2005-2006 PWR

[cth95]

Just found this. http://www.uscho.com/news/id,11635/PairWiseSurprise.html

You Might Think They're NCAA-Bound, But The PairWise Disagrees
No. 15 (tie) Cornell: Simply put, the Big Red haven't played anybody yet. Cornell has a 2-1-1 record against TUCs, and while the winning percentage there (.6250) is good, the problem is that there are only four games in there. That's right — Cornell has only played four of its 17 games this season against teams with a .500 RPI or better (two against Michigan State, one against Harvard, and one against Rensselaer). A quick look at the Big Red's schedule shows why: Cornell's ECACHL schedule is ridiculously back-loaded. The Big Red's remaining opponents include Clarkson (twice), St. Lawrence (twice), Colgate (twice), Harvard and Rensselaer, all currently TUCs. A reasonable record in those games would enhance the Big Red's chances for an at-large bid if the team doesn't win the ECACHL tournament.

[Al DeFlorio]

[quote cth95]A reasonable record in those games would enhance the Big Red's chances for an at-large bid if the team doesn't win the ECACHL tournament.[/quote]
And an "unreasonable" record in those games would mean we deserve to stay home.

[KeithK]

Nothing new or surprising in that quote.  We haven't played the tough part of our conference schedule yet and as Al says if we don't do well in that part we won't deserve to play for the national championship.

Gates ranking surprised me a bit - I guess I hadn't payed close attention to who they've played and beaten.  Also MSU at #5 is a shock, since for a while there it looked like they were playing well below expectations.  In both cases the PWR rank isn't too far away from the KRACH rank, so it makes sense.

Lot's of hockey left to be played.  A timely ten game win streak would do wonders for our tournament chances.  (Especially a ten game streak starting in March...)

[jtwcornell91]

How does the number of games vs TUC affect our PWR, except through the strength of schedule part of the RPI? This isn't lacrosse where SoS is the main criterion.  I guess they're trying to say our RPI is worse than just looking at our record would lead you to believe, but number of games vs TUC is not a criterion.

[Jacob '06]

[quote jtwcornell91]How does the number of games vs TUC affect our PWR, except through the strength of schedule part of the RPI? This isn't lacrosse where SoS is the main criterion.  I guess they're trying to say our RPI is worse than just looking at our record would lead you to believe, but number of games vs TUC is not a criterion.[/quote]

Theres also the winning percentage against TUCs point in the PWR. Looking at the comparisons with all the people above us, we lose TUC and RPI to all of them.

[KeithK]

But John's point is that it doesn't matter whether we are 2-1-1 against TUCs or 8-4-4.  The TUC criteria would be the same either way.  Our RPI would likely be a lot higher though.

[Jacob '06]

[quote KeithK]But John's point is that it doesn't matter whether we are 2-1-1 against TUCs or 8-4-4.  The TUC criteria would be the same either way.  Our RPI would likely be a lot higher though.[/quote]

Yeah, I understood that.

[TCHL8842]

Well I think 2 wins against SLU will really help, they are currently number 3 in the PWR ranking.  Winning against SLU alone will significantly boost our RPI.  Of the ECAC teams with a better RPI then us, it is only Harvard and SLU.  For some reason, I think the 2 weekends against the North Country and the Harvard game will determine if we make it in or not.  If we sweep those 5 games, I think we could be maybe a number 2 seed.  The losses to Princeton and Union are going to kill us if we are a bubble team for making the tourney this year.  Like everyone said there is still a lot of hockey to be played, hopefully we can win the same percentage of the games as we did in the first half.

We currently lose 10 PWR comparisons by having the lower RPI, if this changes for the better which it should based on who left we have left, we can gain back quite a few comparisons.

[KeithK]

For the record, the PWR rankings cited in the article do not all match USCHO's PWR page.  For example, on the ranking page SLU is listed as #6 (tie), not #3.  I suspect the article was written a few days ago and is a little out of date.

[KeithK]

I thikn it's also relevant to point out that if we had played more games against TUCs and had the same record (overall and TUC pct.), our PWR standing might be better due to H2H effects.  The hypothetical extra TUC wins could swing individual comparisons that we currently tie or lose. Not that I think the author is making this point - he's just making the easy SoS argument.

[TCHL8842]

Well the tie breaker on even comparisons is RPI and currently we are currently 21 in this ranking.  With a better RPI even though I have not looked at it closely I think we would take another 5-6 comparisons.  This weekend is just as big of a weekend as the MSU weekend.  If we play as well as we seem them play in stretches, our spot in the PWR ranking will be significantly higher.

[oceanst41]

I think the article already factored in the 3-2-1 bonus

That would make SLU the third ranked PWR team

[Chris 02]

I've noticed that the PWR published on www.siouxsports.com matches exactly with the USCHO one, but the www.rpihockey.net one seems to have some differences.

[Ken \'70]

[quote TCHL8842]Well I think 2 wins against SLU will really help, they are currently number 3 in the PWR ranking.  Winning against SLU alone will significantly boost our RPI. [/quote]

Winning against SLU vs someone else doesn't matter, that's not how the mechanics of RPI work.  If you're 20 - 10 at the end of the year it doesn't matter how those wins and losses are distributed.  Against the same set of opponents you can flip the 10 losses to 10 entirely different teams and still have the same RPI.

[nshapiro]

[Q]
Winning against SLU vs someone else doesn't matter, that's not how the mechanics of RPI work. If you're 20 - 10 at the end of the year it doesn't matter how those wins and losses are distributed. Against the same set of opponents you can flip the 10 losses to 10 entirely different teams and still have the same RPI.
[/Q]

If the RPI is calculated by averaging the winning percentage of your opponents, as stated in the USCHO explanation, then it does matter slightly.

If every team played the same number of games, then who you lose to wouldn't matter.  Since the Ivies play fewer games, the optimal strategy would be to have the 10 losses be against the Ivy teams, since beating them would have more of an effect on an Ivy team's winning percentage.