Q Game Thread

[marty]

Saw the drop in the pwr.  Who wants to take a shot at explaining that?

Game counts as .44 wins and .54 losses, which certainly lowers our win %.
It's 55% win x .8 (home win factor) and 45% loss x 1.2 (home loss factor).

Think also how this affects Q. They feel like crap (or hopefully worse) in the loss but get .54 of a win and only .44 of a loss. So they lose but their pairwise goes up?

And because .8 is not the reciprocal of 1.2 it is only .98 if a game.  Perfect.

[upprdeck]

Saw the drop in the pwr.  Who wants to take a shot at explaining that?

Game counts as .44 wins and .54 losses, which certainly lowers our win %.
It's 55% win x .8 (home win factor) and 45% loss x 1.2 (home loss factor).

the formula should not result in an OT win at home being worth less than a neutral site tie. that's dumb. (not your math, the rules that led to your math).

The home/away formula is dumb, and possibly not rooted in reality (is it actually that much harder to win on the road?), but from what I understand, it isn't actually competitively unfair. The formula might bias more against home teams than home-ice advantage is actually worth in reality. But such bias kicks in whether the home team wins, loses, or ties. So, for instance, Cornell might not have gotten as many points for an OT tie thanks to this bias, but this same bias would have hurt Cornell just the same had it lost in OT, or lost in regulation. The delta between outcomes (win, tie, loss) is no different whether you're at home or on the road. The home/away bias means there is more to lose at home, and more to gain on the road, but the absolute stakes are exactly the same. Therefore, from a competition standpoint, assuming all teams play the same ratio of home and away games, this weird quirk in the formula should be a wash.

The 55/45 breakdown on OT win/OT loss is a subjective weighting of how much 3v3 overtime should count. Personally, I'm okay with this weighting. 3v3 OT is exciting and fans hate ties; but it's a gimmick, and should not be afforded nearly the same weight as a game decided in regulation. But a slight effect on the PWR makes the OT at least somewhat meaningful.

Does it feel strange that Cornell beat a top team (#6 in the PWR prior to tonight) and dropped in the PWR? Sure, but the same quirks in the PWR that caused us to drop with an OT win would have caused us to drop more with an OT loss, and quite a bit more with a regulation loss. Those are the alternatives Cornell was facing tonight, and why the outcome was still just as important as that of any other game (actually slightly more important given the quality win bonus).

explain the logic of 3x3 counting less in the first place.. you can have 3x3 OT even in a game with reg OT rules.. we actually won on a 4x3 play.  I can see saying a shootout is different but 3x3 happens all the time in  a game so no reason to value it any  less.

[Trotsky]

The home/away formula is dumb, and possibly not rooted in reality (is it actually that much harder to win on the road?)

I think the purpose was to provide an incentive for teams to travel -- particularly teams with huge rinks who suffer a non-trivial revenue hit when they sacrifice a home date.

[Dafatone]

The home/away formula is dumb, and possibly not rooted in reality (is it actually that much harder to win on the road?)

I think the purpose was to provide an incentive for teams to travel -- particularly teams with huge rinks who suffer a non-trivial revenue hit when they sacrifice a home date.

Yup. I still think it goes too far, but it makes some sense given that goal.

[upprdeck]

The home/away formula is dumb, and possibly not rooted in reality (is it actually that much harder to win on the road?)

I think the purpose was to provide an incentive for teams to travel -- particularly teams with huge rinks who suffer a non-trivial revenue hit when they sacrifice a home date.

you could also put limits on home.away games to force teams to travel which would allow the formula to not need as much tweaking.

[DL]

Saw the drop in the pwr.  Who wants to take a shot at explaining that?

Game counts as .44 wins and .54 losses, which certainly lowers our win %.
It's 55% win x .8 (home win factor) and 45% loss x 1.2 (home loss factor).

Think also how this affects Q. They feel like crap (or hopefully worse) in the loss but get .54 of a win and only .44 of a loss. So they lose but their pairwise goes up?

And because .8 is not the reciprocal of 1.2 it is only .98 if a game.  Perfect.

Right, someone else may not /need/ to explain this to me, but /I/ need it explained.

[BearLover]

Saw the drop in the pwr.  Who wants to take a shot at explaining that?

Game counts as .44 wins and .54 losses, which certainly lowers our win %.
It's 55% win x .8 (home win factor) and 45% loss x 1.2 (home loss factor).

Think also how this affects Q. They feel like crap (or hopefully worse) in the loss but get .54 of a win and only .44 of a loss. So they lose but their pairwise goes up?

And because .8 is not the reciprocal of 1.2 it is only .98 if a game.  Perfect.

Right, someone else may not /need/ to explain this to me, but /I/ need it explained.

The key is to view OT games as essentially a tie, for PWR purposes. Cornell has played five such games this year: two against Alaska, one against Brown, one against Clarkson, one against Q. In my mind, those games were ties. The ones against Alaska and Brown were poor performances. The ones against Clarkson and Q were fine outcomes (notwithstanding the collapse against Clarkson). Everyone needs to get on the same page with what counts as a win. No more dropping OT wins into the W column.

[Al DeFlorio]

Saw the drop in the pwr.  Who wants to take a shot at explaining that?

Game counts as .44 wins and .54 losses, which certainly lowers our win %.
It's 55% win x .8 (home win factor) and 45% loss x 1.2 (home loss factor).

Think also how this affects Q. They feel like crap (or hopefully worse) in the loss but get .54 of a win and only .44 of a loss. So they lose but their pairwise goes up?

And because .8 is not the reciprocal of 1.2 it is only .98 if a game.  Perfect.

Right, someone else may not /need/ to explain this to me, but /I/ need it explained.

The key is to view OT games as essentially a tie, for PWR purposes. Cornell has played five such games this year: two against Alaska, one against Brown, one against Clarkson, one against Q. In my mind, those games were ties. The ones against Alaska and Brown were poor performances. The ones against Clarkson and Q were fine outcomes (notwithstanding the collapse against Clarkson). Everyone needs to get on the same page with what counts as a win. No more dropping OT wins into the W column.

That is, in fact, now the reality.  Given that, why bother with overtime at all?  And, why do it with a bogus three-on-three format?  Worse, if we should think of it as a tie, anywsy, why add a meaningless shoot-out on top of an unresolved overtime?  That's all now unnecessary.  Why require six numbers to show a team's overall record when three would do?  Classic case of over-thinking something when there's an obvious simple solution.

[BearLover]

Saw the drop in the pwr.  Who wants to take a shot at explaining that?

Game counts as .44 wins and .54 losses, which certainly lowers our win %.
It's 55% win x .8 (home win factor) and 45% loss x 1.2 (home loss factor).

Think also how this affects Q. They feel like crap (or hopefully worse) in the loss but get .54 of a win and only .44 of a loss. So they lose but their pairwise goes up?

And because .8 is not the reciprocal of 1.2 it is only .98 if a game.  Perfect.

Right, someone else may not /need/ to explain this to me, but /I/ need it explained.

The key is to view OT games as essentially a tie, for PWR purposes. Cornell has played five such games this year: two against Alaska, one against Brown, one against Clarkson, one against Q. In my mind, those games were ties. The ones against Alaska and Brown were poor performances. The ones against Clarkson and Q were fine outcomes (notwithstanding the collapse against Clarkson). Everyone needs to get on the same page with what counts as a win. No more dropping OT wins into the W column.

That is, in fact, now the reality.  Given that, why bother with overtime at all?  And, why do it with a bogus three-on-three format?  Worse, if we should think of it as a tie, anywsy, why add a meaningless shoot-out on top of an unresolved overtime?  That's all now unnecessary.  Why require six numbers to show a team's overall record when three would do?  Classic case of over-thinking something when there's an obvious simple solution.

I would guess that 3-on-3 OT plus shootout to determine a winner is much more entertaining to the casual fan than simply ending a game in a tie or playing 5 minutes or 5-on-5.

[JohnF81]

Other than the second period when they had their PP advantage, shots and possession were about even. As you point out, we were missing our top two goal scorers.  And .... we won!  Q is not way better.

[RichH]

Other than the second period when they had their PP advantage, shots and possession were about even. As you point out, we were missing our top two goal scorers.  And .... we won!  Q is not way better.

I think they're better overall, even accounting for our missing players, but it's not a huge gap. Grady & Tim made a good point on the broadcast that QU has the advantage of having played a season last year. These guys had a chance to have continuity among key lines/players and cemented that knowledge of each others' habits. I don't think you can measure that quantitatively, but I think it counts for something.

Cornell seems susceptible to "rust" in coming back from time off.

[adamw]

It's a net negative because of the same home/road weighting that affects every other team. Cornell will benefit from the same weighting when it goes on the road against Q (or any other team). Because of home/road weighting, all possible outcomes of a home game average out to a net negative.

0.8 of a win is not a net negative -- at least not in the way I'm using the term. If your winning percentage for the day is .800 - then that's going to raise your RPI.

[nshapiro]

Saw the drop in the pwr.  Who wants to take a shot at explaining that?

Game counts as .44 wins and .54 losses, which certainly lowers our win %.
It's 55% win x .8 (home win factor) and 45% loss x 1.2 (home loss factor).

Think also how this affects Q. They feel like crap (or hopefully worse) in the loss but get .54 of a win and only .44 of a loss. So they lose but their pairwise goes up?

And because .8 is not the reciprocal of 1.2 it is only .98 if a game.  Perfect.

Right, someone else may not /need/ to explain this to me, but /I/ need it explained.

The key is to view OT games as essentially a tie, for PWR purposes. Cornell has played five such games this year: two against Alaska, one against Brown, one against Clarkson, one against Q. In my mind, those games were ties. The ones against Alaska and Brown were poor performances. The ones against Clarkson and Q were fine outcomes (notwithstanding the collapse against Clarkson). Everyone needs to get on the same page with what counts as a win. No more dropping OT wins into the W column.

That is, in fact, now the reality.  Given that, why bother with overtime at all?  And, why do it with a bogus three-on-three format?  Worse, if we should think of it as a tie, anywsy, why add a meaningless shoot-out on top of an unresolved overtime?  That's all now unnecessary.  Why require six numbers to show a team's overall record when three would do?  Classic case of over-thinking something when there's an obvious simple solution.

I would guess that 3-on-3 OT plus shootout to determine a winner is much more entertaining to the casual fan than simply ending a game in a tie or playing 5 minutes or 5-on-5.

Using the kind of delicate language Trotsky would - Fuck the casual fan.  Scrap the 3-on-3, play a 5-on-5 overtime, in which the winner gets credit for a win, and if you have to, have a shootout for conference purposes that would count as a tie in pairwise.

[adamw]

Think also how this affects Q. They feel like crap (or hopefully worse) in the loss but get .54 of a win and only .44 of a loss. So they lose but their pairwise goes up?

well, to be anal, they went down too - because their RPI was already above .54

[adamw]

Given that, why bother with overtime at all?  And, why do it with a bogus three-on-three format?  Worse, if we should think of it as a tie, anywsy, why add a meaningless shoot-out on top of an unresolved overtime?  That's all now unnecessary.  Why require six numbers to show a team's overall record when three would do?  Classic case of over-thinking something when there's an obvious simple solution.

The answer is the same any time someone is not doing the obvious simple solution ... it's not like no one is aware of the simple solution - they just don't want to. What we have now is a compromise between people who love 3-on-3 and want NCAA to "be like the NHL" - and those who abhor it and don't want it to count for anything.