Spring 2011 PWR discussion

[jimdahl]

As has been discussed elsewhere, the Sioux are in the rare position of being in good shape for a tournament bid before New Years. I hope that means we're looking forward to a Spring of watching the Sioux scrap for a #1 seed, but either way there will be rankings to watch...

UND individual comparisons

...and I found myself wondering about this #2 vs #3 game, so thought it a good time to start a thread.

Though the Bulldogs are currently a spot above UND (PWR), UND is actually winning the comparison by virtue of the RPI tie-breaker. The H2H will clearly take over after tomorrow's game, making it a must-win for UND to remain #3 going into the new year.

As to the possibility of UND climbing if they win, here are the 3 comparisons UND is currently losing--

Boston College has TUC (.6364 to .5769) and COP (.8 to .4167). A win this weekend should only make up about half the difference on TUC, not enough to flip it.

Maine has H2H. No opportunity to flip this one until TUC or COP come into play.

Yale has RPI. UND's last win over UMD contributed something like .696, so it seems like UND would only rise to something like .586 with a win, not enough to take RPI.

Bottom line -- a loss almost certainly will lead to a drop, a win at best maintains the current ranking (all other things equal).

[Big A HG]

Hey Jim, better be careful, you'll get ripped on for talking about Pairwise too early.

Good stuff, like always.

[Big A HG]

That Maine series is going to be a big thorn in our butts to try and maintain a #1 seed. Top Hockey East teams like BC and UNH are going to continue to hold the Common Opponents point against us because of it. If we maintain a good RPI, that gives us a point over each of them, but like Jim said, the TUC percentage will be the deciding factor against both teams. UND must play at a high level against all TUC's from here on in. If so, we'll be a #1 seed. If we slip up, we'll be a #2 seed at best.

[808287]

That Maine series is going to be a big thorn in our butts to try and maintain a #1 seed. Top Hockey East teams like BC and UNH are going to continue to hold the Common Opponents point against us because of it. If we maintain a good RPI, that gives us a point over each of them, but like Jim said, the TUC percentage will be the deciding factor against both teams. UND must play at a high level against all TUC's from here on in. If so, we'll be a #1 seed. If we slip up, we'll be a #2 seed at best.

Good points for sure. Control that which you can control. Beat every greater-than-25-ranked-team and let the tourney seedings fall where they may. Do that, and a #1 seed belongs to the Fighting Sioux for sure. Get Gregoire and Hextall back in the process and go get #8!

[Vimar]

As has been discussed elsewhere, the Sioux are in the rare position of being in good shape for a tournament bid before New Years. I hope that means we're looking forward to a Spring of watching the Sioux scrap for a #1 seed, but either way there will be rankings to watch...

UND individual comparisons

...and I found myself wondering about this #2 vs #3 game, so thought it a good time to start a thread.

Though the Bulldogs are currently a spot above UND (PWR), UND is actually winning the comparison by virtue of the RPI tie-breaker. The H2H will clearly take over after tomorrow's game, making it a must-win for UND to remain #3 going into the new year.

As to the possibility of UND climbing if they win, here are the 3 comparisons UND is currently losing--

Boston College has TUC (.6364 to .5769) and COP (.8 to .4167). A win this weekend should only make up about half the difference on TUC, not enough to flip it.

Maine has H2H. No opportunity to flip this one until TUC or COP come into play.

Yale has RPI. UND's last win over UMD contributed something like .696, so it seems like UND would only rise to something like .586 with a win, not enough to take RPI.

Bottom line -- a loss almost certainly will lead to a drop, a win at best maintains the current ranking (all other things equal).

[Vimar]

Can anyone explain to me how Boston College improves in the pairwise when they beat one of the worst teams in the country.

Would it help the Sioux if next year we schedule

[Goon]

Can anyone explain to me how Boston College improves in the pairwise when they beat one of the worst teams in the country.

Would it help the Sioux if next year we schedule

[Big A HG]

Holy Cow that is an unimpressive schedule that is almost as bad as Yales...

That's not a schedule of anyone's (that I'm aware of)...just a list of bottom-feeders (minus Vermont) that he listed as tongue-in-cheek opponents we could schedule.

[siouxnami]

Can anyone explain to me how Boston College improves in the pairwise when they beat one of the worst teams in the country.

Would it help the Sioux if next year we schedule

[Vimar]

Minnesota

[siouxweet]

That Maine series is going to be a big thorn in our butts to try and maintain a #1 seed. Top Hockey East teams like BC and UNH are going to continue to hold the Common Opponents point against us because of it. If we maintain a good RPI, that gives us a point over each of them, but like Jim said, the TUC percentage will be the deciding factor against both teams. UND must play at a high level against all TUC's from here on in. If so, we'll be a #1 seed. If we slip up, we'll be a #2 seed at best.

right now we lose the comparison to bc and are winning against unh. what will really help us is for maine to play 5 more games against tuc's and we will atomatically flip that comparison as their tuc record is brutal. it wouldn't hurt for yale to start losing a few games. as well to drop their rpi.

[Goon]

That's not a schedule of anyone's (that I'm aware of)...just a list of bottom-feeders (minus Vermont) that he listed as tongue-in-cheek opponents we could schedule.

Actually their schedule really isn't that much better than the one listed for Boston College ->

October 2010

29 Fri W 7 - 4 Brown (nc) Box

30 Sat W 7 - 3 Dartmouth (nc) Box

November 2010

05 Fri W 5 - 3 Princeton Box

06 Sat W 5 - 1 Quinnipiac Box

13 Sat W 5 - 1 at Colorado College (nc) Box

14 Sun L 3 - 4 at Air Force (nc) Box

19 Fri W 4 - 2 at Cornell Box

20 Sat W 6 - 4 at Colgate Box

23 Tue W 5 - 1 Sacred Heart (nc) Box

December 2010

03 Fri W 4 - 2 Rensselaer Box

05 Sun W 5 - 0 Union Box

08 Wed W 3 - 0 Vermont (nc) Box

Beating the weak sisters fo the poor doesn't impress me...

Edited December 31, 2010 by Goon

[jimdahl]

Mankato made a huge leap in RPI and ascension to being a TUC, with wins over Notre Dame and Brown.

With UND's wins over Mankato, that's a nice boost to UND's TUC, and enough to flip the comparison with BC.

It's a nice reminder of how volatile things can be when the team you're watching isn't even playing, particularly when the TUC cliff gets involved.

[Big A HG]

Mankato made a huge leap in RPI and ascension to being a TUC, with wins over Notre Dame and Brown.

With UND's wins over Mankato, that's a nice boost to UND's TUC, and enough to flip the comparison with BC.

It's a nice reminder of how volatile things can be when the team you're watching isn't even playing, particularly when the TUC cliff gets involved.

Why is it that you and USCHO seem to have varying PWR Standings?

[jimdahl]

Why is it that you and USCHO seem to have varying PWR Standings?

Good eye, it took me a long time to even find where they deviated in the teens. It turns out I picked up a bad score for the Bowling Green vs. Holy Cross game. That had such minor implications that it escaped my attention. Entirely my fault, in that I've even got it set up so the system sent me an email (which I hadn't yet read) noting that USCHO and CHN were reporting different scores for the game. Fixed.

[siouxnami]

I was just looking at the PWR matchups and was looking at the Maine matchup. Those H2H losses hurt, but it appears that Maine has sucked against everyone else (different officials?). I also think that the second that Maine plays their 10th TUC game they are going to drop like a rock and UND will instantly flip that comparison. This will leave us 1 short of Yale's total. Assuming we stay strong we will be #1 in the west regardless of where Yale ends up...

[brianvf]

Those H2H losses hurt, but it appears that Maine has sucked against everyone else (different officials?).

Amazing how much harder it is to win games when you aren't given 10 straight power plays.

[Goon]

Jim the Whistler posted a good blog post on Goon's World...

Excluded Yale's 1 wins over Colgate because their game RPIs (0.5694) would have lowered Yale's RPI (0.6057) Excluded Yale's 1 wins over Sacred Heart because their game RPIs (0.5637) would have lowered Yale's RPI (0.6057) Excluded Yale's 1 wins over Holy Cross because their game RPIs (0.5886) would have lowered Yale's RPI (0.6057) Excluded Yale's 1 wins over Harvard because their game RPIs (0.5965) would have lowered Yale's RPI (0.6057)

There are four other team with one exclusion each. Yale's schedule is so weak that they have as many exclusions as the rest of the 57 teams put together.

I'd like to know what Yale's real RPI is. But the problem is, statistics are hard.

What is Yale's real RPI number...

[The Whistler]

Jim the Whistler posted a good blog post on Goon's World...

What is Yale's real RPI number...

Statistics are hard. I got A's in my two classes without learning anything.

[siouxu31]

Jim the Whistler posted a good blog post on Goon's World...

What is Yale's real RPI number...

I ran the numbers after including all games that were dropped due to a team defeating another team and having their R.P.I. go down. Yale is still in front of UND with an R.P.I. of .6057 to UND's .5921. I have an excel sheet that shows all 58 teams with these games included, but can't find a way to paste it on here and have it keep its formatting. If anybody knows of a way to paste an excel file on here let me know, or I can email it to somebody and they can put it on here for me.

[Goon]

I ran the numbers after including all games that were dropped due to a team defeating another team and having their R.P.I. go down. Yale is still in front of UND with an R.P.I. of .6057 to UND's .5921. I have an excel sheet that shows all 58 teams with these games included, but can't find a way to paste it on here and have it keep its formatting. If anybody knows of a way to paste an excel file on here let me know, or I can email it to somebody and they can put it on here for me.

Adam Wodon posted on my blog that the numbers from KRACH is the most accurate.

[MoreSiouxForYou]

Can someone help me? HOw come Wisco is # 8 in PWR but wasn't the win on Sat against Duluth their first against a rated team?

[jimdahl]

Jim the Whistler posted a good blog post on Goon's World...

What is Yale's real RPI number...

SiouxU31 is right, Yale would be .6057 if not for the adjustments. But, in the spirit of teaching a man to fish...

You can see the full details of the RPI calculation for any team by clicking "details" under the team name on the SiouxSports RPI page (when you lose this link, you can also get there by clicking rankings -> RPI under hockey on the front page).

Here is Yale's RPI page

At the top, it shows you the unadjusted calculation and the effect of the adjustments.

One of the really cool things about those RPI details pages are the "future RPI" tables. The top part predicts the team's end-of-season RPI, based on how many of their remaining games they win. The bottom half shows the RPI value of each remaining scheduled game. Those that have the win column grayed out would be negative impact wins, if everything stayed the same. So, if Yale continued at its current pace, 4 more wins would be tossed out by the end of the season.

[The Whistler]

I ran the numbers after including all games that were dropped due to a team defeating another team and having their R.P.I. go down. Yale is still in front of UND with an R.P.I. of .6057 to UND's .5921. I have an excel sheet that shows all 58 teams with these games included, but can't find a way to paste it on here and have it keep its formatting. If anybody knows of a way to paste an excel file on here let me know, or I can email it to somebody and they can put it on here for me.

Well there goes that theory. Still it cuts our deficit in nearly half.

Still if an unbiased person were going to pick a head to head matchup who do you think they'd pick?

[jimdahl]

Since Yale lost but there no was no movement in PWR, a few people have asked what it takes to pass Yale. Here's the short version:

Flipping the comparison with Yale without taking RPI isn't currently possible. RPI is the only criterion currently in use in the comparison between UND and Yale. So, if UND didn't take RPI, it would need to win two other comparison criteria to get a 2-1 comparison win (remember that UND is currently losing 2 comparisons, Yale and Maine, to Yale's 0, so if UND just flipped one it would need the tie-breaker -- RPI).

TUC is likely to come into play for Yale (it's current schedule and rankings have it meet exactly 10 TUCs, the minimum required number for the criterion to be used), but Yale has a commanding 1.000 record vs. TUCs compared to UND's .6750; Yale would pretty much have to lose all 4 for UND to take TUC, but it seems unlikely that Yale could lose those 4 while still beating UND in RPI.

COP will come into play when UND plays CC, but Yale will then have a 1.000 record vs. TUCs. If UND sweeps CC, that will result in a tie for TUC, any other result will give the point to Yale.

RPI is the shortest path to flipping the comparison, and probably necessary to keep the comparison through the end of the year (because of notes on TUC and COP above). What does it take to flip RPI?

* A loss to Clarkson would likely result in an RPI around 0.5965, which would likely be enough for UND to take the RPI lead with a win over UNO.

* A subsequent win over St. Lawrence would probably get dropped as a negative win, even with a loss to Clarkson; though a loss to St. Lawrence would also do the trick.

Bottom-line: One more Yale loss, assuming UND keeps winning, is likely to be enough for UND to take the lead in PWR. The CC series is critical to UND's prospects of holding on to the comparison vs. Yale until the end, no less than a sweep will do.