2013-2014 Women's D-I PairWise Contentions and Affirmations

[robertearle]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

Fair enough. I'm just of the opinion that North Dakota doesn't have much of an argument to say 'we should have been ranked higher than x y and z" last year based on the fact that there are so few data points between the two regions and their biggest out of conference series was a split with Clarkson.

Having paid quite close attention to UND and Wisconsin last year, the conclusion I came to was UND got the #8 slot (as Arlan has said) based on their better RPI; and their RPI was better because of that split with Clarkson.

UND played one more NC series last year than UW did, and it was the Clarkson series that was the 'extra' series for UND. Both played Lindenwood, UND played RIT, Wisconsin played New Hampshire, and then UND played Clarkson while UW 'took the week off'. UND and UW essentially tied in the WCHA season standings, they went 2-2-1 head-to-head.

So one could say 'the system worked' in that UND was rewarded by playing the 'extra' series against a tough opponent.

 

[TonyTheTiger20]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

Now that this thread has rehashed yet again the traditional east vs. west brouhaha, I'm curious why people haven't picked up on the Canada vs USA gauntlet thrown down by the above-quoted post. (do we have an emoticon for ducking while brickbats begin to fly?)

People may have felt that *I* was trolling in my posts, but pokechecker was definitely trolling in that post, so people generally ignored it.

 

[ARM]

If Northeastern was the team last year that was 4th (or whatever) in KRACH and 8th in PWR, I don't think we hear the same frustration coming out of Minnesota about playing the Huskies. Not because they played North Dakota 5 times already but because I think it's perceived that the eastern teams are just inherently weaker.

The bigger problem was playing UND for the sixth time. Minnesota was a confident bunch last year, and even with a banged up Kessel, which did make a huge difference, liked its chances against anyone. But to have to play the same team that you beat a week before in the same building, that posed a difference in motivation for the two sides that wouldn't have existed versus Northeastern or any other non-WCHA opponent. Plus, there was just another bit of history with UND and the Lam's. Look at it another way -- the Lam's are gone, and UND is a top-10 team; Coyne is gone, and Northeastern is not. Who would you rather have played if you are honest?

 

[TonyTheTiger20]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

Having paid quite close attention to UND and Wisconsin last year, the conclusion I came to was UND got the #8 slot (as Arlan has said) based on their better RPI; and their RPI was better because of that split with Clarkson.

UND played one more NC series last year than UW did, and it was the Clarkson series that was the 'extra' series for UND. Both played Lindenwood, UND played RIT, Wisconsin played New Hampshire, and then UND played Clarkson while UW 'took the week off'. UND and UW essentially tied in the WCHA season standings, they went 2-2-1 head-to-head.

So one could say 'the system worked' in that UND was rewarded by playing the 'extra' series against a tough opponent.

That's interesting, I didn't know that; did Wisconsin really not play the maximum number of allowable games?

Is that common? If so I'm guessing it's a result of the travel difficulties in the women's sport compared to the men's?

 

[ARM]

UND and UW essentially tied in the WCHA season standings, they went 2-2-1 head-to-head.

That's incorrect; UW had a 3-2 edge over UND H2H.

 

[TonyTheTiger20]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

Look at it another way -- the Lam's are gone, and UND is a top-10 team; Coyne is gone, and Northeastern is not. Who would you rather have played if you are honest?

Oh, definitely Northeastern, but North Dakota was a better team (and as such, got into the tournament over Northeastern).

What I meant was, if you took the same roster of girls that made the NCAA tournament (MLam et. al) and put them in a Northeastern jersey, that the perception would have been that that team would not have been as good.

 

[robertearle]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

That's incorrect; UW had a 3-2 edge over UND H2H.

Yup. My mistake.

All the more reason to think that the UND-Clarkson RPI 'effect' made the difference in the end.

 

[ARM]

What I meant was, if you took the same roster of girls that made the NCAA tournament (MLam et. al) and put them in a Northeastern jersey, that the perception would have been that that team would not have been as good.

That would depend on their record playing Northeastern's schedule. I do think that there is a benefit to playing a WCHA schedule. Hockey East coaches are always saying that their league is the most competitive top-to-bottom, but that has to either be the party line or based entirely on the fact that they know so little about the WCHA. Brian Durocher seems to have the most respect for the WCHA as a conference, and I think that's in part because he has put more effort into playing those teams than the other top teams have.

But one of the biggest reasons why we want to play a team from the East is because it is fresh and different. There are still teams out there that we've never played. As just a fan of the game, I'd rather seen anybody from the East, no matter how good they are, than a 5th or 6th game against some WCHA team in the first round. Variety is the spice …

 

[robertearle]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

That's interesting, I didn't know that; did Wisconsin really not play the maximum number of allowable games?

Is that common? If so I'm guessing it's a result of the travel difficulties in the women's sport compared to the men's?

I don't know if it is a travel/expense thing, or a preference of Mark Johnson's. But UW has only four (instead of six) NC games again this year; Lindenwood again, and a four team 'mini-tournament' weekend in Vail, where they played Northeastern one night and BU the next. But having played two teams once at least gives them more 'common opponent' data points for the Pairwise to chew on. (Other years, they've played six NC games.)

 

[ARM]

All the reason to think that the UND-Clarkson RPI 'effect' made the difference in the end.

It wasn't just that. Wisconsin's results against last-place Bemidji State hurt its RPI. Also, UND got a little more boost out of the WCHA quarters by playing and defeating a slightly-better Mankato rather than SCSU as UW did.

 

[Eeyore]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

Hey if people want to just use KRACH to determine the field, fine, I'm all aboard that train. The more math the better. But my frustration comes from where it appears that people want to give North Dakota (as the obvious example) a leg up because 'hey, they must be stronger, they're in the WCHA and have to play Minnesota.'

That's a complete mischaracterization of the position of most of the people who participate in this argument on this board. We have reasons to dislike PWR that go far beyond which team or conference is hurt by it. If you don't want to be called a playacting *** then start addressing the arguments that we make about it rather than engaging in bull**** speculation about our motives.

I'm not playacting -- let me put this another way: If Northeastern was the team last year that was 4th (or whatever) in KRACH and 8th in PWR, I don't think we hear the same frustration coming out of Minnesota about playing the Huskies. Not because they played North Dakota 5 times already but because I think it's perceived that the eastern teams are just inherently weaker.

I can't speak for Minnesota, just myself. I probably wouldn't have expressed the same level of frustration but I definitely would have been of the opinion that Northeastern was getting screwed and I'd have said so. I've been involved in arguing about the relative merits of PWR and other ranking systems for a long time before I started following women's hockey and have argued it consistently no matter which way the difference was going.

 

[ARM]

I don't know if it is a travel/expense thing, or a preference of Mark Johnson's.

Both Mark Johnson and Brad Frost seem to want to avoid making their nonconference schedule too tough, but IMO, they err on the side of making it not strong enough. I think they do this because the WCHA schedule is already so grueling, but IMO, UMD has had the better approach. Make that schedule as tough as you can and at least give yourself a chance to prove your worth on the ice. If you play weak teams out of conference or not enough out of conference games, then you are relying on the teams like MSU and BSU to prove the WCHA's strength for you.

 

[mnpokecheck]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

This years OOC Records....

CHA 27-31-4 (.468)
ECAC 35-38-9 (.482)
Hockey East 33-40-9 (.457)
WCHA 25-11-2 (.684)​

...so is it perceived or is it reality that the west is stronger. 3 leagues are under .500 OOC.

 

[ARM]

This years OOC Records....

CHA 27-31-4 (.468)
ECAC 35-38-9 (.482)
Hockey East 33-40-9 (.457)
WCHA 25-11-2 (.684)​

...so is it perceived or is it reality that the west is stronger. 3 leagues are under .500 OOC.

What is the WCHA's record if you subtract all game's versus Lindenwood? It will still be above .500, but not to the inflated degree that it now is.

 

[joegrav]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

This is a really interesting discussion, but I've been having trouble trying to make a post because all of my thoughts on this subject are pretty much a ramble. I'm going to give it a shot, though.

I recently had a long discussion with a coworker about the forthcoming NCAA football 4-team playoff, which is starting up next year. As we all know, college football is pretty much the only sport where we really do decide who plays for the championship based largely on subjectivity - experts deciding "by virtue of watching these teams play non-mutual opponents, I've decided that Team A is better than Team B."

I made the argument that with 5 "power conferences," another set of non-power D1A conferences, plus independents like Notre Dame, BYU, and the service academies, there is no way in hell you should ever put two teams from the same conference into a four team playoff, even if they do come from the almighty SEC. Otherwise, there's no point in having a playoff; you might as well go back to letting pollsters pick the champion, because we are deciding subjectively that it's self-evident that, say, Alabama and Auburn are both better than Big Ten champion Michigan State without ever seeing those teams play one another on the field. While there's no fair, 100% fool-proof way to determine who is the "best team in the country," we have a pretty decent sample size and basis of comparison to determine who the best team is within a particular conference. If you haven't proven, over a large-ish sample size, that you're #1 in your own league, you are obviously not #1 in the country, even if being #2 in your own league subjectively makes you better than the #1 teams in all of the other leagues.

This is all a long-winded way of saying that once you start letting at-large teams in, it's going to be inherently unfair however you slice it or dice it, and we need to accept that as part of the package that comes with playing a single-elimination tournament to determine a champion. This is sports, not brain surgery, philosophy class, or a Presidential election, so we're allowed to have fun with it and do things for entertainment value rather than try to reach some sort of objective truth.

As I read these arguments and think about how to modify the NCAA tournament system going forward, I start with the following assumptions:
1. Single elimination hockey is always unfair.
2. Trying to compare teams that never play one another or have many common opponents is always unfair.
3. Comparing teams based on common opponents on the basis of a small sample size of a couple of OOC games each year is always unfair.
4. Hockey is unfair.
5. Life is unfair.
6. The universe is fleeting and random.

With all of that said, I think the goal should be to develop a fair, objective set of criteria to select the NCAA tournament field, with the knowledge that from a subjective standpoint, you're never really going to get an agreement about who the "top 8" teams are nationally. It's clear that the PWR has flaws because it requires a much bigger sample of non-conference games than women's college hockey provides. However, it's a pretty decent method, and the inherent FAIRNESS in the current system is that EVERYONE has, basically, two shots to 100% ensure themselves a spot in the tournament: either by being the best in their league over the decent sample size of a full season, which virtually guarantees tournament entry (IMO, it SHOULD guarantee tournament entry), OR proving your worth on ice by winning the conference tournament and gaining an auto-bid. No team in an autobid conference is shut out from this possibility.

If I were designing the tournament from scratch, I would probably only include conference tournament and regular season champions, with 2nd-place regular season teams filling in if the same team wins both titles. At that point, they can seed the tournament via polls for all I care. Or by geography. It doesn't really matter to me, because there's no real way of knowing who's #1 other than within conferences.

I hope that all made some semblance of sense. Resume your bickering now!

 

[mnpokecheck]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

What is the WCHA's record if you subtract all game's versus Lindenwood? It will still be above .500, but not to the inflated degree that it now is.

16-11-2

 

[Eeyore]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

What is the WCHA's record if you subtract all game's versus Lindenwood? It will still be above .500, but not to the inflated degree that it now is.

This is a methodologically flawed thing to do. (I assume ARM knows this but it's the kind of thing I see people do all the time so I bring it up.) Two years ago, when Lindenwood probably was the worst team in Division I by a significant margin, it might be justifiable but it isn't now. Otherwise you skew the records of the different conferences by allowing some to keep cheap non-conference wins but not others.

To make any sort of comparison you would have to subtract all of the non-conference games that everyone has played against teams that are comparatively as bad as Lindenwood is. I'd probably argue that you should take out: Brown; Maine; Penn State; Dartmouth; Colgate; and Union. You could put the cutoff in a number of other places but it shouldn't be at Lindenwood and No One Else.

 

[ARM]

If I were designing the tournament from scratch, I would probably only include conference tournament and regular season champions, with 2nd-place regular season teams filling in if the same team wins both titles. At that point, they can seed the tournament via polls for all I care. Or by geography. It doesn't really matter to me, because there's no real way of knowing who's #1 other than within conferences.

Your approach makes nonconference games irrelevant as far as NCAA Tournament selection goes. I agree that any approach is imperfect, and thus, unfair. However, I do think we owe it to the student athletes to attempt to make it as fair as possible given the constraints. They work very hard at their sport, so it should be on as level a playing field as is possible. So to say that the 2nd place team from a six-team CHA is more deserving than the 3rd place team from a 12-team ECAC every season seems to be a flaw that should not exist in a newly-designed process.

 

[joegrav]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

Your approach makes nonconference games irrelevant as far as NCAA Tournament selection goes. I agree that any approach is imperfect, and thus, unfair. However, I do think we owe it to the student athletes to attempt to make it as fair as possible given the constraints. They work very hard at their sport, so it should be on as level a playing field as is possible. So to say that the 2nd place team from a six-team CHA is more deserving than the 3rd place team from a 12-team ECAC every season seems to be a flaw that should not exist in a newly-designed process.

I get that. But to me, if the goal of the tournament is to crown the Champion, if you haven't proven that you're #1 in your league over a full season, you don't really get much sympathy from me on this one. If we're trying to be fair to the student-athletes, it's UNFAIR to let in teams that were objectively worse over a bigger sample size facing the *same competition* and give them a crack at knocking off the top team single-elimination - even if this single-elimination matchup doesn't happen until the title game. While we can make an educated hypothesis that Minnesota is better than whichever team wins Hockey East, we have a lot less basis for that comparison this year than we do a comparison between Minnesota, Wisconsin, and North Dakota, who we get to see play against one another home and away multiple times.

So let's be real, the point of having a tournament isn't to be fair. The point of having a tournament is that tournaments are cool.

Non-conference games under my theoretical system could be used as a way for pollsters to determine who gets seeded where in the tournament. Or you could just get rid of them and let there be some mysterious fun about tournament clashes, sort of like the olden days of the World Series. It's honestly not the worst idea, given that teams are reluctant to travel far for games due to the costs.

 

[Eeyore]

Re: 2013-2014 Women's D-I PairWise Contentions and Affirmations

It's clear that the PWR has flaws because it requires a much bigger sample of non-conference games than women's college hockey provides.

This is not actually the problem with RPI/PWR. Even with a larger slate of non-conference games it's still flawed. It will systematically overrate teams that play in weak conferences. This is an inherent problem with the way that it relies only upon raw winning percentage in calculating strength of schedule. Without some other check on who your opponents have played it can't get this right. Adding opponents' opponents' winning percentage (to say nothing of further iterations of that) to the formula doesn't solve this problem because that approach converges to .500 too rapidly to provide enough information to the calculation.

So the number of non-conference games needed to largely mitigate the flaws of RPI/PWR would have to be large enough that it swamps the effect of teams' conference schedules. And in addition to being plentiful, the schedule would have to involve every team playing a slate of opponents that constitutes a representative sample of the overall talent pool. Just adding more non-conference games that are still skewed in terms of quality of competition does nothing to help RPI.

KRACH and Rutter do a better job of handling this issue.