The 2014 Pairwise, Bracketology and History Thread`

[Numbers]

You are correct. Your winning percentage and opponents records (SOS) are what count, the exact breakdown of your own wins/losses is irrelevant in RPI.

This is debatable. The currently accepted way to do the calculation means that your opponents contribution to you're RPI in each game is weighted according to home/away won/loss. Thus, losing at home to to a strong team is preferable to losing urge road to a weak team.

[ExileOnDaytonStreet]

Re: The 2014 Pairwise, Bracketology and History Thread`

Thanks, blue. Also, my thought is that there is no penalty for losing. So,a schedule weighted with lots of strong opponents has no downside.

I had thought that there were also additional weight given to games where you win on the road and lose at home. So, if you schedule tough opponents to come to your rink, you run the risk of an added bonus weight for the game due to losing at home. Maybe someone can correct me here, but wasn't that another change to the RPI formula that would come into effect this year?

[MaizeRage]

Re: The 2014 Pairwise, Bracketology and History Thread`

Nice. I do think, however, that given the choice, the committee would put either Michigan or Notre Dame in Cincinnati. Other wise, too much potential for an almost empty arena. Do you think otherwise, and if so, why?

Notre Dame didn't help attendance at all in Toledo, and North Dakota didn't help attendance at all in Grand Rapids last year. Neither would do much for Cincinnati. I suspect Michigan wouldn't help much either, if they even made the tournament, which...

More likely is the committee accepts that Cincinnati will be empty regardless, and do as much as possible to help the other three regions.

[Aerman]

Re: The 2014 Pairwise, Bracketology and History Thread`

This is debatable. The currently accepted way to do the calculation means that your opponents contribution to you're RPI in each game is weighted according to home/away won/loss. Thus, losing at home to to a strong team is preferable to losing urge road to a weak team.

Wow, really? That makes no sense to me, if I understand what you are saying. Weighting my wins and losses as a function of where they occur has some obvious merit. But carrying the weighting into the pieces that constitute SOS seems illogical. If I extend what I think you are saying, my "winning-percentage" portion of RPI benefits more from beating Minny at their house than from losing there (duh), AND my SOS benefits more? Is that in fact what you are saying, and if so, how is that logical? I played the same team in the same place.

[Numbers]

As far as I understand, wins losses are weighted 0.8/1.2. Meaning for your own win percentage, a road win or home loss counts 1.2.

There was a big discussion about the SOS part, which seemed to be solved recently, that the SOS party is also weighted. But I am not sure the details of that. It may be as simple as...
If the game is at home, SOS counts 0.8. On the road, 1.2.

Or, it gets weighted according to whether you Erin ours list as well.

Where is Jim Dahl?

[cole_cutz]

Re: The 2014 Pairwise, Bracketology and History Thread`

Notre Dame didn't help attendance at all in Toledo, and North Dakota didn't help attendance at all in Grand Rapids last year. Neither would do much for Cincinnati. I suspect Michigan wouldn't help much either, if they even made the tournament, which...

More likely is the committee accepts that Cincinnati will be empty regardless, and do as much as possible to help the other three regions.

Doesn't matter with Notre Dame they won't come close to making it. They won't get a point at BC this weekend which will probably leave them in 8th place in HE hosting and probably beating UMass for a game then on the road for two quick losses at one of Maine, UML or NU. May as well give out freebie for the games in Cincy. Especially if the UC hoops team is still in NCAA's.

[Stauber1]

Re: The 2014 Pairwise, Bracketology and History Thread`

Doesn't matter with Notre Dame they won't come close to making it. They won't get a point at BC this weekend which will probably leave them in 8th place in HE hosting and probably beating UMass for a game then on the road for two quick losses at one of Maine, UML or NU. May as well give out freebie for the games in Cincy. Especially if the UC hoops team is still in NCAA's.

If you could 3 of Ferris, Wisco, Notre Dame and Michigan to Cinci it might not be a total waste. Tickets are very reasonably priced for that one regional (about $40 for both days) so if they can draw a couple thousand traveling fans, it just might avoid being an attendance disaster.

[Steve_MN]

Re: The 2014 Pairwise, Bracketology and History Thread`

This is debatable. The currently accepted way to do the calculation means that your opponents contribution to you're RPI in each game is weighted according to home/away won/loss. Thus, losing at home to to a strong team is preferable to losing urge road to a weak team.

Good point, and that's a change this year that I hadn't fully thought through. Historically, I'd stand by my statement, but you're absolutely correct that this has changed.

Still thinking through whether the last part of your statement makes sense, though. If the home loss is weighted 1.2, wouldn't you rather lose to the weaker team on the road with the .8 weighting?

I'd still think that the breakdown of who you beat or lose to wouldn't be as important as where the games are played. I wonder how much of an effect this is having on keeping Wisconsin in relatively good PWR position, with their terrible road record... losing on the road only counts .8 in the RPI, but with almost all their wins being at home, those only count .8, that would seem to offset.

[ExileOnDaytonStreet]

Re: The 2014 Pairwise, Bracketology and History Thread`

I'd still think that the breakdown of who you beat or lose to wouldn't be as important as where the games are played. I wonder how much of an effect this is having on keeping Wisconsin in relatively good PWR position, with their terrible road record... losing on the road only counts .8 in the RPI, but with almost all their wins being at home, those only count .8, that would seem to offset.

They would offset with a .500 record, where the home and road results were inverses of each other.

In UW's case (someone could double check for me, but I think I'm right), the records are as follows:

Overall: 19-9-2 (.667)
Home: 17-2-1 (.875)
Away: 2-7-1 (.250)

The two home losses and the two away wins would cancel each other out, but there are still ten more home wins than away losses, so that's still a net benefit in the overall RPI calculation because Wisconsin is (overall) a winning team.

The odd thing is that, IMO, you're wrong when you say "keeping Wisconsin in a relatively good PWR position". Really, if anything the weighting system is keeping UW at a lower RPI than they could be. That difference of having 10 more home wins than road losses is calculated at a 0.8 weighting, instead of the equal weighting it used to have before the home/away bonus was introduced.

[Steve_MN]

Re: The 2014 Pairwise, Bracketology and History Thread`

They would offset with a .500 record, where the home and road results were inverses of each other.

In UW's case (someone could double check for me, but I think I'm right), the records are as follows:

Overall: 19-9-2 (.667)
Home: 17-2-1 (.875)
Away: 2-7-1 (.250)

The two home losses and the two away wins would cancel each other out, but there are still ten more home wins than away losses, so that's still a net benefit in the overall RPI calculation because Wisconsin is (overall) a winning team.

The odd thing is that, IMO, you're wrong when you say "keeping Wisconsin in a relatively good PWR position". Really, if anything the weighting system is keeping UW at a lower RPI than they could be. That difference of having 10 more home wins than road losses is calculated at a 0.8 weighting, instead of the equal weighting it used to have

To be clear, I really don't know what effect it's having. I'd love to see a version of the PWR without the weighting at all (but with the bonus pts and no TUCliff) to see how much, if any, difference there is.

And you're definitely right that it's certainly possible that it's holding Wisconsin back instead of propping them up. My best guess is that the the effect is very minor. Taking a closer look (and putting the knowledge of their road record out of my mind that raised the question in the first palce) I see that Wisconsin is currently 6th in RPI/PWR, 8th in Win Percentage and 7th in KRACH, so, realistically, any RPI/PWR in the 6-8 range would seem to be the proper place, regardless of weighting, etc.

[QUAlum2004]

Re: The 2014 Pairwise, Bracketology and History Thread`

This week’s brackets
West Regional (St. Paul):
13 Colgate vs. 2 Minnesota
10 North Dakota vs. 6 Wisconsin

Midwest Regional (Cincinnati):
14 Michigan vs. 4 Ferris State
12 Vermont vs. 5 St. Cloud State

Northeast Regional (Worcester):
16 Mercyhurst vs. 1 Boston College
11 Cornell vs. 8 Massachusetts-Lowell

East Regional (Bridgeport):
15 Notre Dame vs. 3 Union
9 Northeastern vs. 7 Quinnipiac

This week's USCHO.com brackets. My thoughts are the West Regional is a brutal region with Minnesota, Wisconsin, and two teams gaining steam in North Dakota and Colgate. BC has the easiest region in the East. Northeast regional is a toss up. I think any of those 4 teams could win it. Midwest bracket is not bad either. Likely a Ferris vs. St. Cloud final as Michigan is fading quickly.

http://www.uscho.com/bracketology/20...west-regional/

[mnstate0fhockey]

This week's USCHO.com brackets. My thoughts are the West Regional is a brutal region with Minnesota, Wisconsin, and two teams gaining steam in North Dakota and Colgate. BC has the easiest region in the East. Northeast regional is a toss up. I think any of those 4 teams could win it. Midwest bracket is not bad either. Likely a Ferris vs. St. Cloud final as Michigan is fading quickly.

http://www.uscho.com/bracketology/20...west-regional/

Switch St. Cloud and Wisconsin. Both would be good draws in St. Paul, but I think Wisconsin would draw better than St. Cloud in Cincy.

[johnk]

Re: The 2014 Pairwise, Bracketology and History Thread`

Switch St. Cloud and Wisconsin. Both would be good draws in St. Paul, but I think Wisconsin would draw better than St. Cloud in Cincy.

Cant if you want to avoid interconference matchup.

[mnstate0fhockey]

Cant if you want to avoid interconference matchup.

Ah yes, forgot about that. I'd say move UND, but no way the NCAA does that. Can't say I blame them.

[BBfan13]

Re: The 2014 Pairwise, Bracketology and History Thread`

No Maine? even with a sweep and vermont losing this weekend