Pool C

[Ralph Turner]

Highly inbred...ASC and SCAC!

Two of the lowest three.   SCAC -- Southwestern, Trinity, Colorado College, Austin College, Hendrix, Millsaps and Sewanee all played games versus the ASC schools.

+1 hugenerd!

Correction --

I mis-read the original post that hugenerd had included all of the conferences.  My subsequent review shows me that the ASC and SCAC are in the middle of all of the conferences in D3.

See my next assessment below.

[Mr. Ypsi]

Bravo, Hugenerd! +k

It may take me several days (at least) to assimilate and figure out how to interpret much of this data, but one datum jumped out at me: it should surprise no one that the outlier on WP/OWP was NESCAC.  Their single round-robin 'games' the system perfectly (though probably still not AS much as with the late, unlamented QOWI).

[pabegg]

Pool C update for Wednesday morning

Last night York PA said goodbye. None of the other losers last night had any chance at Pool C.

Occidental locked down their Pool C bid if they don't win the conference tournament, and Cal Lutheran guaranteed that they will be a Pool C contender if they don't. Webster kept their slim Pool C chances alive (probably a moot point, since they are perhaps the most prohibitive favorite in any conference tournament this weekend).

Tonight's tournaments with a Pool C impact:

In all cases, the Pool C candidate is listed first, except that some games now have two Pool C contenders

MAC Commonwealth SF
Widener vs. Albright (both in contention)

MAC Freedom SF
DeSales vs. Wilkes

MIAA QF
Hope vs. Alma
Albion vs. Kalamazoo

NCAC QF
Ohio Wesleyan vs. Hiram

OAC QF
Capital vs. Mount Union
Heidelberg vs. John Carroll

WIAC SF
Stevens Point vs. Eau Claire
Whitewater vs. Superior

Analysis

Widener, Hope, Capital, and Stevens Point would be Pool C locks with a loss tonight.
Heidelberg and Whitewater could probably absorb a loss tonight, but it would push them onto the bubble; on the other hand, a win locks up a Pool C bid.
For Albion, a loss would be nearly fatal (the UW-Platteville category)
Albright, DeSales, and Ohio Wesleyan would be done with a loss tonight.

Current Pool C contenders (no games left to play)

Bates
UW-Platteville

[bionbrit]

Pabegg- I really like this info

Quick question. If Albion was ranked over Heidelberg in the GL region last week, why would Heidelberg be in a class above Albion in your anaylsis. Do you think the berg will jump Albion in the rankings today, or does it have to do with John Carroll being a better loss/win then Kzoo?  Thanks for the info. Also if /when Albion wins tonight do they move into the can absorb a loss to Calvin and get in catergory on Friday?

[Titan Q]

1) OWP and OOWP standard deviations are generally very low for teams within their conference, meaning that teams, on average, have similar OWP within conferences (there are some exceptions).

Great work on this...thank you!

Here are the CCIW's OWP's...

Illinois Wesleyan 0.610
Wheaton: 0.577
Augustana: 0.575
Carthage 0.573
North Central: 0.573
North Park 0.535
Millikin: .515
Elmhurst: 0.505

The CCIW is a pretty "standard" league - 8 teams, a double round robin format, and geographically located in an area with plenty of non-conference, in-region games to play.  It seems clear to me the CCIW teams can influence their OWP significantly via their non-conference schedule - there is a big gap between Illinois Wesleyan (.610) and Elmhurst (.505).

Looking in more detail...

Illinois Wesleyan (in-region non-conf games)
Occidental  .812
Wash U  .800
Webster  .773
Chicago .696
Dominican .583
Claremont-Mudd-Scripps .556
Hanover .522
Illinois College  .381

Elmhurst (in-region non-conf games)
Hope  .882
UW-Oshkosh  .667
Anderson .542
Benedictine  .476
Tri-State  .389
Eureka .348
Mount St. Joseph  .273
Simpson  .261
Manchester  .143
MacMurray .056

[pabegg]

Pabegg- I really like this info

Quick question. If Albion was ranked over Heidelberg in the GL region last week, why would Heidelberg be in a class above Albion in your anaylsis. Do you think the berg will jump Albion in the rankings today, or does it have to do with John Carroll being a better loss/win then Kzoo?  Thanks for the info. Also if /when Albion wins tonight do they move into the can absorb a loss to Calvin and get in catergory on Friday?

'Berg beat Capital last week, so I expect them to move ahead of Albion. I had them essentially tied in my numbers through 2/17, but there's a notable gap in the 2/24 numbers. I will be surprised if the NCAA rankings that come out today don't reflect this.

I think Albion needs to win their first two games to stay on the good side of the bubble. Since they are only 14-3 in region, each additional loss hits them harder than others.

[HopeConvert]

I haven't carefully parsed hugenerd's data, but this conversation does raise a question for me that I'm confident one of you can answer. Let's take Hope's game against Alma tonight as an example. Given Alma's poor record, a Hope win tonight would actually lower Hope's OWP, although that might be offset somewhat by the OOWP (given that Alma's opponents records are inflated by having played Alma), even if the OOWP numbers are dissipated over a larger n. Then too, I would think the OWP would remain constant win or lose (that is, Alma's WP counts either way). I'm wondering, however, if it's possible that Hope's Pool C prospects would actually be improved by a loss to Alma rather than a win. In terms of their regional ranking, a win against Alma doesn't really help them that much - but how much would a loss hurt them? Given that Hope is a "lock" anyway, it might be a moot point, but is it unthinkable that there might be a team that's more borderline than Hope that could benefit more from a loss than a win? Just wondering...

[Flying Dutch Fan]

HC - I think I get your point, but you're leaving out the fact that the loss affects regional record, which is at least equal in consideration to OWP and OOWP. 

[HopeConvert]

Oh sure - that's why I asked how much a loss would hurt them. That's the calculation, and while I can't imagine a team intentionally losing a game, at the same time I wonder if there could be situations where the loss would affect the regional rankings less than a win would. I don't know the answer to that, although I suspect it would be highly unlikely. I suppose it goes to my suspicions about using OWP and OOWP. In part it's a two-edged sword: if the idea is to get teams to play a tougher non-conference schedule, since the conference schedule will zero out pretty much, then - assuming a tougher non-con would lead to more losses -  I'm not sure a team would actually be that motivated. I'm not sure there is a good answer to this dilemma, but neither am I sure that OWP and OOWP are the best solutions. Usually when organizers resort to such arcane formluae, it's because they're trying to substitute some faux objectivity for judgment. That may be because the logistics of judgment are too difficult, or because they're just trying to immunize themselves against criticism - but in either case, I don't see OWP and OOWP as being much of a measure.

[pabegg]

Let's be clear: Hope would end up with the same OOP/OOWP win or lose tonight.

The OOP is the average of the Opponent's Winning Percentage in the games not involving the team in question. So in the Hope-Alma case, Alma is 3-15 in region, with 2 losses to Hope, so they are 3-13 as far as Hope's OWP calculation is concerned (or .1875). So tonight's game will add another .1875 to Hope's total, reducing their OWP to .492 from .508.

So it's not a matter of winning or losing; Hope drops in this category simply by playing the game.

Of course, all of that is dwarfed by the potential of dropping the RWP from .882 to .833 if they lose the game (it goes up to .889 if they win).

This happened to Mass-Dartmouth last night as they beat their last place sister school Mass-Boston and dropped in the numbers.

D1 hockey has a provision for ignoring games like these, where the net impact on the RPI (a weighted average of RWP, OOP, and OOWP) is negative, then the game is ignored.

[Titan Q]

I haven't carefully parsed hugenerd's data, but this conversation does raise a question for me that I'm confident one of you can answer. Let's take Hope's game against Alma tonight as an example. Given Alma's poor record, a Hope win tonight would actually lower Hope's OWP, although that might be offset somewhat by the OOWP (given that Alma's opponents records are inflated by having played Alma), even if the OOWP numbers are dissipated over a larger n. Then too, I would think the OWP would remain constant win or lose (that is, Alma's WP counts either way). I'm wondering, however, if it's possible that Hope's Pool C prospects would actually be improved by a loss to Alma rather than a win. In terms of their regional ranking, a win against Alma doesn't really help them that much - but how much would a loss hurt them? Given that Hope is a "lock" anyway, it might be a moot point, but is it unthinkable that there might be a team that's more borderline than Hope that could benefit more from a loss than a win? Just wondering...

Anytime you start asking if it'd be better to lose than win, you're over-thinking!

[Hugenerd]

Like pabegg said, upping alma's win total by one will hardly effect Hope's overall OWP (since it will be 1 of 20^2 or so games, raising their OWP with a loss compared to a win potentially by about 0.002).  However, it would cause a much larger drop in their regional winning percentage, as pabegg said, by about 0.05.  There is really no way of controlling your OWP and OOWP, other than scheduling teams at the top of traditionally strong conferences.  Also, it will never be beneficial to lose a game, compared to winning a game, for your own case to make the NCAAs.  For example, take the case of Babson, they played a really difficult out of conference schedule and have the highest OWP in the country, but they have a losing record and have no chance at the tourney.

[smedindy]

I'm wondering, however, if it's possible that Hope's Pool C prospects would actually be improved by a loss to Alma rather than a win.

Must I get my Herm Edwards quote out again??

"You play to WIN the game!"

[Hugenerd]

Highly inbred...ASC and SCAC!

Two of the lowest three.   SCAC -- Southwestern, Trinity, Colorado College, Austin College, Hendrix, Millsaps and Sewanee all played games versus the ASC schools.

+1 hugenerd!

Correction --

I mis-read the original post that hugenerd had included all of the conferences.  My subsequent review shows me that the ASC and SCAC are in the middle of all of the conferences in D3.

See my next assessment below.

I think your assessment was correct.  The ASC and SCAC have two of the 3 lowest OWP HI/LO, which is the criteria I created to judge the extent of "inbreeding" in the schedule.  They both are in the middle of the pack for overall OWP, but that is for the reason we discussed a few days ago (nearly all their games are in conference or against common opponents say their OWP and OOWP are very close to 0.500).

[Ralph Turner]

1) OWP and OOWP standard deviations are generally very low for teams within their conference, meaning that teams, on average, have similar OWP within conferences (there are some exceptions).

Great work on this...thank you!

Here are the CCIW's OWP's...

Illinois Wesleyan 0.610
Wheaton: 0.577
Augustana: 0.575
Carthage 0.573
North Central: 0.573
North Park 0.535
Millikin: .515
Elmhurst: 0.505

The CCIW is a pretty "standard" league - 8 teams, a double round robin format, and geographically located in an area with plenty of non-conference, in-region games to play.  It seems clear to me the CCIW teams can influence their OWP significantly via their non-conference schedule - there is a big gap between Illinois Wesleyan (.610) and Elmhurst (.505).

Looking in more detail...

Illinois Wesleyan (in-region non-conf games)
Occidental  .812
Wash U  .800
Webster  .773
Chicago .696
Dominican .583
Claremont-Mudd-Scripps .556
Hanover .522
Illinois College  .381

Elmhurst (in-region non-conf games)
Hope  .882
UW-Oshkosh  .667
Anderson .542
Benedictine  .476
Tri-State  .389
Eureka .348
Mount St. Joseph  .273
Simpson  .261
Manchester  .143
MacMurray .056

Titan Q illustrates the geographic isolation of the SCAC and the ASC.

The SCAC plays a 16-game schedule.  They play double round robin inside each 6-team division and single round robin in the inter-division schedule.  (Because the NCAA does not count the games with first-year provisional team, Birmingham- Southern, in the in-region records, you only get 14 or 15 in-region games from the conference play.)

SCAC teams also usually play one ASC team as part of their in-region schedule as I cited above.  the ASC's OOWP is .5047; the SCAC's .4985.  I interpret the "OOWP HI/LO" of the ASC 1.0947 and the SCAC 1.0987 to be a measure of the geographic isolation.

(I interpret the Centennial Conference's OOWP HI/LO of 1.0885 to reflect "league parity" in a geographic area of robust statistical activity.)

Comments are appreciated.

hugenerd answered some of my concerns in his post just previous to this one.

hugenerd is really helping us understand what the numbers are telling us!  +1! Again!  Thanks!