Ivy Title
Posted by Brunke
Ivy Title
Posted by: Brunke (216.133.203.---)
Date: February 09, 2002 04:54PM
So with Cornell's win last night and Dartmouth's loss we now have a 2 point lead for the ivy title (right?, according to ivyleaguesports.com) so if we win tonight does that give us the outright title even without having played the Big Green up in their house?
Re: Ivy Title
Posted by: jtwcornell91 (---.utb.edu)
Date: February 09, 2002 05:01PM
No, Dartmouth is 4-2-1, so if they win out they will be 7-2-1 and if we beat Yale and lose to Dartmouth we will be 6-3-1. (For that matter, Harvard is currently 5-3-1.)
Re: Ivy Title
Posted by: Craine (---.nyc.rr.com)
Date: February 09, 2002 05:05PM
A win tonight does not seal it for us... IF we win tonight, we'd have 13 points with one game (Dartmouth) to go... Dartmouth has 9 points, but they get to play us, along with Princeton and Yale... So, they could end up with 15 points...
Right now, the standing are (based on Points Per Game):
Cornell - 1.375 (11 Points in 8 Games)
Dartmouth - 1.2857 (9 Points in 7 Games)
Harvard - 1.125 (8 Points in 9 Games)
Yale - 0.833 (5 Points in 6 Games)
Brown - 0.75 (6 Points in 8 Games)
Princeton - 0.5714 (4 Points in 7 Games)
A win tonight would allow us to control our own destiny when we go to Hanover next Friday...
Right now, the standing are (based on Points Per Game):
Cornell - 1.375 (11 Points in 8 Games)
Dartmouth - 1.2857 (9 Points in 7 Games)
Harvard - 1.125 (8 Points in 9 Games)
Yale - 0.833 (5 Points in 6 Games)
Brown - 0.75 (6 Points in 8 Games)
Princeton - 0.5714 (4 Points in 7 Games)
A win tonight would allow us to control our own destiny when we go to Hanover next Friday...
Re: Ivy Title
Posted by: Graham '02 (---.twcny.rr.com)
Date: February 09, 2002 05:49PM
So a win for us tonight makes next weekends Dartmouth game the de facto Ivy League championship game, right?
Re: Ivy Title
Posted by: kaelistus (---.sbo.ma.webcache.rcn.net)
Date: February 09, 2002 05:54PM
If we beat Dartmouth next week we are guaranteed at least a share of the title irregardless of what happens tonight. If we don't beat Dartmouth then the odds aren't so great.
Re: Ivy Title
Posted by: mjh40 (---.twcny.rr.com)
Date: February 09, 2002 09:51PM
Greg, Correction on Yale...2-4-1.
Re: Ivy Title
Posted by: Greg Berge (---.dial.spiritone.com)
Date: February 09, 2002 09:54PM
Done. Thx.
Re: Ivy Title
Posted by: kingpin248 (---.twcny.rr.com)
Date: February 09, 2002 10:28PM
Cornell wins the Ivy title outright if it beats Dartmouth next Friday night. The Red would then have 15 league points, and no other team would be able to get as high as 13.
Ivy Magic Numbers
Posted by: KeithK (---.dsl.snfc21.pacbell.net)
Date: February 09, 2002 10:56PM
The following are a set of Magic Numbers for Ivies, ala John Whelan's ECAC Nutshell page at slack.net. Of course, since there are no Ivy tiebreakers things are a little easier:
so even a tie Friday clinches a share of the title.
And if that doesn' sufficiently demonstrate my hockey geekdom... Cornell gets a share of the Ivy title in 1944 of 2187 possible outcomes, or 88.9%.
Cor Dar Har Brn Yal Pri
Cornell 13- 5-2 - 3 1 0 0 0
Dartmouth 9- 5-6 7 - 5 2 3 2
Harvard 9- 7-4 7 7 - 2 3 2
Brown 6-10-4 xx 10 8 - 6 5
Yale 5- 9-6 xx 11 9 6 - 6
Princeton 4-10-6 xx 12 10 7 8 -
Remember that getting a MN down to one is enough to win a share of the title
so even a tie Friday clinches a share of the title.
And if that doesn' sufficiently demonstrate my hockey geekdom... Cornell gets a share of the Ivy title in 1944 of 2187 possible outcomes, or 88.9%.
Re: Ivy Magic Numbers
Posted by: Robb (---.156.78.127.dial1.dallas1.level3.net)
Date: February 09, 2002 11:24PM
Keith,
Does your script just do an exhaustive search of the possibilities? I've been working on some tools of my own, but it gets harder when you look at the ECAC as a whole. E.g. with only two weekends left, there are 3^24 = 282,429,536,481 cases to check. If you could check 100 cases per second, that's about 89.49 years to run all the combinations (yes, accounting for leap years
). Anyway, just curious if you had figured out something more clever, or if your algorithm only works for small numbers of games (12 games remaining would only take 88 minutes). So you can't do an exhaustive search until the very last weekend of the season - pretty pointless by then.
Geeky minds want to know!
Does your script just do an exhaustive search of the possibilities? I've been working on some tools of my own, but it gets harder when you look at the ECAC as a whole. E.g. with only two weekends left, there are 3^24 = 282,429,536,481 cases to check. If you could check 100 cases per second, that's about 89.49 years to run all the combinations (yes, accounting for leap years
). Anyway, just curious if you had figured out something more clever, or if your algorithm only works for small numbers of games (12 games remaining would only take 88 minutes). So you can't do an exhaustive search until the very last weekend of the season - pretty pointless by then.Geeky minds want to know!
Re: Ivy Title
Posted by: DeltaOne81 '03 (---.resnet.cornell.edu)
Date: February 10, 2002 01:53AM
Glad to find some other stat freaks on here ![B-]](http://elf.elynah.com/smileys/11.gif "B-]")
Anyhow, after tonight's last second (well, last 15.6 seconds) win this is how the Ivy league title plays out...
If we win at Dartmouth, we win the Ivy title outright - we're up by 6 on them and Harvard (both would then have two games left)
If we tie at Dartmouth, we're up by 4 on them (they having two games to play), so we at worst tie for the title if Dartmouth beats Yale and Princeton. We'd be up with 5 on Harvard and they'd have 2 games left, so it would at worst be a two way tie for the Ivy title.
If we loose at Dartmouth, well, this gets sticky... we're only up by two on them, 4 on Harvard - each of them having two games left. Harvard can at best tie us, Dartmouth can pass us w/ 3 or more pts vs. Yale/Princeton, tie us w/ 2 or we can still win the Ivies if they only get one. Either way, Ivy title wise, it's big that we at least don't loose next Friday.
Oh, did I mention that Harvard's remaining two Ivy games are also Yale/Princeton... those guys get around . And as a final, obvious, note: Yale, Princeton, and Brown have no chance at the title.
Either way, Let's Go Red!
Anyhow, after tonight's last second (well, last 15.6 seconds) win this is how the Ivy league title plays out...
If we win at Dartmouth, we win the Ivy title outright - we're up by 6 on them and Harvard (both would then have two games left)
If we tie at Dartmouth, we're up by 4 on them (they having two games to play), so we at worst tie for the title if Dartmouth beats Yale and Princeton. We'd be up with 5 on Harvard and they'd have 2 games left, so it would at worst be a two way tie for the Ivy title.
If we loose at Dartmouth, well, this gets sticky... we're only up by two on them, 4 on Harvard - each of them having two games left. Harvard can at best tie us, Dartmouth can pass us w/ 3 or more pts vs. Yale/Princeton, tie us w/ 2 or we can still win the Ivies if they only get one. Either way, Ivy title wise, it's big that we at least don't loose next Friday.
Oh, did I mention that Harvard's remaining two Ivy games are also Yale/Princeton... those guys get around
. And as a final, obvious, note: Yale, Princeton, and Brown have no chance at the title.Either way, Let's Go Red!
Re: Ivy Magic Numbers
Posted by: KeithK (---.dsl.snfc21.pacbell.net)
Date: February 10, 2002 02:37AM
Well, it's a lot easier looking at Ivy standings because the pesky tiebreaker problems go away since the Ivies don't bother breaking ties. Adn when there are only 7 games and 3^7 = 2187 possibilities left it's real easy to slap together a simple spreadsheet and Perl script to compute possibilities.
My attempts at the ECAC tiebreakers haven't worked yet - I can never seem to wrap my mind around coding all of the possibilities, let alone handling the number of cases. And flooding slack.net with 3^24 possibilities via Perl script would probablynot be appreciated
.
My attempts at the ECAC tiebreakers haven't worked yet - I can never seem to wrap my mind around coding all of the possibilities, let alone handling the number of cases. And flooding slack.net with 3^24 possibilities via Perl script would probablynot be appreciated
.
Re: Ivy Magic Numbers
Posted by: Robb (---.56.120.14.dial1.dallas1.level3.net)
Date: February 10, 2002 07:22PM
Yeah - I'm not even sure if the ECAC tiebreakers are fully defined. I seem to remember that there can be cases that are ambiguous and/or definitely unresolvable with the official tie-breaking procedure. Thank goodness we've never seen one arise. Which is a worse system: the electoral college or the ECAC tiebreakers? You make the call!
Re: Ivy Title
Posted by: DeltaOne81 '03 (---.resnet.cornell.edu)
Date: February 10, 2002 09:58PM
I emailed the ECAC about this about a year ago actually. Here was the response:
Thank you for your inquiry and interest in the ECAC.
The tie-breaking procedures for Division I men's ice hockey are:
In the event that there are ties in the final regular-season standings, the
following tie-breaking procedures will be used to seed teams for the
championship:
1. Comparison of game results between tied teams (head to head)
2. Comparison of results of games against the top five teams
3. Comparison of results of games against the top 10 teams
4. Goal differential in head-to-head competition
5. Goal differential in games against the top five teams
6. Goal differential in games against the top 10 teams
If you have any questions, please let me know.
Stephen R. Hagwell
Assistant Commissioner
Eastern College Athletic Conference (ECAC)
Though what I'm not exactly clear on is how this applied to multiple-way ties? How do you do head-to-head in that situation, etc?
Thank you for your inquiry and interest in the ECAC.
The tie-breaking procedures for Division I men's ice hockey are:
In the event that there are ties in the final regular-season standings, the
following tie-breaking procedures will be used to seed teams for the
championship:
1. Comparison of game results between tied teams (head to head)
2. Comparison of results of games against the top five teams
3. Comparison of results of games against the top 10 teams
4. Goal differential in head-to-head competition
5. Goal differential in games against the top five teams
6. Goal differential in games against the top 10 teams
If you have any questions, please let me know.
Stephen R. Hagwell
Assistant Commissioner
Eastern College Athletic Conference (ECAC)
Though what I'm not exactly clear on is how this applied to multiple-way ties? How do you do head-to-head in that situation, etc?
Re: Ivy Title
Posted by: kingpin248 (---.twcny.rr.com)
Date: February 10, 2002 10:06PM
Hopefully the coming change in playoff format will give the ECAC an opportunity to rethink its tiebreaker system. The tiebreakers used to be record against the top four and top eight, then were changed to top five and ten. So what's next - top six and top (or bottom) twelve?
This is what I would like to see:
(1) Head-to-head record.
(2) Wins. (This is actually the first tiebreaker in the NHL, before H2H.)
(3) Goal difference in conference games.
(4) Goals scored in conference games.
I'm not sure the ECAC would go for (3) or (4), as it might be seen to encourage running up scores. Also, there's a problem with using wins as a tiebreaker - it follows that when comparing two teams with equal points and equal games played, a team with more wins will also have more losses.
And besides, the conference doesn't exactly have a track record of sound thinking on these matters, does it?
This is what I would like to see:
(1) Head-to-head record.
(2) Wins. (This is actually the first tiebreaker in the NHL, before H2H.)
(3) Goal difference in conference games.
(4) Goals scored in conference games.
I'm not sure the ECAC would go for (3) or (4), as it might be seen to encourage running up scores. Also, there's a problem with using wins as a tiebreaker - it follows that when comparing two teams with equal points and equal games played, a team with more wins will also have more losses.
And besides, the conference doesn't exactly have a track record of sound thinking on these matters, does it?
Re: Ivy Title
Posted by: KeithK (---.dsl.snfc21.pacbell.net)
Date: February 10, 2002 10:21PM
The issue with the ECAC is not what the tiebreakers are, it's how to apply them in crazy situations. For instance, in a four way tie, if in H2H tiebreaker Team A has 7 points, Team D has 5 and Teams B&C have 6 each, does Team A win the tiebreaker and recheck from the beginning with A-D? Does D lose and have A-C start over? Or does A win, D lose and B&C start over?
And what do you do if there's a tie for 5th which depends on Top-10 record and a tie for 10th which depends on Top-5? This has been explained before (Whelan? Bill Fenwick?) but it's not exactly a clean system. I [I}think[\I] there is a completely deterministic way of doing the tiebreakers but I've never been able to code it.
I would assume that the ECAC will go back to Top-4 and Top-8 tiebreakers with the new playoff system next year.
And what do you do if there's a tie for 5th which depends on Top-10 record and a tie for 10th which depends on Top-5? This has been explained before (Whelan? Bill Fenwick?) but it's not exactly a clean system. I [I}think[\I] there is a completely deterministic way of doing the tiebreakers but I've never been able to code it.
I would assume that the ECAC will go back to Top-4 and Top-8 tiebreakers with the new playoff system next year.
ECAC Tiebreakers
Posted by: jtwcornell91 (---.utb.edu)
Date: February 11, 2002 12:53AM
The ECAC did resolve the "infinite loop" situation a few years ago, and it was on the website back when Tim Danehy was running it and it was called ECAC HockeyNet. Here it is (and yes, this is implemented in my ECAC playoff possibilties scripts):
Actually, when you think about it, it's the most sensible and straightforward solution to the problem.
If there is a tie involving 5th/6th that requires application of the third tie-breaker, record against Top 10, and there is also a tie involving 10th/11th that requires application of the second tie-breaker, record against Top 5, then the following procedure shall be used: The tie involving 10th/11th shall be broken using record against all teams above and included in the tie involving 5th/6th. If this fails the tie involving 5th/6th shall be broken using record against all teams above and included in the tie involving 10th/11th. If this also fails the tie involving 10th/11th shall be broken using record against all teams above the tie involving 10th/11th. If this also fails the tie involving 5th/6th shall be broken using head-to-head goal differential. If this also fails the tie involving 10th/11th shall be broken using head-to-head goal differential. If this also fails the tie involving 5th/6th shall be broken using goal differential against all teams above the tie involving 5th/6th. If this also fails the tie involving 10th/11th shall be broken using goal differential against all teams above and included in the tie involving 5th/6th. If this also fails the tie involving 5th/6th shall be broken using goal differential against all teams above and included in the tie for 10th/11th. If this also fails the tie involving 10th/11th shall be broken using goal differential against all teams above the tie involving 10th/11th. Again once one or more teams are separated from any others the tie-breakers are re-applied in each group of remaining tied teams starting with head-to-head record.
Actually, when you think about it, it's the most sensible and straightforward solution to the problem.
Sorry, only registered users may post in this forum.