2017 Pairwise thread

[Tater]

Re: 2017 Pairwise thread

This many affect the WMU getting a #2/#3 seed issue.

<blockquote class="twitter-tweet" data-lang="en"><p lang="en" dir="ltr">We're re-running The Matrix b/c of glitch found affecting NCHC. Only major difference is North Dakota's chances of missing NCAAs <5% now</p>— College Hockey News (@chnews) <a href="https://twitter.com/chnews/status/841315124350722049">March 13, 2017</a></blockquote>
<script async src="//platform.twitter.com/widgets.js" charset="utf-8"></script>

 

[Numbers]

Re: 2017 Pairwise thread

Much better from CHN now.

Jim Dahl,
I have a question for you. Do Thursday's results offer any possibility of definite conclusions? In other words, Do PSU and OSU both winning shut AFA out of an at-large bid? Or, some other such thing?

 

[Crimson on the Glass]

Re: 2017 Pairwise thread

I can't get to work, either. I looked at Harvard's schedule, and chose everyone on their schedule to lose . . .

I hope nothing quite so drastic as that occurs. Other than the #4/#5 swapping we've seen recently, it seems like the top 4 tier is pretty well set, with reasonably different RPIs. Though with 3 of the 4 teams playing each other, clearly some losses are inevitable. For obvious reasons, I am rooting for Minnesota to move up to #4 by week's end.

Down lower, of course, the RPIs are so packed together that larger than I'd expect moves can happen based on just a game or two. I am surprised, for example, by the contention (above) that North Dakota not only falls out of the #3 seed level, but all the way out, if it does not win its tournament. I would argue that if you can win the NCHC tournament, your team is about the best there is. I don't want to see North Dakota fall into the #4 seed level, again because of what happens to the Crimson travel plans, and was starting to think they were pretty securely in the 9-12 group unless they decide to lose 2 this weekend. But perhaps not.

 

[brassbonanza]

Re: 2017 Pairwise thread

I hope nothing quite so drastic as that occurs. Other than the #4/#5 swapping we've seen recently, it seems like the top 4 tier is pretty well set, with reasonably different RPIs. Though with 3 of the 4 teams playing each other, clearly some losses are inevitable. For obvious reasons, I am rooting for Minnesota to move up to #4 by week's end.

Down lower, of course, the RPIs are so packed together that larger than I'd expect moves can happen based on just a game or two. I am surprised, for example, by the contention (above) that North Dakota not only falls out of the #3 seed level, but all the way out, if it does not win its tournament. I would argue that if you can win the NCHC tournament, your team is about the best there is. I don't want to see North Dakota fall into the #4 seed level, again because of what happens to the Crimson travel plans, and was starting to think they were pretty securely in the 9-12 group unless they decide to lose 2 this weekend. But perhaps not.

Harvard would only go west to play UND if all three of WMU, DU, and UMD are 1s. If BU or Minnesota finish on the 1 line and UND on the 4, Harvard likely stays east unless it swaps BU and Harvard for attendance purposes.

 

[Crimson on the Glass]

Re: 2017 Pairwise thread

Harvard would only go west to play UND if all three of WMU, DU, and UMD are 1s.

Yes, am aware! It seems less like a perfect storm these days, and more like something that could actually happen.

I can't imagine anyone making an informed decision about attendance that would favor Harvard staying close to home. BU and UML outdraw the Crimson at Bright-Landry, regrettably.

 

[UML Puck Hawk]

Re: 2017 Pairwise thread

Technically possible since CHN Probability Matrix gives Lowell 0.8% chance of getting a 4th but extremely unlikely. The following scenario where BU, Minnesota, Western Michigan (loses 2 games), and Union all lose. UML is still in 5th, but only by 0.0001 RPI. You can tinker and I'm sure find how UML makes up that small deficit, but I wouldn't place any bets on this.

http://pwp.uscho.com/rankings/pairwise-predictor/?uniq=pwp_58c6b44dadeb5

http://pwp.uscho.com/rankings/pairwise-predictor/?uniq=pwp_58c60859c2e90

 

[brassbonanza]

Re: 2017 Pairwise thread

Yes, am aware! It seems less like a perfect storm these days, and more like something that could actually happen.

And I really can't imagine anyone making a decision about attendance that would favor Harvard staying close to home. BU and UML outdraw the Crimson at Bright-Landry, regrettably.

What would also throw a wrench into that is if Wisconsin or (far less likely) one of the Michigans wins the B1G, then UND missing completely becomes a real possibility, in the scenario that UND loses two and falls to the 4 line territory.

 

[icehawk]

Re: 2017 Pairwise thread

Percentage of making tourney
http://collegehockeyranked.com/forecast/pwrtournament/

 

[SJHovey]

Re: 2017 Pairwise thread

Ok, I'm working on it with CHN. I'm going to keep somewhat quiet until we've got it figured out for sure and agree, but meanwhile here's a preview of what I'm seeing:

http://collegehockeyranked.com/forecast/pwrtournament/

Does a loss and a tie by North Dakota change the numbers at all?

 

[JimDahl]

Does a loss and a tie by North Dakota change the numbers at all?

Absolutely. Unfortunately, I keep those records by number of wins, so the 0 win scenario includes both loss+loss and loss+tie. Let me think if I can come up with an easy way to tease out that information.

 

[Scarlet]

Re: 2017 Pairwise thread

Does a loss and a tie by North Dakota change the numbers at all?

How can they tie? Aren't we at a point in the season where it's either win or lose?

 

[cetihcra]

Re: 2017 Pairwise thread

How can they tie? Aren't we at a point in the season where it's either win or lose?

NCHC has a third place game, which can end in a tie.

r

 

[UML Puck Hawk]

Re: 2017 Pairwise thread

How can they tie? Aren't we at a point in the season where it's either win or lose?

Question for pairwise people. In postseason tournaments with continuous OT isn't the game still considered a tie if it is scoreless after 5 min for pairwise purposes? I feel like for the sake of fairness that must be true, but I haven't seen it discussed anywhere.

 

[meauxmo]

Re: 2017 Pairwise thread

Harvard would only go west to play UND if all three of WMU, DU, and UMD are 1s. If BU or Minnesota finish on the 1 line and UND on the 4, Harvard likely stays east unless it swaps BU and Harvard for attendance purposes.

What happens to BU if they lose to BC on Friday? I assume BC is out if they lose?

 

[Scarlet]

Re: 2017 Pairwise thread

Question for pairwise people. In postseason tournaments with continuous OT isn't the game still considered a tie if it is scoreless after 5 min for pairwise purposes? I feel like for the sake of fairness that must be true, but I haven't seen it discussed anywhere.

Huh? You mean if a game goes into double OT and one team wins it, because neither team scored within the first five minutes of the first OT the game counts as a tie?? That's truly ridiculous if that's the case. Why bother playing OT then?

 

[tape]

Re: 2017 Pairwise thread

Question for pairwise people. In postseason tournaments with continuous OT isn't the game still considered a tie if it is scoreless after 5 min for pairwise purposes? I feel like for the sake of fairness that must be true, but I haven't seen it discussed anywhere.

I feel like I would have read that somewhere over the last 15 years or so and I've definitely never heard of that being a thing.

What happens to BU if they lose to BC on Friday? I assume BC is out if they lose?

Jim Dahl's chart has an extremely slim chance they make it in without winning one game this weekend. I've been playing with the predictor and can't find a way to make it happen yet. The big sticking point is that somehow they have to get past Penn State and Ohio State and I can't figure out how to get that to happen.

 

[UML Puck Hawk]

Re: 2017 Pairwise thread

Huh? You mean if a game goes into double OT and one team wins it, because neither team scored within the first five minutes of the first OT the game counts as a tie?? That's truly ridiculous if that's the case. Why bother playing OT then?

Because tournaments need a winner. But the Pairwise doesn't?

Why in the eyes of a computer should the games played this weekend (and any game 3's of series) count more than a game earlier in the season. Seems weird to me to count a win now if it wouldn't have counted two weeks ago...

 

[Rhett Hot]

Re: 2017 Pairwise thread

Question for pairwise people. In postseason tournaments with continuous OT isn't the game still considered a tie if it is scoreless after 5 min for pairwise purposes? I feel like for the sake of fairness that must be true, but I haven't seen it discussed anywhere.

No. In the postseason there are no ties. You could go 7 OT, and the winner is still the winner, including for PWR purposes.

 

[UML Puck Hawk]

Re: 2017 Pairwise thread

No. In the postseason there are no ties. You could go 7 OT, and the winner is still the winner, including for PWR purposes.

Except in consolation games...

Can you link to any actual statement which backs this up, because I can't find one anywhere.

 

[tape]

Re: 2017 Pairwise thread

Except in consolation games...

Can you link to any actual statement which backs this up, because I can't find one anywhere.

Honestly I think your side of the argument is the one that needs positive proof. I don't know how logic suggests that there be a tie for ranking purposes when there is no possibility a given game ends in a tie.

I think the reason you haven't seen it discussed anywhere is because that's not how it works.