This Weekend's Games Rescheduled

[Trotsky]

Well, not if you're a cheap ILDN subscriber like me, and only pay for Cornell games.

And me.

I may finally take the big plunge next year, but I do most of my non-Cornell watching during the playoffs which, IIRC, aren't on ILDN.

What we need is a fucking ECACDN.

[andyw2100]

Well, not if you're a cheap ILDN subscriber like me, and only pay for Cornell games.

And me.

I may finally take the big plunge next year, but I do most of my non-Cornell watching during the playoffs which, IIRC, aren't on ILDN.

What we need is a fucking ECACDN.

I'd settle for a "by sport" instead of "by school" option. I'd be more likely to watch other Ivy League hockey games than I am to watch other Cornell sports during the four months of the year I subscribe.

[Trotsky]

Well, not if you're a cheap ILDN subscriber like me, and only pay for Cornell games.

And me.

I may finally take the big plunge next year, but I do most of my non-Cornell watching during the playoffs which, IIRC, aren't on ILDN.

What we need is a fucking ECACDN.

I'd settle for a "by sport" instead of "by school" option. I'd be more likely to watch other Ivy League hockey games than I am to watch other Cornell sports during the four months of the year I subscribe.

You know the problem with that.  Not a lot of interest/traffic for the Field Hockey package.  Though maybe if they bring back the skirts...

[jkahn]

I take the perspective that there is no outcome of that game that doesnt help us in some way. its really unclear which one helps us the most. my head says a harvard win does, since we still play union again, but my heart has a hard time with that thought.

A Union win can help our quality win points, since the higher they finish the more quality win points we get.  And we also have a second shot at getting more.

I thought quality win bonus was based solely on their Pairwise position at the time of the game, not where they end up.

It's got to be based on where they end up.  First of all, it wouldn't make sense to give a team credit for beating a highly positioned team two weeks into the season, when that team is ultimately mediocre.  And, what would you do with wins in the first game of the season.  Plus, I just checked a couple of examples, including Cornell, it it seems to agree with where Union and St. Lawrence are now positioned.

[Beeeej]

I take the perspective that there is no outcome of that game that doesnt help us in some way. its really unclear which one helps us the most. my head says a harvard win does, since we still play union again, but my heart has a hard time with that thought.

A Union win can help our quality win points, since the higher they finish the more quality win points we get.  And we also have a second shot at getting more.

I thought quality win bonus was based solely on their Pairwise position at the time of the game, not where they end up.

It's got to be based on where they end up.  First of all, it wouldn't make sense to give a team credit for beating a highly positioned team two weeks into the season, when that team is ultimately mediocre.  And, what would you do with wins in the first game of the season.  Plus, I just checked a couple of examples, including Cornell, it it seems to agree with where Union and St. Lawrence are now positioned.

You make a compelling argument, and I'm sure you're right about how it does work, but that system still values "quality wins" differently at different times of the season. Let's say Penn State's decline wasn't just a regression to the mean, but they lost half their players to injury, and lost the rest of their games. Team X, who beat them early in the season when they were in the top few slots in the Pairwise, really did get a quality win over a strong team, but if Penn State ends up #25, Team X gets screwed out of that bonus regardless of the reason for the decline.

I guess no system is perfect, but I wonder if there's a mathematical way to find a middle-ground solution.

[Trotsky]

You make a compelling argument, and I'm sure you're right about how it does work, but that system still values "quality wins" differently at different times of the season. Let's say Penn State's decline wasn't just a regression to the mean, but they lost half their players to injury, and lost the rest of their games. Team X, who beat them early in the season when they were in the top few slots in the Pairwise, really did get a quality win over a strong team, but if Penn State ends up #25, Team X gets screwed out of that bonus regardless of the reason for the decline.

I guess no system is perfect, but I wonder if there's a mathematical way to find a middle-ground solution.

You could wait until the end of the season, then retroactively go back and calculate every opponent's strength at the time of your game against them based on their results to that point against their opponents end-season SOS.

Very simple example:

Let's say Penn State goes 13-2 losing games 5 and 15, then has a Horrible Zamboni Accident, and goes 2-13 winning games 20 and 30.  You step through the schedule day by day adjusting opponent strength to what it is at the moment the puck drops on each game, so the teams who beat PSU on games 5 and 15 get a VERY sweet boost.  The teams that start beating them on game 21 still get a good boost for a while, but it sinks until the win against PSU on game 29 and the loss against them on game 30 are +/- virtually the same amount, since at that point they are effectively a .500 team.

It would be a weird hybrid with basically no higher theoretical statistical justification.  But it would try to capture both concerns.

it might be better to front-load the schedule with say 4 dummy results of .500 ties to wash out the unduly high or low effects of the very low denominators early in the season.

Anyway, it could be done, but it winds up looking like those 11th grade Coronet films of spiders spinning their webs on acid.

[Beeeej]

You make a compelling argument, and I'm sure you're right about how it does work, but that system still values "quality wins" differently at different times of the season. Let's say Penn State's decline wasn't just a regression to the mean, but they lost half their players to injury, and lost the rest of their games. Team X, who beat them early in the season when they were in the top few slots in the Pairwise, really did get a quality win over a strong team, but if Penn State ends up #25, Team X gets screwed out of that bonus regardless of the reason for the decline.

I guess no system is perfect, but I wonder if there's a mathematical way to find a middle-ground solution.

You could wait until the end of the season, then retroactively go back and calculate every opponent's strength at the time of your game against them based on their results to that point against their opponents end-season SOS.

Very simple example:

Let's say Penn State goes 13-2 losing games 5 and 15, then has a Horrible Zamboni Accident, and goes 2-13 winning games 20 and 30.  You step through the schedule day by day adjusting opponent strength to what it is at the moment the puck drops on each game, so the teams who beat PSU on games 5 and 15 get a VERY sweet boost.  The teams that start beating them on game 21 still get a good boost for a while, but it sinks until the win against PSU on game 29 and the loss against them on game 30 are +/- virtually the same amount, since at that point they are effectively a .500 team.

It would be a weird hybrid with basically no higher theoretical statistical justification.  But it would try to capture both concerns.

it might be better to front-load the schedule with say 4 dummy results of .500 ties to wash out the unduly high or low effects of the very low denominators early in the season.

Anyway, it could be done, but it winds up looking like those 11th grade Coronet films of spiders spinning their webs on acid.

Somehow I knew you'd be the first to submit a proposal.

[Trotsky]

Somehow I knew you'd be the first to submit a proposal.

"I GOT NOWHERE ELSE TO GO!!!!"   :`-(

[ugarte]

You make a compelling argument, and I'm sure you're right about how it does work, but that system still values "quality wins" differently at different times of the season. Let's say Penn State's decline wasn't just a regression to the mean, but they lost half their players to injury, and lost the rest of their games. Team X, who beat them early in the season when they were in the top few slots in the Pairwise, really did get a quality win over a strong team, but if Penn State ends up #25, Team X gets screwed out of that bonus regardless of the reason for the decline.

I guess no system is perfect, but I wonder if there's a mathematical way to find a middle-ground solution.

You could wait until the end of the season, then retroactively go back and calculate every opponent's strength at the time of your game against them based on their results to that point against their opponents end-season SOS.

Very simple example:

Let's say Penn State goes 13-2 losing games 5 and 15, then has a Horrible Zamboni Accident, and goes 2-13 winning games 20 and 30.  You step through the schedule day by day adjusting opponent strength to what it is at the moment the puck drops on each game, so the teams who beat PSU on games 5 and 15 get a VERY sweet boost.  The teams that start beating them on game 21 still get a good boost for a while, but it sinks until the win against PSU on game 29 and the loss against them on game 30 are +/- virtually the same amount, since at that point they are effectively a .500 team.

It would be a weird hybrid with basically no higher theoretical statistical justification.  But it would try to capture both concerns.

it might be better to front-load the schedule with say 4 dummy results of .500 ties to wash out the unduly high or low effects of the very low denominators early in the season.

Anyway, it could be done, but it winds up looking like those 11th grade Coronet films of spiders spinning their webs on acid.

NCAA: We need a simple way to reward quality wins.
Beeeej: How about a complicated way?
Trotsky: Always here for you, man.

[Trotsky]

NCAA: We need a simple way to reward quality wins.
Beeeej: How about a complicated way?
Trotsky: Always here for you, man.

It's actually no more complicated because once it's coded you press one fucking button. ::banana::

[nshapiro]

NCAA: We need a simple way to reward quality wins.
Beeeej: How about a complicated way?
Trotsky: Always here for you, man.

It's actually no more complicated because once it's coded you press one fucking button. ::banana::

There is only one more change worth making to the selection process - KRACH

[Trotsky]

There is only one more change worth making to the selection process - KRACH

Agreed.  No, wait.

(confirms Cornell is higher in KRACH than PWR)

Agreed.

The PlayoffStatus.com "Pairwise" seems to actually be KRACH.

[Swampy]

Well, that sucks.  I was looking forward to coming up.

Fortunately they switched my tickets to Senior Weekend (RPI/Union).

"Senior" as in "about to graduate" or "senior" as in "can't remember when or even if I graduated?"

[billhoward]

One good thing about the change- if you have ILDN you can see Union @ Harvard Friday night.  A big game for all three teams.  Might be a trap game for Harvard looking forward to Beanpot final. ::banana::

All three teams? Am I missing something?

[billhoward]

I was thinking the miserable bus rides on Friday to Colgate and Cornell would add some home-ice advantage.