The 2014 Pairwise, Bracketology and History Thread`

[Slap Shot]

Re: The 2014 Pairwise, Bracketology and History Thread`

Meh. Play who you have to, but I don't think the NCAA cares as much about attendance as some assume. Regionals are bid by cities and hosts which guarantee revenue to the NCAA. When revenues come in under the bid amount, the host is responsible for the difference; if they are over the bid amount, the host gets the bulk of the reward.

 

[Happy]

Re: The 2014 Pairwise, Bracketology and History Thread`

Like Holy Cross and Yale?
You'd think you western folk would want your regional stuffed with western teams... that way at least one of you has a chance of advancing!

Man... nothing like a 70 mph fastball over the heart of the plate.

yep, just as soon as one of the weak Eastern confernces can do this,

http://www.insidecollegehockey.com/6History/ncaa_05.htm

you can come and talk to me about who has a chance of advancing to the Frozen Four.

 

[Onion Man]

Re: The 2014 Pairwise, Bracketology and History Thread`

Havent t teams from the East won 5 of the last 6 titles?

 

[Numbers]

Re: The 2014 Pairwise, Bracketology and History Thread`

I see. Wisco is also a common opponent that BC has a win over. BC lost to Mich. Lowell has wins over Mich, Mich State and Penn St. How's that figure in? LOL... You number nerds are awesome. I mean that in a good way.

Well, the figuring goes like this:
Minnesota v BC head-to-head: Minnesota won, and there was a tie. For PWR purposes, this is +1 to Minny.

Let's assume Minnesota loses all games against teams who are not common opponents to Minny and BC, then..
Minnesota v BC RPI: BC has a better RPI, so for PWR purposes, this is +1 to BC.

That leaves the common opponents to decide the comparison and the overall #1 seed. There are 6 common opponents: New Hampshire, Notre Dame, RPI, Michigan, Wisconsin, and Penn State. ONLY games against those teams count for this part of the comparison, so every game that Lowell played all year is immaterial here.

Let's look at these common opponents. The way the calculation is done is that you figure the winning percentage of both teams (MN and BC) against that opponent. This means that 4-0 is the same as 2-0. Then, add these numbers for all the opponents. Higher total is the winner.

Against RPI, both MN and BC won, so this doesn't figure in. All the rest figure in some way, potentially....

Against NoDame, Minn was 1-1, and BC is currently 1-1. If NoDame loses to BU, this stays that way... But, let's assume ND beats BU, and then BC sweeps ND. That means BC ends up 3-1. So, for NoDame, Minnesota is 1-1 (.500) an BC is 3-1 (.750).

Against New Hampshire, Minnesota is 1-0, and BC is 2-0. Let's assume BC either does not play UNH in the HEA semis, or defeats them. That gives both teams 1.000 here.

At this point, let's access. BC is slightly ahead - by .250 altogether.

Now, Michigan. We assume Minnesota sweeps, because I am stating that Minnesota clinches the #1 with a sweep in this case. That sweep gives Minnesota 1.000 against Mich, while BC was 0-1. Minnesota now leads by .750

Now, Penn State. BC was 2-0, Minnesota was 4-0, but Minnesota could lose to Penn State in the Big10 playoffs. Losing to PennState makes Minny 4-1. BC 1.000, Minnesota .800, and Minnesota still leads by .550

Finally, Wisconsin. BC was 1-0, Minnesota 2-2. BC 1.000, Minnesota .500. Add those in and Minny wins by the slightest of margins - .050.

Note that after sweeping Michigan, Minnesota could lose one game to PSU, Wisc, or Mich. None are worse for them than PSU, so I chose that one.

And, you could go back up to the top and do the same work-through assuming ND loses to BU, in which case, BC can't add to their split with ND right now, and then all Minny needs is a split against Michigan next week.

And, again, assuming ND wins, but Merrimack does, too. Then, BC can only get one win max against ND, and therefore Minny needs a W and a T at Yost next week.

Cheers. It's just simple math (copyright Scooby)

 

[Priceless]

Re: The 2014 Pairwise, Bracketology and History Thread`

Anyway, back to conferences that still exist:

Notre Dame could be upset by BU this weekend, miss the Hockey East tournament entirely and still qualify for the national tournament.

 

[FiveHole12]

Re: The 2014 Pairwise, Bracketology and History Thread`

Well, the figuring goes like this:
Minnesota v BC head-to-head: Minnesota won, and there was a tie. For PWR purposes, this is +1 to Minny.

Let's assume Minnesota loses all games against teams who are not common opponents to Minny and BC, then..
Minnesota v BC RPI: BC has a better RPI, so for PWR purposes, this is +1 to BC.

That leaves the common opponents to decide the comparison and the overall #1 seed. There are 6 common opponents: New Hampshire, Notre Dame, RPI, Michigan, Wisconsin, and Penn State. ONLY games against those teams count for this part of the comparison, so every game that Lowell played all year is immaterial here.

Let's look at these common opponents. The way the calculation is done is that you figure the winning percentage of both teams (MN and BC) against that opponent. This means that 4-0 is the same as 2-0. Then, add these numbers for all the opponents. Higher total is the winner.

Against RPI, both MN and BC won, so this doesn't figure in. All the rest figure in some way, potentially....

Against NoDame, Minn was 1-1, and BC is currently 1-1. If NoDame loses to BU, this stays that way... But, let's assume ND beats BU, and then BC sweeps ND. That means BC ends up 3-1. So, for NoDame, Minnesota is 1-1 (.500) an BC is 3-1 (.750).

Against New Hampshire, Minnesota is 1-0, and BC is 2-0. Let's assume BC either does not play UNH in the HEA semis, or defeats them. That gives both teams 1.000 here.

At this point, let's access. BC is slightly ahead - by .250 altogether.

Now, Michigan. We assume Minnesota sweeps, because I am stating that Minnesota clinches the #1 with a sweep in this case. That sweep gives Minnesota 1.000 against Mich, while BC was 0-1. Minnesota now leads by .750

Now, Penn State. BC was 2-0, Minnesota was 4-0, but Minnesota could lose to Penn State in the Big10 playoffs. Losing to PennState makes Minny 4-1. BC 1.000, Minnesota .800, and Minnesota still leads by .550

Finally, Wisconsin. BC was 1-0, Minnesota 2-2. BC 1.000, Minnesota .500. Add those in and Minny wins by the slightest of margins - .050.

Note that after sweeping Michigan, Minnesota could lose one game to PSU, Wisc, or Mich. None are worse for them than PSU, so I chose that one.

And, you could go back up to the top and do the same work-through assuming ND loses to BU, in which case, BC can't add to their split with ND right now, and then all Minny needs is a split against Michigan next week.

And, again, assuming ND wins, but Merrimack does, too. Then, BC can only get one win max against ND, and therefore Minny needs a W and a T at Yost next week.

Cheers. It's just simple math (copyright Scooby)

Thanks for all that.

 

[JimDahl]

Re: The 2014 Pairwise, Bracketology and History Thread`

Jim, could you please clarify the home/road weighting part of the Oppwin% and Oppoppwin%? thanks. I see 2 possibilities:
a) both are weighted according to whether you won or not: 0.8 for homewins % roadlosses, etc
b) the weighting goes strictly by home or road.

In the case of a), the calculation for each game would be: WF*(.25*result of game + .21*(oppwin%) + .54*(oppoppwin%)) and the pre-QWB RPI is SUM (over all games of those calculations) divided by SUM(over all games of the WF) WF=weighting factor.

In the case of b), the calculation could not be done game by game, but each game would have 3 pieces: WF1*.25*result, WF2*.21*oppwin%, and WF2*.54*oppoppwin% And, the pre-QWB RPI requires averaging all 3 of those pieces, and then adding them together.

A) looks neater in that it allows each game to be considered individually, and it's easy to see which games have to be removed because of being wins vs weak teams.

B) is perhaps conceptually better, because A weights your SOS according to whether you won or lost the game, and theoretically, SoS is independent of results.

Sorry, I've been away.

What we're all doing is mathematically equivalent to A**. The reason I originally implemented what we've come to call the USCHO formula (which they no longer use) was the same reasoning as yours -- thinking of win% as stand-alone and the other two as SOS makes you not want to weight SOS on game outcomes.

However, if one thinks of it like you stated -- each game has an RPI and those are weighted by location/outcome -- I actually come to prefer it. Plus it gets rid of a lot of nasty quirks caused by the components having different weightings.

**rather than calculating each game as having an RPI that gets weighted averaged with all the others, you can also still calculate each component separately with weights and divide by total potential points then combine at the end, which makes it a little more clear how it gets weird when the components have different weights.

 

[Numbers]

Re: The 2014 Pairwise, Bracketology and History Thread`

Anyway, back to conferences that still exist:

Notre Dame could be upset by BU this weekend, miss the Hockey East tournament entirely and still qualify for the national tournament.

In fact, how could Notre Dame miss the tournament at all presently. Providence, Northeastern and Vermont could all pass them. Michigan could pass them, and then Cornell maybe. Which would leave them 15th. WCHA upset??

That is no great long shot of a scenario, really, but at this point, I give them better than 50% even with a loss to BU.

 

[Numbers]

Re: The 2014 Pairwise, Bracketology and History Thread`

Sorry, I've been away.

What we're all doing is A. The reason I originally implemented what we've come to call the USCHO formula was the same reasoning as yours -- thinking of win% as stand-alone and the other two as SOS (so shouldn't be weighted on game outcome).

However, if one thinks of it like you do -- each game has an RPI and those are weighted by outcome/location -- I actually come to prefer it. Plus it gets rid of a lot of nasty quirks caused by the components having different weightings (rather than think of each game as having an RPI that gets weighted then averaged with all the others, you can also still calculate each component separately with weights and divide by total potential points).

Thanks for the reply. I initially thought I liked this way, because the idea of each game having its own component seems right.

However, I have come to think that it is not correct, because my SoS should not depend on whether I won or lost. But, really, however you do it, the whole RPI calculation is kind of bogus anyway. The 25%, 21% and 54% are just pulled out of a hat (it's my opinion that they were chosen by an actuary employed by the NCAA to get results close to KRACH without using KRACH). The QWB is just a guess. Etc. So, I guess the committee can calculate however they want....

 

[Aerman]

Re: The 2014 Pairwise, Bracketology and History Thread`

... I have come to think that it is not correct, because my SoS should not depend on whether I won or lost ...

This

 

[LTsatch]

Re: The 2014 Pairwise, Bracketology and History Thread`

No. They are equivalent to Yale though.

Can any team that has ever won the NC ever beaten three one seeds and a two? How is CC equal to Yale?

 

[Tipsy McStagger]

Re: The 2014 Pairwise, Bracketology and History Thread`

Can any team that has ever won the NC ever beaten three one seeds and a two? How is CC equal to Yale?

Read the rest of the conversation, slappy.

 

[Stauber1]

Re: The 2014 Pairwise, Bracketology and History Thread`

Meh. Play who you have to, but I don't think the NCAA cares as much about attendance as some assume. Regionals are bid by cities and hosts which guarantee revenue to the NCAA. When revenues come in under the bid amount, the host is responsible for the difference; if they are over the bid amount, the host gets the bulk of the reward.

I have to disagree. Going forward, you have a brand new host in Fargo and a host in South Bend because no other sites bid in the west. That's directly tied to the attendance woes of most of the western regionals. If regionals continue to be attendance disasters in the "west" I think the NCAA will be forced to change the format, and I think there is a lot of resistance to that from most of the member schools.

 

[LTsatch]

Re: The 2014 Pairwise, Bracketology and History Thread`

Read the rest of the conversation, slappy.

I have read it Tipsy McHappy, never been more upsets in the tourney, in a row, by a 15 seed, than last year.

 

[Dutchman]

Re: The 2014 Pairwise, Bracketology and History Thread`

I have to disagree. Going forward, you have a brand new host in Fargo and a host in South Bend because no other sites bid in the west. That's directly tied to the attendance woes of most of the western regionals. If regionals continue to be attendance disasters in the "west" I think the NCAA will be forced to change the format, and I think there is a lot of resistance to that from most of the member schools.

Very good point. The East is geographically very condensed and its (IMO) only about a max 4 hr drive from New York, Buffalo or Boston to Albany. Why not one Regional in the Albany area in place of Worcester and Bridgeport and add another out west? Another idea is to scrap the Regionals and use the home barn of the #1 -#4 seeds, or if that is not large enough, a near by facility that is? Ticket sales can be shared and costs would be very low. Last year in Providence, IMO, Canisius, Union and especially QU fans showed up in force despite having the furthest to travel. BC fans did not despite all the talk about there being high speed public transportation, practically door to door, for them. (The BC band did show up, which was nice). Also, IMO, lots of college hockey fans in general not affilitaed with any of the schools and lots of alumni from the schools that lived in the area seemed to be on hand (Brown, Providence College, URI). Attendance should never be a factor in deciding which team plays where in the east. The schools are just too close together to care. Putting another Regional out West might help.

My biggest problem is that I don't believe that the NCAA has enought real data to understand the paying customer/viewer because it does not really care and this is a reflection that the NCAA is, as an entity, very very poorly managed. My other problem is that the NCAA massages the NCAA hockey tournament rules so much each year that there are no rules. IMO The fans (many alumni) and the players always come last and the $$ and school administrations always come first, with rare exception. College hockey is a sport that is growing in popularity and the NCAA should have an understanding of why and where and help those schools that want to understand its appeal.

 

[claver2010]

Re: The 2014 Pairwise, Bracketology and History Thread`

Or like when CC smoked BC?

If we've learned one thing, it's that BC clearly has struggled in the NCAAs against western teams of late.

As for attendance, they're pretty much disasters everywhere. From dumpster fires (Cincy, St. Louis, etc.) to not good (Worcester). The whole model needs to be blown up.

 

[Nick Papagiorgio]

Re: The 2014 Pairwise, Bracketology and History Thread`

If we've learned one thing, it's that BC clearly has struggled in the NCAAs against western teams of late.

That 15-3 record in the NCAAs against CCHA/WCHA/Air Force since 2006 really could be better!! How did we ever lose 3 games to those clowns?

 

[SJHovey]

Re: The 2014 Pairwise, Bracketology and History Thread`

How is CC equal to Yale?

The year after winning their only national championship, they were a bust and didn't even make the tournament?

 

[Happy]

Re: The 2014 Pairwise, Bracketology and History Thread`

That 15-3 record in the NCAAs against CCHA/WCHA/Air Force since 2006 really could be better!! How did we ever lose 3 games to those clowns?

If you ever had any good enough players to go pro early, I could understand those three loses. but, since you don't you guys just suck.

 

[SJHovey]

Re: The 2014 Pairwise, Bracketology and History Thread`

That 15-3 record in the NCAAs against CCHA/WCHA/Air Force since 2006 really could be better!! How did we ever lose 3 games to those clowns?

Had to play somewhere other than Worcester in the regional?

It's kind of funny, if you look at it. The road to the championship, since going to 16 teams, certainly favors a home regional.

2003 -- MN (Minneapolis regional)
2004 -- DU (Colorado Springs)
2005 -- DU (Amherst, MA)
2006 -- Wis (Green Bay)
2007 -- MSU (Grand Rapids)
2008 -- BC (Worcester)
2009 -- BU (Manchester)
2010 -- BC (Worcester)
2011 -- UMD (Bridgeport)
2012 -- BC (Worcester)
2013 -- Yale (Grand Rapids)

Kind of makes BU's win in 2009 more impressive, coming out of a regional where the host was also in the tournament.