Pairwise and Bracketology 2013 Edition

[aweise]

Updated bracketology:

Manchester (UNH)	Providence (Brown)	Toledo (BGSU)		Grand Rapids (Michigan)
New Hamp		Quinnipiac		Miami			Minnesota
Boston C		Mankato			No Dakota		W Michigan
Denver			Yale			Niagara			St Cloud
Notre Dame		Merrimack		Boston U		Lowell

Hockey East gets five teams in, so one could go to Manchester (should be BU) but Notre Dame can fill in for them (and can't go to Toledo). Lowell and BU remain interchangeable - Miami v Lowell is the natural 3-14 pairing but Minnesota is the overall #2 and should get the lower-ranked of Lowell and BU. An argument can be made either way. W Michigan and Mankato aren't placed perfectly but that will hopefully help attendance. Otherwise the WMU v SCSU and Mankato v Yale pairings can be switched.

Curious...why take Merrimack over Union, when Union is ahead of them in the PWR and actually wins the comparison vs. Merrimack?

[ericredaxe]

Curious...why take Merrimack over Union, when Union is ahead of them in the PWR and actually wins the comparison vs. Merrimack?

Merrimack in 1st in HE so I'd assume he figures they he the auto bid

[QUAlum2004]

Merrimack in 1st in HE so I'd assume he figures they he the auto bid

Yes but are they the favorite to win the HE tourney? I'd prolly say no st this point.

[sterlippo1]

Re: Pairwise and Bracketology 2013 Edition

Yes but are they the favorite to win the HE tourney? I'd prolly say no st this point.

no, prolly not (but you wouldn't want to play them right now, i can tell you that)................... , but bracketology is for now, this point in time

[Ralph Baer]

Re: Pairwise and Bracketology 2013 Edition

no, prolly not (but you wouldn't want to play them right now, i can tell you that)................... , but bracketology is for now, this point in time

Using the top team in the AHA in years when none are in line to get an autobid has always made sense to me. Using it in a league that is already getting four bids strikes me as quite odd. OTOH, the whole process of predicting the field in advance each time some games are played is an exercise and just shows how fluid things are.

[FlagDUDE08]

Re: Pairwise and Bracketology 2013 Edition

Curious...why take Merrimack over Union, when Union is ahead of them in the PWR and actually wins the comparison vs. Merrimack?

Union is also higher than Merrimack in RatingsPI.

My guess is Hockey Rast Bias.

[sterlippo1]

Re: Pairwise and Bracketology 2013 Edition

Union is also higher than Merrimack in RatingsPI.

My guess is Hockey Rast Bias.

yeah, what i meant was it (bracketology) doesn't consider what may or may not happen after today as he mentioned MC may not even be the favorite to win HE............though i wouldn't bet against them at this point

[aweise]

yeah, what i meant was it (bracketology) doesn't consider what may or may not happen after today as he mentioned MC may not even be the favorite to win HE............though i wouldn't bet against them at this point

And that's the most important point...it's the selections AS OF TODAY. Take the current conference leader and assume they're the conference tourney champion because we don't have the tourney results. We can't assume someone other than Merrimack would be the HE champs, regardless of what we think will happen.

[kingdobbs]

Re: Pairwise and Bracketology 2013 Edition

Curious...why take Merrimack over Union, when Union is ahead of them in the PWR and actually wins the comparison vs. Merrimack?

Because currently, Merrimack is leading Hockey East, both in points accumulated and in in-conference winning percentage.

Bracketology has always been an informal simulation of the tournament selection process, operating under the assumption that there are no more games remaining to play, not even those in conference tournaments. In the absence of such, bracketologists have to choose based on another method, and the next-most-logical (and the one that requires zero computations of probability) is "current regular season standings winner". Bracketology could use "highest ranking in the PWR or RPI" to determine this, but the fact that it's a simulation of the process itself leads bracketologists to use a method which would be used by conferences in such a process*.

Ordinarily, with the usual exception of Atlantic Hockey**, the team leading a conference in the standings is usually ranked in the Top 16 (i.e. the "cut-off" line), and so this question would be moot.

*And by NCAA regulations, there's really only two ways a conference can select its automatic participant: highest ranked at the end of round-robin conference play, or a knockout tournament. Technically, there's also a provision for an "in case of emergency" if a tournament can't be conducted in time for the tournament selection.

**Niagara's performance this year being, of course, against the trend.

[Numbers]

Re: Pairwise and Bracketology 2013 Edition

Priceless,

It seems like slack.net is back up, in the condition it was in before it went down. No results posted since Jan 12. Is that what you are seeing, too?

[RENCEB]

Re: Pairwise and Bracketology 2013 Edition

Because currently, Merrimack is leading Hockey East, both in points accumulated and in in-conference winning percentage.

Bracketology has always been an informal simulation of the tournament selection process, operating under the assumption that there are no more games remaining to play, not even those in conference tournaments. In the absence of such, bracketologists have to choose based on another method, and the next-most-logical (and the one that requires zero computations of probability) is "current regular season standings winner". Bracketology could use "highest ranking in the PWR or RPI" to determine this, but the fact that it's a simulation of the process itself leads bracketologists to use a method which would be used by conferences in such a process*.

Ordinarily, with the usual exception of Atlantic Hockey**, the team leading a conference in the standings is usually ranked in the Top 16 (i.e. the "cut-off" line), and so this question would be moot.

*And by NCAA regulations, there's really only two ways a conference can select its automatic participant: highest ranked at the end of round-robin conference play, or a knockout tournament. Technically, there's also a provision for an "in case of emergency" if a tournament can't be conducted in time for the tournament selection.

**Niagara's performance this year being, of course, against the trend.

Excellent explanation! Absent the 1st sentence, this post should be inserted every 25 posts or so as a reminder of the approach to determining the field. Would love to see MC make the tournament, auto-bid or otherwise (and by otherwise, I mean by improving their pairwise postion to a point where they are safely "in" and not in danger of being knocked out by some other conference auto-bid).

[Jim]

Re: Pairwise and Bracketology 2013 Edition

I have two questions about the field. 1. Is it theoretically or practically possible that AHA could get 2 bids if Niagara loses in the AHA finals? 2. If that were to somhow happen, would the NCCA change the rules again to prevent it from ever happening again?

[Priceless]

Re: Pairwise and Bracketology 2013 Edition

I have two questions about the field. 1. Is it theoretically or practically possible that AHA could get 2 bids if Niagara loses in the AHA finals? 2. If that were to somhow happen, would the NCCA change the rules again to prevent it from ever happening again?

Very possible.

No, the NCAA isn't going to change the rule. This has happened before (2010, Bemidji State) and the NCAA didn't act then.

[Numbers]

Re: Pairwise and Bracketology 2013 Edition

I have two questions about the field. 1. Is it theoretically or practically possible that AHA could get 2 bids if Niagara loses in the AHA finals? 2. If that were to somhow happen, would the NCCA change the rules again to prevent it from ever happening again?

To answer both questions:

1) Yes, it is very possible. Niagara's PWR ranking is almost totally dependent on their RPI. If they were to, let's say, win all their games from now until the AHA finals, and then lose, their RPI might well hang right where it is right now. I believe that would be a 7-1 record, which is better than their current win %age, so their RPI couldn't suffer too bad. Then, whoever beat them in the finals would also qualify.

2) There would be no reason to change the rules. The rules are that each conference has one auto-bid they can give out however they want. AHA does that by their tourney. So do all the other conferences. That makes it always possible that someone gets hot and wins the tourney, even though their regular season may have been less than stellar.

2a) In this case, the likely candidate is Robert Morris, who is almost a TUC or on the bubble anyway. I am just guessing here, but if they won the auto bid, their PWR rank might end up somewhere around 20 - 22. I am sure the NCAA would have no problem with that.

Sidenote: There was a situation a few years ago where the definition of TUC was: All teams with RPI at or above .500 + all teams who won their conference tourney. That was a problem that one year, because the regular season champ from AHA actually would have benefitted in the PWR by intentionally losing the tourney to Bentley. Bentley had a low RPI, and Holy Cross (can someone verify that???) played them 7 times or something like that, winning them all. If they had intentionally lost in the tourney, that would have been 6-1 or 7-1, and Bentley, by winning the tourney, would have been TUC. That would have raised Holy Cross' TUC record by 6-1, and vaulted them over about 4 or 5 other teams in the PWR.
Thankfully, HC won the game. And, the NCAA changed the rule about who is a TUC to only include at or above .500 RPIs, regardless of conference championships, after that happened.

Sorry for the history lesson....

[Numbers]

Re: Pairwise and Bracketology 2013 Edition

Very possible.

No, the NCAA isn't going to change the rule. This has happened before (2010, Bemidji State) and the NCAA didn't act then.

Priceless, did you see that slack.net seems to be back????