29 Nov 2004 19:22:33
George W. Harris
Unbalanced Schedules And Tiebreakers

How should tiebreakers work in conferences
where a complete double round-robin isn't played (if
there aren't divisions), or all teams in a division don't play
identical schedules?

Traditionally, when two (or more) teams tie in
the standings you first look at their head-to-head records,
and then look at their records against subsets of the
entire conference. While this may not always make
sense (why should the team that lost to Michigan State
lose the tiebreaker to the team that lost to Northwestern?),
it's at least fair, because the teams played the same
games. Not any more.

Suppose, for instance, that UNC and Wake Forest
tied for first in the ACC regular season. This year, they only
play each other once, at Wake Forest. This means that it
would be impossible for the two to split, since they don't
play home-and-home. That seems a little unfair.

But it can be even worse. Suppose NC State and
Wake finished tied for first in the regular season, having
split their regular season games, and UNC finished third.
Wake played UNC at home, winning, while NC State beat
UNC in the ESA while losing in the Dean Dome. NC State
loses the tiebreaker because they lost a game that Wake
didn't even have to play!

How should it work? What seems fair to me is you
just look at each team's *schedule*. In an 11-team
conference with a 16-game schedule, each team will play
six teams twice and four teams once. You just add the
conference records of the six teams Wake played twice,
and then add the conference records of the six teams NC
State played twice, and whichever was tougher, that team
wins the tiebreaker.

--
Doesn't the fact that there are *exactly* fifty states seem a little suspicious?

George W. Harris For actual email address, replace each 'u' with an 'i'


29 Nov 2004 14:39:13
Drew
Re: Unbalanced Schedules And Tiebreakers

"George W. Harris" <gharrus@mundsprung.com > wrote in message news:ursmq0hni04gqq2q7b897u1ut7g3b509oe@4ax.com...
>
> How should it work? What seems fair to me is you
> just look at each team's *schedule*. In an 11-team
> conference with a 16-game schedule, each team will play
> six teams twice and four teams once. You just add the
> conference records of the six teams Wake played twice,
> and then add the conference records of the six teams NC
> State played twice, and whichever was tougher, that team
> wins the tiebreaker.

That still leaves you open to the scenario where, for example,
Wake Forest played Duke twice and split, and Clemson
once, but Georgia Tech played Duke once, and Clemson
twice, and swept Clemson. No two teams play the same
set of 6 teams twice, so any such comparison has to favor
one team over the other.

I don't necessarily have a better idea, mind you. I've been
aggravated myself by teams that have managed to finish atop
the Big Ten by virtue of not playing the other top teams on
the road (yes, Wisconsin, I'm talking about you). For me,
such scheduling imbalance is about the only reason I like
having a conference's automatic bid go to the winner of its
tournament. At least everybody has a shot on a neutral
court, even if some brackets are easier than others. You
still have the chance to put up or shut up.

Drew




29 Nov 2004 20:30:05
Donnie Barnes
Re: Unbalanced Schedules And Tiebreakers

On Mon, 29 Nov, George W Harris wrote:
> How should tiebreakers work in conferences
> where a complete double round-robin isn't played (if
> there aren't divisions), or all teams in a division don't play
> identical schedules?

Isn't it obvious? Why, the MLB way, of course! I'm talking about TV
revenue here. The "league office" just matches up whoever will draw the
highest ratings. Makes perfect sense in the new football conference we
voluntarily joined.


--Donnie

--
Donnie Barnes http://www.donniebarnes.com 879. V.


29 Nov 2004 14:46:55
Drew
Re: Unbalanced Schedules And Tiebreakers


"Drew" <yeah@right.not > wrote in message news:cog1dp$dfr@netnews.proxy.lucent.com...
> "George W. Harris" <gharrus@mundsprung.com> wrote in message news:ursmq0hni04gqq2q7b897u1ut7g3b509oe@4ax.com...
> >
> > How should it work? What seems fair to me is you
> > just look at each team's *schedule*. In an 11-team
> > conference with a 16-game schedule, each team will play
> > six teams twice and four teams once. You just add the
> > conference records of the six teams Wake played twice,
> > and then add the conference records of the six teams NC
> > State played twice, and whichever was tougher, that team
> > wins the tiebreaker.
>
> That still leaves you open to the scenario where, for example,
> Wake Forest played Duke twice and split, and Clemson
> once, but Georgia Tech played Duke once, and Clemson
> twice, and swept Clemson. No two teams play the same
> set of 6 teams twice, so any such comparison has to favor
> one team over the other.

Let me clarify what I'm saying here. I didn't express that well.
Yes, in the above scenario (assuming equal win-loss), Wake
would rightly get the edge as Duke's conference records would
presumably be better than Clemson's. But, that would require
them to need your tie-breaker. To get to the point of needing
that tiebreaker, though, Wake has to do as well against Duke
as GT did against Clemson (in the simple case where their
other 5 two game opponents, and results, were the same). And
requiring Wake to do as well against Duke as GT must do against
Clemson is not fair, even if the tiebreaker used in that case would
then favor Wake.

Hopefully that makes more sense.

Drew




29 Nov 2004 21:20:54
George W. Harris
Re: Unbalanced Schedules And Tiebreakers

"Drew" <yeah@right.not > wrote:

:
:"Drew" <yeah@right.not > wrote in message news:cog1dp$dfr@netnews.proxy.lucent.com...
: > "George W. Harris" <gharrus@mundsprung.com> wrote in message news:ursmq0hni04gqq2q7b897u1ut7g3b509oe@4ax.com...
: > >
: > > How should it work? What seems fair to me is you
: > > just look at each team's *schedule*. In an 11-team
: > > conference with a 16-game schedule, each team will play
: > > six teams twice and four teams once. You just add the
: > > conference records of the six teams Wake played twice,
: > > and then add the conference records of the six teams NC
: > > State played twice, and whichever was tougher, that team
: > > wins the tiebreaker.
: >
: > That still leaves you open to the scenario where, for example,
: > Wake Forest played Duke twice and split, and Clemson
: > once, but Georgia Tech played Duke once, and Clemson
: > twice, and swept Clemson. No two teams play the same
: > set of 6 teams twice, so any such comparison has to favor
: > one team over the other.
:
:Let me clarify what I'm saying here.

Please (IDHTPASH,DI?)

:I didn't express that well.
:Yes, in the above scenario (assuming equal win-loss), Wake
:would rightly get the edge as Duke's conference records would
:presumably be better than Clemson's. But, that would require
:them to need your tie-breaker. To get to the point of needing
:that tiebreaker, though, Wake has to do as well against Duke
:as GT did against Clemson (in the simple case where their
:other 5 two game opponents, and results, were the same). And
:requiring Wake to do as well against Duke as GT must do against
:Clemson is not fair, even if the tiebreaker used in that case would
:then favor Wake.
:
:Hopefully that makes more sense.

Yes, unbalanced schedules are inherently
unfair, and hence anathema to college basketball
fans everywhere I'm addressing the issue of making
tiebreakers more reasonable.

:Drew
:

--
They say there's air in your lungs that's been there for years.

George W. Harris For actual email address, replace each 'u' with an 'i'.


29 Nov 2004 17:03:52
=?ISO-8859-1?Q?=BBQ=AB?=
Re: Unbalanced Schedules And Tiebreakers

Donnie Barnes <djbSPAMSUCKS@donniebarnes.com > wrote in
<news:slrncqn1kq.g75.djbSPAMSUCKS@localhost.localdomain >:

> On Mon, 29 Nov, George W Harris wrote:
>> How should tiebreakers work in conferences
>> where a complete double round-robin isn't played (if
>> there aren't divisions), or all teams in a division don't play
>> identical schedules?
>
> Isn't it obvious? Why, the MLB way, of course! I'm talking about
> TV revenue here. The "league office" just matches up whoever will
> draw the highest ratings. Makes perfect sense in the new football
> conference we voluntarily joined.

Oh, yes, please. Can we have 'pods' for the tournament, too? With all
these teams, we should be able to fill more than one venue.

--
Q


30 Nov 2004 01:12:06
Charlie Board
Re: Unbalanced Schedules And Tiebreakers

George W. Harris wrote:

> How should it work?

Reverse alphabetically, of course.



29 Nov 2004 21:04:33
Edward M. Kennedy
Re: Unbalanced Schedules And Tiebreakers

"Q" <boxcars@gmx.net > wrote:

> > On Mon, 29 Nov, George W Harris wrote:
> >> How should tiebreakers work in conferences
> >> where a complete double round-robin isn't played (if
> >> there aren't divisions), or all teams in a division don't play
> >> identical schedules?
> >
> > Isn't it obvious? Why, the MLB way, of course! I'm talking about
> > TV revenue here. The "league office" just matches up whoever will
> > draw the highest ratings. Makes perfect sense in the new football
> > conference we voluntarily joined.
>
> Oh, yes, please. Can we have 'pods' for the tournament, too? With all
> these teams, we should be able to fill more than one venue.

I think it should be an 8 team tournament and the last 3,
next year four, teams go can screw themselves. This makes
the regular season count for more. Also, the bottom seeds
don't really have much shot at four straight wins in four days.
Since teams either play each other once or twice, a total
points standard would be a fair tiebreak but would encourage
RUTSing, which is allegedly bad. (I think you pay a decent
price for RUTSing, instead of letting the practice team play,
m'self -- more chance of an injury to a good player, and the
morale of the necessary but gloryless practice squad plummets)

Therefore, after head-to-head, I'd go with the best record
against the best team each played twice, then the best record
against the second best...if that ties out, repeat for the teams
each played once. If that ties out you can flip a coin, or go
for the point total -- if the point total is that far down the
ladder, RUTSing isn't a problem.

--Tedward