PDA

View Full Version : 2012 NCAA Tournament: Bracketology



Pages : 1 2 3 [4] 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

JimDahl
01-23-2012, 09:21 AM
Hmmm there is a discrepancy between sites.

Not sure where the difference is...

Comparing the PWR Details for Minnesota (http://siouxsports.com/hockey/rankings/pwrdetails.php?teamid=7), USCHO seems to have a different RPI for Minnesota, such that Miami wins the comparison. Digging into the RPI details (http://siouxsports.com/hockey/rankings/rpidetails.php?teamid=7), it looks like USCHO isn't dropping the wins over Sacred Heart (see also USCHO RPI (http://www.uscho.com/rankings/rpi/d-i-men/) and CHN RPI (http://www.collegehockeynews.com/ratings/rpi.php)).

I don't know if they know something we don't or are messing up.

Bakunin
01-23-2012, 10:42 AM
Funny thing about that is the W hosting both western regions and probably would be favored to come out of them.
I wouldn't bet on MN to advance out of their region unless they figure out how to do something other than split every effing weekend.

Fighting Sioux 23
01-23-2012, 10:45 AM
I wouldn't bet on MN to advance out of their region unless they figure out how to do something other than split every effing weekend.

Agreed. Minnesota has been far too inconsistent over the past two months to be a legitimate threat for the National Title. Obviously, there are two months left for the Gophers to correct that. They have the tools to do so, and I think it is just a mental hurdle that they need to pass before they are a legit title contender.

JDUBBS1280
01-23-2012, 11:03 AM
Interesting fact. The Gophers have never won their conference the same year they won an NCAA national championship. They have always gotten hot at the right time at the end of the season. This Gopher team has the talent to win it all, they just need to be more consistent (as previously mentioned).

WCHA Records of the National Championship Teams
1974 14-9-2 2nd place
1976 18-13-1 3rd place
1979 20-11-1 2nd place
2002 18-7-3 3rd place
2003 15-6-7 2nd place

SJHovey
01-23-2012, 11:30 AM
I wouldn't bet on MN to advance out of their region unless they figure out how to do something other than split every effing weekend.With MN hosting, I see very little chance of them being upset before the Frozen Four.

Osorojo
01-23-2012, 12:03 PM
The mid-season rationales for predicting season's-end NCAA tournament brackets are based upon mid-season teams' records - ignoring the possibility that prejudice could play a part in such predictions. In the past five or ten years how close a correlation has there been between best performing mid-season teams and NCAA playoff seeds?

FlagDUDE08
01-23-2012, 12:08 PM
The mid-season rationales for predicting season's-end NCAA tournament brackets are based upon mid-season teams' records - ignoring the possibility that prejudice could play a part in such predictions. In the past five or ten years how close a correlation has there been between best performing mid-season teams and NCAA playoff seeds?

Obviously anything could happen, but given the system used to pick the teams is objective and chances of things happening can be predicted based upon the KRACH predictor, probably about 75% of the time you don't see a correlation is when there is an auto-bid involved.

Patman
01-23-2012, 12:20 PM
Obviously anything could happen, but given the system used to pick the teams is objective and chances of things happening can be predicted based upon the KRACH predictor, probably about 75% of the time you don't see a correlation is when there is an auto-bid involved.

More to the point... what do we have that could be better? Not much. Its not a perfect situation and if there are changes in chemistry etc that can't be handled in a model (especially with sparse data)... but, its often better than going on a straight hunch.

FlagDUDE08
01-23-2012, 12:30 PM
More to the point... what do we have that could be better? Not much. Its not a perfect situation and if there are changes in chemistry etc that can't be handled in a model (especially with sparse data)... but, its often better than going on a straight hunch.

And it certainly beats not being able to be mostly accurate because of committee votes (like some other sports out there ;)).

Osorojo
01-23-2012, 12:36 PM
Obviously anything could happen, but given the system used to pick the teams is objective and chances of things happening can be predicted based upon the KRACH predictor, probably about 75% of the time you don't see a correlation is when there is an auto-bid involved.

Thanks, I neglected to account for the auto-bid team(s)! Still, the chance of "things happening" - like the second half of the season - bemuse me when I see predictions made at or before the season's halfway over. I'm not sure I understand the accuracy of the KRACH predictor. Does it ( at midseason) correctly predict the NCAA tournament seeding of non-autobid teams only 25% of the time, or is it correct 75% of the time?

Priceless
01-23-2012, 12:48 PM
The mid-season rationales for predicting season's-end NCAA tournament brackets are based upon mid-season teams' records - ignoring the possibility that prejudice could play a part in such predictions. In the past five or ten years how close a correlation has there been between best performing mid-season teams and NCAA playoff seeds?

I have the pairwise data since 2003, taken at key points throughout the season, and mid-January teams in the Pairwise end up in the field 79.4% of the time. The top 8 seeds have made the tournament 64 of 72 times (88.9%) and no #1 or #2 in the mid-January rankings has ever failed to make the field. Of the top 4 overall seeds, only 2010 UMD (#3) and 2004 St Cloud (#4) have failed to make the field.

FlagDUDE08
01-23-2012, 12:51 PM
Thanks, I neglected to account for the auto-bid team(s)! Still, the chance of "things happening" - like the second half of the season - bemuse me when I see predictions made at or before the season's halfway over. I'm not sure I understand the accuracy of the KRACH predictor. Does it ( at midseason) correctly predict the NCAA tournament seeding of non-autobid teams only 25% of the time, or is it correct 75% of the time?

KRACH has nothing to do with tournament teams. The only thing KRACH does is to help predict, based upon past performances, who is the most likely to win a specific game. If you'd like information on tournament teams, PWR is the way to go. Obviously the stats change from game to game, but unless some team really goes to heck or a team around the bubble really starts to get hot, not much changes. I don't know what the percentages are, though. The "bubble" should be very large at this point since we still have two months worth of games to be played (playoffs DEFINITELY count toward the PWR), but it will get smaller the closer we get to the end of March.

Fighting Sioux 23
01-23-2012, 01:41 PM
College Hockey Weekly has their Bracketology up...

http://www.collegehockeyweekly.com/Current_Pairwise.html

Also, here is their forecast of what the tournament will look like...

http://www.collegehockeyweekly.com/Tournament-Forecast.html

JimDahl
01-23-2012, 01:53 PM
I have the pairwise data since 2003, taken at key points throughout the season, and mid-January teams in the Pairwise end up in the field 79.4% of the time. The top 8 seeds have made the tournament 64 of 72 times (88.9%) and no #1 or #2 in the mid-January rankings has ever failed to make the field. Of the top 4 overall seeds, only 2010 UMD (#3) and 2004 St Cloud (#4) have failed to make the field.

I wrote an article (When to start looking at PWR (http://blog.siouxsports.com/2009/12/31/when-to-start-looking-at-pwr/)) a couple years ago that tackled this from a different angle -- I looked at the average week-to-week movement in PWR and the average deviation of each week's ranking from the final rankings. The latter had this to say about the January PWR, "each team has been ranked an average of 3-8 slots differently on April 1 than in January over the past six years." That fits pretty well with your statistic that the top 8 in January almost always make it, and reminds us that there is still hope for those teams in the mid-20s.

ItalianSioux
01-23-2012, 05:20 PM
Also, here is their forecast of what the tournament will look like...

http://www.collegehockeyweekly.com/Tournament-Forecast.html

5 WCHA teams making it?!?!? :eek:

Snively65
01-23-2012, 05:40 PM
Wow, the east is getting pounded this season with only 3 HE and 1 ECAC teams in the rankings today and in the predicted post-season, along with AFA predicted for the AHA auto-seed. :(

Priceless
01-23-2012, 06:09 PM
Wow, the east is getting pounded this season with only 3 HE and 1 ECAC teams in the rankings today and in the predicted post-season, along with AFA predicted for the AHA auto-seed. :(

BU, Lowell, Merrihack, Boston College...I think that's four...

Fighting Sioux 23
01-23-2012, 06:31 PM
5 WCHA teams making it?!?!? :eek:

:p

I was a little surprised, although I know one guy who still thinks the Sioux will get a #2 seed, which I keep telling him is extremely unrealistic. Oh well...

FWIW, Here was my ballot (one of 8 that gets tallied up, 4 points for #1 seed, 3 for #2 seed...etc.)

#1 Seeds: Minnesota Duluth, Michigan, Boston University Notre Dame
#2 Seeds: Ohio State, Lowell, Boston College, Minnesota
#3 Seeds: Miami, Cornell, Ferris State, Denver
#4 Seeds: Western Michigan, Union, Colorado College, Air Force (AHA Champ)

Fighting Sioux 23
01-23-2012, 06:32 PM
BU, Lowell, Merrihack, Boston College...I think that's four...

I think they are referring to CHW's Tourney Forecast. Merrimack is not slated to make the tournament in their forecast.

mookie1995
01-23-2012, 06:34 PM
College Hockey Weekly has their Bracketology up...

http://www.collegehockeyweekly.com/Current_Pairwise.html

Also, here is their forecast of what the tournament will look like...

http://www.collegehockeyweekly.com/Tournament-Forecast.html

first week they did this, they had bu falling from #5 to OUT!
now they have bu staying in and retaining their #1 seed in worcester.

what changed?!?!?!?! :p