Union @ Cornell

[Trotsky]

Cornell 1969-70 44x122 .3607

[Trotsky]

At the risk of sounding like a Luddite or an idiot, we win 10-1 and drop in the PWR. This suggests that what other teams do is more important than what you do. SMH

It seems dumb but what it means is we beat somebody bad so we didn't get very much credit for it.  Also, some other stuff probably happened to shift around our SOS.

[arugula]

I know I just wish I could see from the scores what that would be.

[Trotsky]

I know I just wish I could see from the scores what that would be.

UMD lost, that probably didn't help.

[Dafatone]

At the risk of sounding like a Luddite or an idiot, we win 10-1 and drop in the PWR. This suggests that what other teams do is more important than what you do. SMH

It seems dumb but what it means is we beat somebody bad so we didn't get very much credit for it.  Also, some other stuff probably happened to shift around our SOS.

It's a little more complicated than that. It would help if I could find the pairwise from before tonight's games. But what I think happened is this:

There's us, Michigan Tech, and Harvard. We had the best RPI, which is the tiebreaker. Then Harvard, then Michigan Tech. Cause Harvard beat us twice they win their comparison to us. And we lose ours with Minnesota Duluth as well. Harvard wound up 1st of the three and two comparisons ahead of us, then we were tied with Tech, so we were ahead cause of the tiebreaker.

Tonight, Tech passed Harvard in RPI and flipped that comparison. Now they're each one comparison ahead of us and tied with each other, and Tech wins that tiebreak. So it's Tech, Harvard, us.

We have a good chance of flipping the Harvard comparison back if we either play well enough to win the common opponent category or beat Harvard head to head. UMD, oh well, that ship has sailed.

But we are 10th in RPI, which is a good place to be in.
that Michigan Tech passed Harvard in RPI and leapfrogged us because they flipped their pairwise with Harvard.

[Trotsky]

We are also on the wrong end of a weird situation.  PWR tracks almost perfectly with RPI, but we are behind not just Harvard but also Michigan Tech in PWR despite being ahead of them in RPI.  We are actually #10 in RPI.

[BearLover]

At the risk of sounding like a Luddite or an idiot, we win 10-1 and drop in the PWR. This suggests that what other teams do is more important than what you do. SMH

Not really. First of all, score differential doesn't factor in. Second, our RPI improved. That creates more separation between us and the teams below us, which makes it more likely we qualify for the NCAAs. After tonight we are more likely to make the NCAAs than we were before tonight.

We dropped in the PWR because we win fewer "comparisons" than Michigan Tech and Harvard, who are tied for #10, despite us having better RPI than them both. They each win 50 comparisons and we win 49. The reason for this is that we lost two games to Harvard and two to Minnesota-Duluth, so we lose comparisons to those teams.

There will be weird fluctuations in the Pairwise for the rest of the season. It definitely isn't great that we lost to Harvard and Duluth twice. But if we maintain our current level of winning, we will make the NCAAs. We are in good position now, with quite a bit of RPI separation between us and the teams below us.

[Trotsky]

(jinx)

[upprdeck]

Can we play well down the stretch..  we should be top 4..
So that means 5 home games left, need them all.

then road vs colg/brown/yale..
Colgate has played us tough lately.

Brown/Yale 2 of the bottom 3 with little to play for, can we be mentally tough to end the yr.

if we got 7-1 to get to the ECAC final 4 we should be in good shape.

[Trotsky]

We are in good position now, with quite a bit of RPI separation between us and the teams below us.

We are actually closer to 9 than 11 and to 8 than 12 in RPI:

Code Select Expand


08  .562  +.007  St. Cloud
09  .558  +.003  Western Michigan
[color=#FF0000]10  .555         Cornell[/color]
11  .550  -.005  Michigan Tech
12  .547  -.008  Harvard


[arugula]

Based on the idea that our opponents doing well should help our PWR, I get why UMD and Clarkson faltering hurts us, but shouldn't BU and Q winning help us? Doesn't seem to. I'm just exasperated.  I know theres logic to this but the logic seems to always go the other way.

[Trotsky]

Based on the idea that our opponents doing well should help our PWR, I get why UMD and Clarkson faltering hurts us, but shouldn't BU and Q winning help us? Doesn't seem to. I'm just exasperated.  I know theres logic to this but the logic seems to always go the other way.

Q winning doesn't help because it's in conference against a team we have played the same number of times.  So robbing Peter pays Paul.

[BearLover]

Based on the idea that our opponents doing well should help our PWR, I get why UMD and Clarkson faltering hurts us, but shouldn't BU and Q winning help us? Doesn't seem to. I'm just exasperated.  I know theres logic to this but the logic seems to always go the other way.

In-conference games are basically a wash since we play both the loser and the winner. As mentioned above, our RPI, which is a function of win% and SOS, improved tonight. The Pairwise is largely RPI, but not exactly (the difference was summarized earlier in this thread, but there are also more involved primers on the ranking that can be found online). We are in better position to make the NCAAs now than we were before tonight, because our RPI improved and we avoided a bad loss. Don't pay too much attention to small fluctuations in the Pairwise or you will miss the forest for the trees.

[arugula]

Got it. Still doesn't explain why BU shooting up the ratings hasn't helped. Or perhaps it has just not as much as other factors have hurt us.

The idea that the conference games are a wash supports the idea that we should schedule as strong an OOC slate as possible.  We had this debate a few years ago and the perspective was decidedly split.  I guess UMD is theoretically an opponent to strengthen that aspect but they have had a down year. Instead of AIC and SHU perhaps Hockey East  opponents?  Otoh you want to win the OOC games. Science!

[Trotsky]

Don't pay too much attention to small fluctuations in the Pairwise or you will miss the forest for the trees.

PWR is one of those things that has meaning for any given reading, but has no meaning when you look at local fluctuations around a given reading.  I'm sure there is a clever word for this in metrics.