Bracketology and the PairWise Rankings II

[BC/HE]

Re: Bracketology and the PairWise Rankings II

Nope. Again, read the history. Especially the Maine/UMass/SLU one mentioned. And the link in my previous post.

Yes, it's being double counted - yes, it puts too much weight on the RPI - but yes, that's how they do it.

Whether they should or shouldn't - that's another story. They didn't used to.

The USCHO listing is not wrong, per se - because, like I said, it's not written in stone. But in practice, it is wrong. CHN started listing it differently (rightly?) after that UMass/Maine/SLU thingy.

Like said before it is nowhere in the selection manual. The RPI tiebreaker is only quoted to break ties in the comparison count. The fact that it was done one way before is not a requirement to do it again. The selection committee can decide to do it anyway they want. Precedence may or may not be a factor. I think they like to have that leeway to tweak things a bit. It is all that they are given the chance to do so I don't think the practice they followed before will definately be the outcome. Therefore when teams are tied we should accept the fact that it could go either way and not say that a team is locked in one way or the other.

[Patman]

Re: Bracketology and the PairWise Rankings II

Like said before it is nowhere in the selection manual. The RPI tiebreaker is only quoted to break ties in the comparison count. The fact that it was done one way before is not a requirement to do it again. The selection committee can decide to do it anyway they want. Precedence may or may not be a factor. I think they like to have that leeway to tweak things a bit. It is all that they are given the chance to do so I don't think the practice they followed before will definately be the outcome. Therefore when teams are tied we should accept the fact that it could go either way and not say that a team is locked in one way or the other.

The RPI has always and consistently broken all ties. Ties in pairs... ties in Pairwise standings. Could it change? Possibly. But this is as it is known and as it is understood.

[Priceless]

Re: Bracketology and the PairWise Rankings II

Over on the other site, Adam has done a similar breakdown, but he says, "Obviously, if Vermont loses, it is done." Not true. There are numerous possibilities for the Catamounts even with a loss to BC on Friday. Logically, it would seem a loss by the 14th team in the PWR would knock them down, but their non-conference record keeps them afloat.

Earlier I posted that Hockey East could get one team in if BC won the autobid. If the BU/Maine winner gets the autobid, HEA could get 4 teams in the dance. That's how fluid the situation is with just a weekend left.

[UML]

Re: Bracketology and the PairWise Rankings II

Just wondering what the chances are for St. Cloud playing in one of the eastern regionals?

[Patman]

Re: Bracketology and the PairWise Rankings II

Over on the other site, Adam has done a similar breakdown, but he says, "Obviously, if Vermont loses, it is done." Not true. There are numerous possibilities for the Catamounts even with a loss to BC on Friday. Logically, it would seem a loss by the 14th team in the PWR would knock them down, but their non-conference record keeps them afloat.

Earlier I posted that Hockey East could get one team in if BC won the autobid. If the BU/Maine winner gets the autobid, HEA could get 4 teams in the dance. That's how fluid the situation is with just a weekend left.

i haven't been looking in depth at all... are you playing out everybody else as well? Again, I wish I had my simulator written... I know I did it last year... despite flaws... but admittedly I had more time.

[Wol4ine]

Re: Bracketology and the PairWise Rankings II

like i said... i don't claim to be the authority...

do you have the final rpi for that year? all i could find was the 3/17/07 on slack.net which had the following (not sure if it is final)

STL - .5384
Massachusetts - .5328
Maine - .5306

if those are in fact, the season ending rpis, i wonder if they took liberties in seeding in this instance, as all three teams were in the field... if they were all tied for the last spot in the tournament i wonder if it would have played out differently

These were the final RPI going inot selection sunday;

12 St Lawrence (EC) 13 0.5384
13 Maine (HE) 13 0.5366
11 Mass-Amherst (HE) 13 0.5328

All were tied for 11th with 13 comparisons, I beleive UMass won the comp's over both Maine and StLaw, but ended up with the lowest RPI after their loss to UNH in the HE semis.

Most of us put UMass at 11 because they won the comp's with the other 2, and then StL because they won the comp with Maine.

[Farce Poobah]

Re: Bracketology and the PairWise Rankings II

Just wondering what the chances are for St. Cloud playing in one of the eastern regionals?

Can you define "playing"? Hint: this is St Cloud you're asking about.

OK, but seriously, given they are in the range from #3 to about #8, seems like they are about 50-50. Unless the NC$$ needs butts in St Paul, in which case the odds go up.

[UML]

Re: Bracketology and the PairWise Rankings II

my cousin went to clood so since UML is out I told him I'd go with him if it was in the east.

[Priceless]

Re: Bracketology and the PairWise Rankings II

i haven't been looking in depth at all... are you playing out everybody else as well? Again, I wish I had my simulator written... I know I did it last year... despite flaws... but admittedly I had more time.

I'm using the PWR Predictor.

[Red Cloud]

Re: Bracketology and the PairWise Rankings II

Just wondering what the chances are for St. Cloud playing in one of the eastern regionals?

I'd say their chances are pretty decent, without getting too much involved in the nitty gritty (I have a regular Husky Bracketology column at DaHuskies).

[ScoobyFanClub]

Re: Bracketology and the PairWise Rankings II

The RPI has always and consistently broken all ties. Ties in pairs... ties in Pairwise standings. Could it change? Possibly. But this is as it is known and as it is understood.

Not true. It wasn't like that prior to 2003.

[brianvf]

Re: Bracketology and the PairWise Rankings II

Stolen from PWR guru Jim over on ss.com, he ran scenarios to figure out the potential max and min PWR for the teams:

Denver 1-3
Miami 1-3
UW 1-5
UND 3-6
SCSU 4-11
Boston College 3-9
UMD 5-19
Northern Michigan 6-19
Ferris State 5-18
CC 18-25
New Hampshire 10-18
UMN 20-25
Bemidji State 5-11
Alaska 10-18
Vermont 6-19
Michigan 10-25
Michigan State 12-21
Nebraska-Omaha 18-24
Cornell 5-19
Maine 11-25
Boston University 12-25
Mass.-Lowell 18-25
Ohio State 24-N/TUC
Yale 7-13
Northeastern 25-N/TUC
Massachusetts 16-20 or N/TUC
Union 11-N/TUC
St. Lawrence 23-N/TUC
RIT 20-N/TUC

[yalehockeyfan]

Re: Bracketology and the PairWise Rankings II

Stolen from PWR guru Jim over on ss.com, he ran scenarios to figure out the potential max and min PWR for the teams:

Given that they won't play another game before the tournament, and thus don't control their own destiny, am quite please with the position Yale is in.
Every team with the exception of Denver, Miami, UW, UND, SCSU, CB, and Bemidji has a lower floor to fall to than Yale. Looks like they will most likely be in the tournament as a 3 seed. Now if we just had our best player for the tournament I would like our chances a lot more

[BUPhD]

Re: Bracketology and the PairWise Rankings II

How does he get Maine to an 11 and BU to a 12 when no one can seem to replicate such a thing on the PWR predictor?

I smell math shenanigans here...

[brianvf]

Re: Bracketology and the PairWise Rankings II

How does he get Maine to an 11 and BU to a 12 when no one can seem to replicate such a thing on the PWR predictor?

I smell math shenanigans here...

He's smart.