The 2015 Pairwise, Bracketology and History Thread

[Numbers]

Re: The 2015 Pairwise, Bracketology and History Thread

AS FOR MINNESOTA....

I have examined all the possibilities using the Predictor on this site. My results are different than JimDahl and also, apparently, GPL.

Here is what I get, with reasoning... assuming Minnesota loses (if they win they are in, and a high #3 seed),
Any in by BU (knocks out Lowell and Providence{i think}), Harvard (knocks out Colgate), or StCloud (keeps Minnesota's RPI high because we played them twice and thus keeps us ahead of Harvard if they lose and ahead of Providence) puts Minnesota in. This is a 7/8 chance.

Now, for the other 1/8....The games in question are RIT/Mercyhurst;UND/DU;MTU/MSUM. There is no pecking order here. The results are very close in RPI. I found one scenario where the Gophers were two spots below the bubble, and Harvard qualified by leading BOTH Providence and Minnesota in the 5th decimal place of RPI.

This is what I got:
If RIT wins and MSUM wins: Gophers are in if DU does not win (means UND wins or tie)
If RIT wins and MTU wins: Gophers are in if UND does not win (means DU wins or tie)
If Mercyhurst wins and MSUM wins: Gophers are out
If Mercyhurst wins and MTU wins: Gophers are in if UND does not win (means DU wins or tie)

If someone could compare this to CHN, it would be nice....

 

[Priceless]

Re: The 2015 Pairwise, Bracketology and History Thread

So the Gophers cheering for North Dakota and Yale cheering for Harvard....

 

[The Rube]

Re: The 2015 Pairwise, Bracketology and History Thread

AS FOR MINNESOTA....

I have examined all the possibilities using the Predictor on this site. My results are different than JimDahl and also, apparently, GPL.

Here is what I get, with reasoning... assuming Minnesota loses (if they win they are in, and a high #3 seed),
Any in by BU (knocks out Lowell and Providence{i think}), Harvard (knocks out Colgate), or StCloud (keeps Minnesota's RPI high because we played them twice and thus keeps us ahead of Harvard if they lose and ahead of Providence) puts Minnesota in. This is a 7/8 chance.

Now, for the other 1/8....The games in question are RIT/Mercyhurst;UND/DU;MTU/MSUM. There is no pecking order here. The results are very close in RPI. I found one scenario where the Gophers were two spots below the bubble, and Harvard qualified by leading BOTH Providence and Minnesota in the 5th decimal place of RPI.

This is what I got:
If RIT wins and MSUM wins: Gophers are in if DU does not win (means UND wins or tie)
If RIT wins and MTU wins: Gophers are in if UND does not win (means DU wins or tie)
If Mercyhurst wins and MSUM wins: Gophers are out
If Mercyhurst wins and MTU wins: Gophers are in if UND does not win (means DU wins or tie)

If someone could compare this to CHN, it would be nice....

But what if there's a full moon?

I love trying to figure out PWR, and I hate trying to figure out PWR.

 

[Riz]

So the Gophers cheering for North Dakota and Yale cheering for Harvard....

If MN can't take care of their own business, I guess. Yale will be wearing Crimson tomorrow!

 

[Red Cows]

Re: The 2015 Pairwise, Bracketology and History Thread

But, the frozen four is always well attended. I get what your are saying if you are talking about regionals; but, the frozen four? Furthermore, frankly, I wouldn't be surprised to see a Florida school startup a program (especially with the recent ASU announcement & if any more schools follow). I just moved down to Tampa area from Wisconisn; and they have nearly as many high school programs,& more importantly, broadcast a high school hockey game every friday night on the local sports cable channel (ahead of bouncy ball). From what I hear, they actually had good promo & local deals for the frozen four last time it was in Tamps. Granted, I am biased; but, I think Tampa is a good frozen four destination (though I'm not sure you can really go wrong with that).

You are right that the regionals are a serious problem. The Frozen fours are not/will not change/ prove anything (IMO).

What tangible benefits, as a sport, does college hockey derive from having a Frozen Four in Tampa?

Promoting the sport (college hockey, not hockey, in general) in a place where the closest team is in Huntsville, Alabama and the next closest is a lot farther away than that?

I'd be willing to bet you everything I own that there is no D-1 college hockey in Florida anytime in the foreseeable future and I am quite dubious that the ASU announcement portends anything happening anytime soon at any other school, either. You need the money, the facilities, you have to comply with Title IX, and hockey is monstrously expensive. I have some doubts about the future success of the ASU foray into this. Huge travel costs and they have no place of their own to play. We'll find out in the next few years just how real their commitment is when they start footing actual bills. I hope they can make it work but this start-up seems woefully underfunded and the press conference where this was announced was vague on details, to say the least. The school officials involved barely answered a single pointed question from the press. I sure didn't have warm fuzzies about the future of hockey there after seeing it.

 

[The Rube]

Re: The 2015 Pairwise, Bracketology and History Thread

What tangible benefits, as a sport, does college hockey derive from having a Frozen Four in Tampa?

Promoting the sport (college hockey, not hockey, in general) in a place where the closest team is in Huntsville, Alabama and the next closest is a lot farther away than that?

I'd be willing to bet you everything I own that there is no D-1 college hockey in Florida anytime in the foreseeable future and I am quite dubious that the ASU announcement portends anything happening anytime soon at any other school, either. You need the money, the facilities, you have to comply with Title IX, and hockey is monstrously expensive. I have some doubts about the future success of the ASU foray into this. Huge travel costs and they have no place of their own to play. We'll find out in the next few years just how real their commitment is when they start footing actual bills. I hope they can make it work but this start-up seems woefully underfunded and the press conference where this was announced was vague on details, to say the least. The school officials involved barely answered a single pointed question from the press. I sure didn't have warm fuzzies about the future of hockey there after seeing it.

Iowa is more likely, given the new arena they are building.

 

[Red Cows]

Re: The 2015 Pairwise, Bracketology and History Thread

And Nebraska as well, given they have already built their arena and started construction on another (practice) arena right next door this past December.

And, I think this season's "something of a fiasco" season (so far) in Big 10 hockey will move the conference further in the direction of the two more teams (at least) that they need, badly.

The Huskers have the money, a brand new arena to play in, and a practice facility that will be completed by this fall.

 

[Numbers]

Re: The 2015 Pairwise, Bracketology and History Thread

I would like to comment on the statistical analysis of PatMan and others above.

There is wonderful agreement on this message board that RPI is a rough equivalent of 'how to make sausage.' It's obvious that the metric was chosen in the beginning as a rough way to compare basketball teams, then adapted to hockey. Then, in the minds of people who use it, it became somewhat of a gold standard. And, then it has been tweaked often in its use for hockey. Obviously, the off-and-on tweaking merely exposes it for what it is: A not-very-exact way to measure how a team played against its schedule.

The question is: How would anyone make a metric that worked better? On this site, KRACH has often been touted. I have done this myself. The reason is that KRACH can post-predict everyone's record up until the point at which is used. And, it has several other intuitive results which are correct, such as "if you lose, your KRACH goes down. If you win, it goes up." It does not need tweaking.

However, this season has exposed KRACH as well. It's a perfect metric for what it does. However, when the vast majority of teams still play a very insular schedule, what KRACH does is over-emphasize the importance of early season, non-conference games. As a metric, it's perfect. What it is attempting to measure, however, is a bit more ghost-like and hard to actually put your finger on. As much as a respect the NCHC for having great teams this year, it defies logic, rational thought, history, and in some ways the eye test, to believe that something like 6 of the best 9 teams in the country all play in the same conference.

How do you devise a system that does what you want it to do in a year like this? It's a big question, with no perfect answer.

If you guarantee a more even split of berths in the tournament, how to you do that non-artificially?
You can't mandate more out-of-conference games....
etc.

 

[The Rube]

Re: The 2015 Pairwise, Bracketology and History Thread

There is no perfect metric in sports. None. You just have to find the BEST one that comes closest to your needs.

 

[MagnessMan]

Re: The 2015 Pairwise, Bracketology and History Thread

What tangible benefits, as a sport, does college hockey derive from having a Frozen Four in Tampa?
Promoting the sport (college hockey, not hockey, in general) in a place where the closest team is in Huntsville, Alabama and the next closest is a lot farther away than that?

I've been attending the FF's since I went to one here in my home town in 2008. Florida was hugely popular with the attendees. Mainly because it was warm and the complex is built for the weekend. Everything is centralized. Walk from your hotel to the game. Good restaurants etc. That's why were all headed back there after Chicago. The site location isn't really about local promotion, never has been, never will be. The target audience are a select group of nuts who love the game. Their love crosses generations so they come in groups and they come year after year. Promotion of the game happens at the youth hockey level.

 

[The Rube]

Re: The 2015 Pairwise, Bracketology and History Thread

I've been attending the FF's since I went to one here in my home town in 2008. Florida was hugely popular with the attendees. Mainly because it was warm and the complex is built for the weekend. Everything is centralized. Walk from your hotel to the game. Good restaurants etc. That's why were all headed back there after Chicago. The site location isn't really about local promotion, never has been, never will be. The target audience are a select group of nuts who love the game. Their love crosses generations so they come in groups and they come year after year. Promotion of the game happens at the youth hockey level.

That's why I loved Denver. Light rail to and from downtown (or rather, the Mall/etc). I have ZERO sense of direction, and get lost QUITE easily. Denver? Within a day, I could tell you where everything is, no problem. Even some parts of the outskirts of Denver I am familiar with at this point.

 

[Numbers]

Re: The 2015 Pairwise, Bracketology and History Thread

There is no perfect metric in sports. None. You just have to find the BEST one that comes closest to your needs.

And how would yours look for college hockey?

 

[Handyman]

Re: The 2015 Pairwise, Bracketology and History Thread

The Tampa Frozen Four was great...second only to St. Paul in the 4 I have attended. The people were friendly, the arena was great and the bars were awesome. You could walk everywhere.

And i love the conspiracy theories that show up about how the NCAA doesnt want an all X Frozen Four. The paranoia people have about this is laughable. The NCAA does not care. The fans do because it would suck (The All WCHA Frozen Four was horrid) but the the games sell the same tickets either way. The NCAA isnt going to waste 2 seconds caring or acting out to stop it...

 

[The Rube]

Re: The 2015 Pairwise, Bracketology and History Thread

And how would yours look for college hockey?

Honestly, the PWR is fine with me. I cannot fully figure it out, granted, but I generally get it (not for lack of intellect, it's a lack of effort).

 

[The Rube]

Re: The 2015 Pairwise, Bracketology and History Thread

The Tampa Frozen Four was great...second only to St. Paul in the 4 I have attended. The people were friendly, the arena was great and the bars were awesome. You could walk everywhere.

And i love the conspiracy theories that show up about how the NCAA doesnt want an all X Frozen Four. The paranoia people have about this is laughable. The NCAA does not care. The fans do because it would suck (The All WCHA Frozen Four was horrid) but the the games sell the same tickets either way. The NCAA isnt going to waste 2 seconds caring or acting out to stop it...

The St Paul one was the best, obviously. /tArrogance.

 

[HarleyMC]

Re: The 2015 Pairwise, Bracketology and History Thread

I would like to comment on the statistical analysis of PatMan and others above.

There is wonderful agreement on this message board that RPI is a rough equivalent of 'how to make sausage.' It's obvious that the metric was chosen in the beginning as a rough way to compare basketball teams, then adapted to hockey. Then, in the minds of people who use it, it became somewhat of a gold standard. And, then it has been tweaked often in its use for hockey. Obviously, the off-and-on tweaking merely exposes it for what it is: A not-very-exact way to measure how a team played against its schedule.

The question is: How would anyone make a metric that worked better? On this site, KRACH has often been touted. I have done this myself. The reason is that KRACH can post-predict everyone's record up until the point at which is used. And, it has several other intuitive results which are correct, such as "if you lose, your KRACH goes down. If you win, it goes up." It does not need tweaking.

However, this season has exposed KRACH as well. It's a perfect metric for what it does. However, when the vast majority of teams still play a very insular schedule, what KRACH does is over-emphasize the importance of early season, non-conference games. As a metric, it's perfect. What it is attempting to measure, however, is a bit more ghost-like and hard to actually put your finger on. As much as a respect the NCHC for having great teams this year, it defies logic, rational thought, history, and in some ways the eye test, to believe that something like 6 of the best 9 teams in the country all play in the same conference.

How do you devise a system that does what you want it to do in a year like this? It's a big question, with no perfect answer.

If you guarantee a more even split of berths in the tournament, how to you do that non-artificially?
You can't mandate more out-of-conference games....
etc.

All valid concerns my friend. I can guarantee you those teams that get locked out because of this statistical disaster should also be very concerned IF they understand that after all their *** breaking work on and off the ice, this system essentially ruined their season. This is probably not the appropriate venue for deep academic discussions on stats procedures. But I'll give it a go.

John Whalen, a physics prof at RIT helped design the pairwise tools on this site. BTW, he did a fine job, but I'll bet he abhors this beast now and prefers the Bradley-Terry model. But the BT model (as you stated) has similar flaws in sports applications in the substrata of latent factorials as well.

In order to fix this colossal mess a hybrid probability model must be designed, which could be based upon a variation of the Bradley-Terry-Luce logistic regression or Thurstone's model for scaling dominant matrices, but in addition it absolutely needs a semi-parametric approach that includes recursive partitioning and a homogenous scaling of covariates. This results in a tree-structured partition in the covariate space that allows for valid discriminatory analysis and then the parameters can be determined as usual with Maximum Likelihood Estimation (MLE). This will require field testing to determine where the parameter instabilities are in the model in a time series under the null hypothesis. But that's only a part of the problem. The NCAA SOS determinations may need to be revamped and the distribution of NC and league games must be restructured to improve equality of variance or the model itself will replicate some of the current statistical idiosyncrasies. My .02.

 

[goblue78]

Re: The 2015 Pairwise, Bracketology and History Thread

HarleyMC: The flaw in your thoughts about this come from trying to precisely measure something that the data are too noisy to allow precise measurement of. The NCAA is well aware, for example, that it could get more precise estimates of whether Team A is better than Team B by using the scores of the games rather than just using the results of the games. They refuse to do this because it has the unfortunate side effect of teams running up scores to improve rankings. (It has a second side effect that taking a high risk strategy to win a game may lower your ratings as well.) So when you play 40 games or so, and a few of them are obvious mismatches of talent, you just can't learn that much from who wins a game. Refining the methodology for discrimination with an MLE technique doesn't help because the results themselves are unstable. But there's a fundamental issue you're missing: the teams from, say, 10-25 in any rating system just aren't that different one another. They almost can't be and have a sport anyone wants to watch. So all the NCAA wants is a system that (a) takes some set of teams that are OK; and (b) takes conference winners automatically. Any such system has to leave out teams who are statistically indistinguishable from some of the teams that are in -- but that's because those teams aren't just statistically indistinguishable; they're actually indistinguishable. You can discriminate using a model, but the model isn't stable to perturbation, as Patman said above.

One more thing. You say you need to revamp scheduling to allow better estimation: that's exactly what the NCAA is uninterested in. Conferences are what the thing is about, and the Tournament at the end is gravy. The less that is known about interconference strengths come tournament time, the happier the NCAA is.

 

[Tipsy McStagger]

"in this world"? It's done all the time, pal. Think casuality for a change. Academically, this is a statistical nightmare and violates some of the most basic assumptions necessary to establish internal validity. If an undergrad....its probably publishable? I'm not talking at all about the model. I conduct a research study using a probability model involving gender, I run the procedures but I treat the sample without any discrimination. I challenge you to find ONE peer reviewed research journal that will publish something like this. You'll end up with a pile of rejection letters. With this system forget about statistical power because as it stands it's non-parametric.

So, you must have been banging this drum last year too, right?

 

[HarleyMC]

Re: The 2015 Pairwise, Bracketology and History Thread

HarleyMC: The flaw in your thoughts about this come from trying to precisely measure something that the data are too noisy to allow precise measurement of. The NCAA is well aware, for example, that it could get more precise estimates of whether Team A is better than Team B by using the scores of the games rather than just using the results of the games. They refuse to do this because it has the unfortunate side effect of teams running up scores to improve rankings. (It has a second side effect that taking a high risk strategy to win a game may lower your ratings as well.) So when you play 40 games or so, and a few of them are obvious mismatches of talent, you just can't learn that much from who wins a game. Refining the methodology for discrimination with an MLE technique doesn't help because the results themselves are unstable. But there's a fundamental issue you're missing: the teams from, say, 10-25 in any rating system just aren't that different one another. They almost can't be and have a sport anyone wants to watch. So all the NCAA wants is a system that (a) takes some set of teams that are OK; and (b) takes conference winners automatically. Any such system has to leave out teams who are statistically indistinguishable from some of the teams that are in -- but that's because those teams aren't just statistically indistinguishable; they're actually indistinguishable. You can discriminate using a model, but the model isn't stable to perturbation, as Patman said above.

One more thing. You say you need to revamp scheduling to allow better estimation: that's exactly what the NCAA is uninterested in. Conferences are what the thing is about, and the Tournament at the end is gravy. The less that is known about interconference strengths come tournament time, the happier the NCAA is.

No, you're wrong. On the contrary the "noise" you refer to is why the discussion exists. Your command of stats is obviously weak and you missed key points of my post. Recursive partitioning has been tested for quite sometime and used in the exact the manner I described. No, the MLE is not unstable in the model I suggested. Typically, the MLE is not derived by direct maximization of the multinomial likelihood but instead by fitting a surrogate log-linear Poisson model for the aggregated frequencies and this procedure can be performed in any number of statistical software packages.

Whether the NCAA chooses to recognize and change this current debacle is not the reason for my post nor is it my problem. I would add the tournaments in their current form should be axed and other post season parameters adjusted according to a new equitable format. I'm not going to take anymore time on this thread. With consideration for others, this is not an appropriate place for this type of discussion.

So, you must have been banging this drum last year too, right?

Doesn't matter. The change would have still been a welcome relief to many other teams and with most likely a minor effect on last year's Gopher team.

 

[goblue78]

Re: The 2015 Pairwise, Bracketology and History Thread

First, this isn't a B1G Tourney thread; second, rather than getting is a ****ing match over whose command of statistics is better, I will respectfully disagree. If you think there is no noise, ie that there is informative content worth discriminating over, well, we disagree. And your comments about the method for calculating the MLE rather than the important question, which is whether the maximum likelihood has any particular claim on our attention show that we're discussing the problem at different levels, which is fine. I didn't mean to suggest instability in the estimation algorithm which,as you point out, can be stabilized through partitioning. I meant that there is actual noise in the estimate which makes the maximum of the likelihood not particularly compelling as a discriminator. It discriminates, sure, but in a method that is at best marginally farther from nonarbitrary than the simplest Bradley-Terry models, ie, KRACH which in turn are at best marginally better than PWR. And my point above is that discrimination has a very limited purpose: to pick a set of 16 teams in a deterministic way that (a) bears some relationship to quality and that (b) encourages, marginally, better out of conference scheduling.