tournament: Division I Bracketology

I think the only possible scenario for the Hockey East and ECAC to only get 1 is not realistic.
Notice I said "or", not "and". The second team in each league could finish outside of the top six in at large rankings. So Harvard, Dartmouth, et al could conceivably be out if Cornell, BU, and UMD took auto bids. Similarly, say Northeastern has a tough time with Maine and BC takes the HEA tourney while someone like Dartmouth gets the ECAC. Given these rankings aren't very static, it wouldn't take a lot to change the picture as we know it.
 
Re: 2012 NCAA Tournament: Division I Bracketology

:rolleyes:So, last season there was a question as to Dartmouth deserving to be in the tournament. Many saw UND as the superior club. And they may have been. But I think MOST eventually felt Dartmouth earned their selection.

Yes, I agree Dartmouth earned the selection in the sense that there was a specific criteria in place and Dartmouth was better in the criteria.

UND failed down the stretch/in the end, especially, to continue to 'surge', win games, perform on the ice. Maintain their electability if you will- when they could have. Thus sealing the deal (and their own fate). Alas and alack, they didn't (surge/win/be deserving/continue their electability, etc.) and the air went out of their balloon and Dartmouth got the bid.

That's not really why Dartmouth made the tournament over North Dakota. The ability to win games later rather than earlier in the season is no longer a selection criterion. Dartmouth earned the spot over North Dakota because they did better over the balance of the season in the selection criterion. It's simply not the case that a choice of Dartmouth here was better than North Dakota because only the former choice represented "deciding on the ice."

In the end we will agree the deserving teams were selected for tournament play.

Well that depends on how we define "deserving." If deserving means who did best with the rules in place, then of course Dartmouth was deserving -- it's a tautology. If deserving means the criteria should be whatever we feel like, then we can rationalize whichever choice we want. Neither of these approaches are intellectually insightful.

I approach the question of "deserving" in the following way:
(1) What are the first principles that should be used to select to NCAA tournament teams?
(2) Do the particular NCAA criteria chosen line up with these first principles?
(3) Does the NCAA choose the best measures to implement the criteria?
(4) Is the NCAA committee objectively following the measures that have been chosen?
Whether a team is deserving is based on whether (2),(3),(4) have answers of "yes" leading to an effective implementation of the first principles.

Here is how I approach these questions:
(1) What are the first principles that should be used to select to NCAA tournament teams? I think tournament choice should balance a number of factors
(a) who has had the most success during the course of the season
(b) which teams on the margin are most likely to be successful during the tournament -- this means looking at who did better against the best teams
(c) encourage some degree of broad representation across leagues to both encourage the growth of the sport and ensure that every team has an opportunity to compete for the championship

(2) Do the particular NCAA criteria historically chosen line up with these first principles? I think they actually match up pretty well with my first principles.
(a) We need some broad measure that takes into account all games of the season (e.g. the RPI)
(b) We need some measure that looks at teams ability to perform against the best (e.g. the record vs. the RPI top 12 or TUC)
(c) We have automatic bids for each conference champion (except the CHA, but we're getting there...)

(3) Does the NCAA choose the best measures to implement the criteria?
NO!!!! Here is where I have an objection
(a) The RPI is subject to all kinds of abnormalities in terms of comparing teams' overall balance of play. I think common opponents and head-to-head are good tiebreakers and reality checks, but they have too much importance in the current criteria. In UND vs. Dartmouth, UND did not got enough credit for beating Minnesota 3 of 4 times, or its success against Bemidji, a team that was underrated in the RPI.
(b) The record vs. RPI top 12 isn't adjusted for strength-of-schedule and the RPI selects the wrong top 12 teams. UND played a much stronger set of top 12 teams than Dartmouth did.
(c) I think autobids are fine, and they to some degree recognize that there are always errors in the process and teams do always have a clear opportunity to get in the tournament regardless of what the criteria are.

(4) Is the NCAA committee objectively following the measures that have been chosen?
The answer here is Yes. The immediate reaction of a lot of people was that there was an anti-WCHA bias from the committee. That's missing the point. The problem is what measures are used to implement the selection criteria, not human bias.

So to summarize, the philosophy behind the criteria is sound, the execution of the committee is sound -- it's the pure technocratic problem of implementing the NCAA's selection philosophy that is problematic, and where the NCAA should seek to improve. The failure in the technocratic problem is why I thought North Dakota was more deserving than Dartmouth in 2011, where the standard for "deserving" is whether the particular measures chosen for the criteria actually implement the NCAA's selection principles.
 
Last edited:
Re: 2012 NCAA Tournament: Division I Bracketology

Notice I said "or", not "and". The second team in each league could finish outside of the top six in at large rankings. So Harvard, Dartmouth, et al could conceivably be out if Cornell, BU, and UMD took auto bids. Similarly, say Northeastern has a tough time with Maine and BC takes the HEA tourney while someone like Dartmouth gets the ECAC. Given these rankings aren't very static, it wouldn't take a lot to change the picture as we know it.

I would not be in the least surprised if the Hockey East Tourney results created a scenario entirely different than what has been discussed. Expectations are that Northeastern ,BC, or BU will end up in the finals, but realistically, Maine or Providence could very well make it, and if one or the other does, game on.
 
Re: 2012 NCAA Tournament: Division I Bracketology

I just ran some estimates of the probability of conference tournament upsets.

ECAC:
Chance someone other than Cornell wins: 37%
Chance someone other than Cornell/Harvard wins: 22%
Chance someone other than Cornell/Harvard/Dartmouth wins: 12%

Hockey East:
Chance someone other than BC wins: 56%
Chance someone other than BC/Northeastern wins: 25%
Chance someone other than BC/BU/Northeastern wins: 9%

WCHA:
Chance someone other than Wisc wins: 39%
Chance someone other than Wisc/Minn wins: 8%
Chance someone other than Wisc/Minn/UND wins: 3.3%

Of course these models are assuming that team success is consistent with what has been observed in KRACH throughout the season to date, and it's possible a lower-tier team could step up more than a higher-ranked team. That said, we haven't observed a whole lot of upsets since 2005 to date.

There's a 43% chance an autobid team outside the current top 8 wins a conference tourney, and there's about a 7% chance there are two autobids outside the top 8.
 
Re: 2012 NCAA Tournament: Division I Bracketology

There's a 43% chance an autobid team outside the current top 8 wins a conference tourney, and there's about a 7% chance there are two autobids outside the top 8.
I wonder if that 43% is higher than normal, because I don't remember it happening much. Of course, when it does happen, the team often moves into the Top 8 in the process, so the event becomes much less memorable.

Thanks for the numbers.
 
Re: 2012 NCAA Tournament: Division I Bracketology

Right, that 43% includes potentially Dartmouth getting an autobid to knock out Harvard or BU getting an autobid to knock out Northeastern, and neither case is really earthshattering.

One that comes to mind is Harvard winning the ECAC in 2006 knocked out BC (who'd beaten Harvard in the Beanpot), but these were close teams to begin with.

We haven't really ever had a situation where a solid team got unexpectedly got bounced by a surprising and inferior autobid winner.
 
Re: 2012 NCAA Tournament: Division I Bracketology

One that comes to mind is Harvard winning the ECAC in 2006 knocked out BC (who'd beaten Harvard in the Beanpot), but these were close teams to begin with.
Harvard beat Brown to earn the auto bid, and if I'm remembering right, the Bears were a little farther down and a win by them would have raised more eyebrows. RPI in 2009 is another one I remember where a team that was on the road the whole way came close.
 
Re: 2012 NCAA Tournament: Division I Bracketology

Let the musical chairs continue.

Here is a first in a while. No movement at all after last nights games. Everyone in the pairwise that played won their game last night.
 
Re: 2012 NCAA Tournament: Division I Bracketology

Big news is Northeastern is at 10th in the PWR. No real sense behind the movement. If BU gets back in the top 12.

The Top 8 mostly follows the RPI, except Northeastern which falls all the way to 10th. But if BU passes Dartmouth for 12th place, I believe Northeastern gets back up to 7th place, since they gain 3-1 in record vs. RPI top 12 and lose a loss to Dartmouth vs. the RPI top 12. What a wonderful system.

Clarkson & SLU solidified their position in the top 12 with wins over a reasonably tough pairing of Princeton & Quinnipiac. The big winners from this are Mercyhurst & Harvard, who are 2-0 and 4-0 against these schools.
 
Last edited:
Re: 2012 NCAA Tournament: Division I Bracketology

Big news is Northeastern is at 10th in the PWR. No real sense behind the movement. If BU gets back in the top 12.

The Top 8 mostly follows the RPI, except Northeastern which falls all the way to 10th. But if BU passes Dartmouth for 12th place, I believe Northeastern gets back up to 7th place, since they gain 3-1 in record vs. RPI top 12 and lose a loss to Dartmouth vs. the RPI top 12. What a wonderful system.

Clarkson & SLU solidified their position in the top 12 with wins over a reasonably tough pairing of Princeton & Quinnipiac. The big winners from this are Mercyhurst & Harvard, who are 2-0 and 4-0 against these schools.

The UND score is now counted in, and the Clarkson win was counted in late, and this moved NEU back up into a tie in 8th, and moved UMD back down. UND dropped a spot due to their tie, and MC was the benefactor.

Clarkson and SLU are playing Cornell next weekend. The other ECAC teams in the top 12, Harvard and Dartmouth, are playing a light schedule to close out the regular season.
 
Re: 2012 NCAA Tournament: Division I Bracketology

SLU & Clarkson stand to lose about .0050 in RPI on Dartmouth/UMD/BU is they split and the others win out. Think of .0100 in RPI as a full game in the standings, while .0050 is a half a game. Basically, they only lose half a game because Cornell is such a good team. It'll be enough for the RPI to tighten but not enough for BU to enter the top 12.

I expect BU will end up in the top 12 because one of those 5 ECAC teams in the top 12 has to lose an ECAC quarterfinal series.
 
Re: 2012 NCAA Tournament: Division I Bracketology

Grandaddy, Unless UMD losses to MSU they belong. BSU and OSU are strong teams at least one game on the weekend and just need to find that consistancy and they would be deserving.
16-13-1 @ .5500 wining %, RPI rank 12th in a class of 12 @ .5474, TUC record 4-9-1 for TUC % of .3214

Based on that I am unable to agree they belong...unless they get the auto bid for the WCHA. If winning consistently in each series is the measuring stick, Minnesota Duluth has exhibited similar difficulty as those mentioned. All regardless of WCHA strength, IMO. Would beating the last place team in the WCHA twice on the final weekend make a big difference?

Still a couple of weeks to go.

It is looking to me more and more like UMD needs an autobid to even get into the tournament.

Which, to be fair, I believe they are capable of achieving. It is a difficult task, but . . .
That's the way I see it. Wouldn't count them out before they lose in the WCHA tournament...if they lose.

Quick thought: if the top 4 WCHA teams do not make it into the tourney, i will be deeply disappointed. I respect the teams out east, but they sometimes don't respect the strength of the WCHA..
Maybe they're better than some teams ranked above them in the PWR. I don't necessarily buy into that based on the WCHA / strength argument...or buy that it can be assumed they're better than the HE or ECAC teams ranked above them. With the current system / criteria I don't think you could justify their participation without the auto bid. On the home ice...I'd never count them out 'til the final buzzer. Sans an auto bid I can't say they've proven their case.

Yes, I realize things have changed a lot since the above quoted posts were originally posted.

Sorry Doc...my honest thoughts.
 
Re: 2012 NCAA Tournament: Division I Bracketology

On the subject of where UMD deserves to be ranked, it's interesting to consider why KRACH has UMD above Boston College. Think about whether you agree with this logic if you think KRACH is a better system.

Let's break down the seasons in the following way, aside from their head-to-head split.

UMD
1-7-0 vs. Wisconsin/Minnesota (Top 3 in RPI/KRACH)
2-1-1 vs. North Dakota (KRACH #4, RPI #7)
4-4-0 vs. Bemidji/OSU (KRACH #10-11, RPI 14-15)
8-0-0 vs. UConn/MSU/St. Cloud

BC
0-0-0 vs. Games against RPI/KRACH top 4
7-4-2 vs. Games against RPI/KRACH top 15
5-3-0 vs. Maine/Quinnipiac/Providence (KRACH 16-18, RPI 17-19)
8-0-1 vs. Teams outside of top 20 in RPI/KRACH

So how do you compare these two teams with very different schedules?

The way the current system works, UMD gets crucified for going 1-7 against UMD/Minnesota. KRACH basically infers BC would've done about the same (17% win pct. vs Minnesota, 11% win pct. vs. Wisconsin). But because UMD actually played these games, UMD gets crucified.

Now look at the rest of schedules.

Both BC & UMD basically took care of below-average teams (except BC's tie against UNH, but let's ignore that).

UMD underachieved going 4-4 against Bemidji and OSU, but not by much. KRACH infers BC would've had about a 60% win pct. against these teams. But BC also underachieved against that Maine/QPac/Providence group, posting a 62.5% when they should have been (and UMD projects to) a win percentage over 75%. So really that's a wash.

Which of these teams is better comes down to whether UMD's 2-1-1 against North Dakota is better than BC's impressive 7-4-2 against a mix of teams that includes Northeastern/Harvard/SLU/Clarkson/BU/Mercyhurst/Dartmouth.

KRACH thinks the 2-1-1 vs. North Dakota is more impressive because it believes North Dakota is better than any of those teams I listed. KRACH projects BC would have a 41% win pct. against North Dakota, but UMD posted a 62.5% win percentage. That is the main reason why UMD is ranked higher than BC in KRACH. But if you believe North Dakota's RPI ranking is closer to it's true ranking, then again we have a wash.

If you compare the body of work, UMD and BC are fairly similar teams, and KRACH gives a slight edge to UMD. At the very least, I think we can all agree that the current pairwise system is absurd to place UMD an order of magnitude below BC. Again, the current standing in the PWR is all about the penalty UMD gets for going 1-7 against Wisconsin and Minnesota.
 
Last edited:
Re: 2012 NCAA Tournament: Division I Bracketology

Let's break down the seasons in the following way, aside from their head-to-head split.

UMD ...
4-4-0 vs. Bemidji/OSU (KRACH #10-11, RPI 14-15) ...

BC ...
5-3-0 vs. Maine/Quinnipiac/Providence (KRACH 16-18, RPI 17-19) ...
I'd say that it is a bit of a stretch to equate Maine/QU/PC with BSU and OSU. Maybe I'm biased by having seen the latter two live, and the eastern teams only on webcasts, but BSU was 3-0-1 H2H with that group. Maybe the better record by BC evens the difference out -- two games over .500 versus level. I don't feel that there is much separation between UMD, BSU, and OSU, and I don't see BC as far ahead of that group as RPI does.

At the very least, I think we can all agree that the current pairwise system is absurd to place UMD an order of magnitude below BC. Again, the current standing in the PWR is all about the penalty UMD gets for going 1-7 against Wisconsin and Minnesota.
It may be two orders of magnitude: BC is a host team, while UMD doesn't even make the field.
 
Re: 2012 NCAA Tournament: Division I Bracketology

I'd say that it is a bit of a stretch to equate Maine/QU/PC with BSU and OSU. Maybe I'm biased by having seen the latter two live, and the eastern teams only on webcasts, but BSU was 3-0-1 H2H with that group.
Right, I was not intending to equate the groups. Both rankings agree that BSU & OSU are better -- though KRACH thinks BSU & OSU are much better.
Maybe the better record by BC evens the difference out -- two games over .500 versus level.
Yes, that's precisely the argument I was making. You could also through Princeton into that group instead of the lower-tier group and make BC look a little better, but probably the tie with UNH makes it all a wash.

I don't feel that there is much separation between UMD, BSU, and OSU, and I don't see BC as far ahead of that group as RPI does.

I don't think it's fair to say there isn't much separation between UMD and BSU/OSU. I know the H2Hs were all 2-2, but the big differences are pretty clear
(1) UMD split BC
(2) UMD went 2-1-1 vs. North Dakota while the other two were 1-3 each against UND.
(3) BSU & OSU each lost once against the league's bottom tier.
Otherwise the results are pretty much the same.

UMD is expected to be three full games ahead of OSU by the end of next week in the WCHA standings, plus they have that split of BC. That's the difference between being a team that should be in the top 8 and one that's clearly not -- if those results hold.

If you want to make a good argument that BC is better than UMD, it has to be that the BC 13 games against the top 15 is a more reliable body of work than UMD's 4 games against North Dakota.
 
Last edited:
I'd say that it is a bit of a stretch to equate Maine/QU/PC with BSU and OSU. Maybe I'm biased by having seen the latter two live, and the eastern teams only on webcasts, but BSU was 3-0-1 H2H with that group. Maybe the better record by BC evens the difference out -- two games over .500 versus level. I don't feel that there is much separation between UMD, BSU, and OSU, and I don't see BC as far ahead of that group as RPI does.

It may be two orders of magnitude: BC is a host team, while UMD doesn't even make the field.

You are definitely biased towards WCHA. Don't get me wrong....I think OSU is headed in the right direction under Nate Handrahan's guidance but they tied Princeton twice. Princeton lost to Niagara, St. Lawrence twice, Quinnipiac twice, Clarkson twice, split with Colgate. At best I think OSU would be 6th in the ECAC.
 
Re: 2012 NCAA Tournament: Division I Bracketology

I don't think it's fair to say there isn't much separation between UMD and BSU/OSU.
I'm saying this as a human having watched the teams, not as one looking at the results and trying to analyze them. I realized that by any measure, UMD has had better results. BSU is coming off a road trip where they went to Minneapolis and lost a one-goal game and tied, went to Madison and lost two one-goal games, went to Grand Forks and didn't play the best, but by then I think they were getting worn down, then came home and held UMD to one goal on the weekend. I don't think there are a lot of teams in the country that could do that. So I see UMD as better, but not by very much. I realize we don't design ranking system to mirror human perceptions. Humans aren't bound by fairness doctrines, at least as far as I can tell from reading this particular message board. :D

(3) BSU & OSU each lost once against the league's bottom tier.
BSU actually lost twice to MSU, both by 2-1 scores, once in OT, once when the Mavericks scored twice in 20 seconds. The Beavers don't score much, so they can be very vulnerable over the long haul. MSU is a screwy team. I can't figure out why the Mavs don't have better results, other than they share a league with six teams that are better than them, and a seventh that probably gets up to play them because it is their best chance to win. However, BSU also has some notable advantages in results. It did have a better record than UMD versus Minnesota. Also, BSU didn't lose in non-conference play. Given the performance of the teams, I think going 5-0-1 against PC, Maine, and RMU is more impressive than going 3-1 against BC and UConn. Rankings systems won't agree, but then they rate BC higher than I do. So IMO, UMD is better than BSU, just not by much. As for OSU and UMD, the team that gets better goaltending on that particular day is the better team. Usually, but not always, that is UMD. Both are vulnerable on D, both have some great forwards, just not enough of them.
 
Re: 2012 NCAA Tournament: Division I Bracketology

You are definitely biased towards WCHA. Don't get me wrong....I think OSU is headed in the right direction under Nate Handrahan's guidance but they tied Princeton twice. Princeton lost to Niagara, St. Lawrence twice, Quinnipiac twice, Clarkson twice, split with Colgate. At best I think OSU would be 6th in the ECAC.
I admit that I'm biased towards the WCHA; I see about a dozen games played by WCHA teams for every one game I see played by the rest of the country. That does warp my perception.

However, don't use the Princeton series as the gold standard for judging OSU. They were missing three players for the Meco Cup, including one of their two best forwards. While not as extreme a case, it is a little like Dartmouth beating Cornell during 4 Nations. I think OSU belongs somewhere in that grouping of the ECAC that stretches from Harvard to Quinnipiac, and on any given day, no result is that surprising. We could replay the games between those teams and get entirely different results. OSU might fall sixth, fifth, even fourth. Who knows.
 
ARM, in terms of forecasting future performance, I agree with you: no big difference between osu/umd/bsu.

In terms evaluating the body of work this season as you would for ncaa selection purposes, I stand by my what I already said.
 
Back
Top