[ER] Electoral College reform

[Mike Ossipoff]

donald at mich.com writes:

[Mike writes: I'm replying to most of this letter, except for some
paragraphs near the beginning. So my reply begins a little way down
in the letter]




> 
> 10-10-96 Steve wrote:
> >
> >That was me, and since it was a "technical" proposal I posted it in
> >the EM list (election-methods-list at eskimo.com), not this list.
> 
> Don: Good - someone knows the proper place to post technical stuff.
> 
> >I'd suggest that ER folks interested in discussing these details
> >further subscribe to EM and direct their replies there.
> 
> Don: Yes - listen up all you guys and pay attention.(he means you) oops.
> All right already! I will take your suggestion Steve - but first listen to
> my tale of woe. (listen to your excuses)
> 
> One: I posted to [ER] by mistake.(won't wash - that is the same excuse
> Steve made to you when Kevin posted in the [ER] list.)
> 
> Two: I followed the subject on [EM] then filed my report on the subject to
> the [ER] list.(he's not going to buy that)
> 
> Three: I held up the report until there was a slow day on the list. Tuesday
> was a Slow News Day on the [ER] list. (nobody should have been overwhelmed
> by the two messages on that morning)
> 
> Four: I'm afraid to post on the [EM] list because I fear that the subject
> will flounder there and never come out. I understand that at this very

In other words, you're afraid that what you're saying can't hold up
in a discussion. So you just dump it directly on ER. But it isn't
just a question of your arguments holding up--it's worse than that.
The things that you're saying are obvious misquotations, showing
that you refuse to take the trouble to carefully read what you're
replying to, before writing your reply. As this letter to which I'm
replying exemplifies, you're wasting our time., because you won't
do us the respect of sending us _considered_ replies, instead of
careless, sloppy mis-quotations of what we've said.

Lest anyone think that, by saying that you're wasting our time,
I'm trying to usurp the list-owner's authority, let me emphasize that
I said that as one member to another, and that I have no authority
to tell anyone what they can or can't say on this list, or to set
its standards for discussion. But a member has a right to speak as
a member, and that comment seemed necessary.

It isn't that you're wasting our time because I disagree with what
you're arguing for. It just that you're replying so carelessly that
the things you quote us as saying bear no relationship to what we've
said. Get your act together (said as one member to another). It isn't
just one statement, or even one posting that has that fault. Incredibly,
it's many of the statements in pretty much all of your recent postings.

But it isn't for me to tell you what to say or not say. But it would
be a dis-service to you not to let you know what you've been doing.
So "get your act together" was merely a friendly suggestion.

> point in time there is a big report lost somewhere in the [EM] list rolling
> around like a loose cannon. If that ever comes out it'll be like dropping a

If only it _were_ loose. If anything, it's held too tight, and we're
erring in the side of caution in a conscientious effort to ensure
that the posting of the report to ER is democratically done.

> bomb on the [ER] list. (You're afraid those technical guys on [EM] will eat
> you alive.)

In what sense would it be like dropping a bomb on ER? Do you mean
dropping a bomb on Instant Runoff? :-)

Does your question in parentheses mean that you've posted about this
to ER again? I though we'd all agreed that this debate was more
suited for EM.

If ER is receiving this, I assure you that it isn't anyone's
intention to eat you alive, or to swamp you with technical discussion.
The technical discussion is confined to EM. What you on ER will receive
is a concise summary of our recommendation, and of our vote on
that recommendation. Later that will be followed by a posting that
will combine various people's statements, arranged orderly. That,
we feel, is preferable to sending you an ongoing debate.

> 
> Five: I have not been feeling well lately. (You never felt better - you
> have no more excuses.)
> 
>   NO MORE EXCUSES!! - You Mean! - (Yes - you must post this on the [EM] list)
> 
>                god have mercy on all of us
> 
> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
> Steve Wrote:
> 
> Steve:>                  - the algorithm which awards delegates must
> >take into account the voting in other states.
> >
> Don: Very good - I agree. But the question is how is this to be done?
> Instant Runoff Method of single winner can handle this very well. Instant

The single-winner method used by the reform states to pick its favorite
doesn't have to be what takes into account national winnability.
You want to do that with a sloppy hodge-podge that tries to combine
the reform states' ranking election count with some (poor) measure
of national winnability.

National winnability is a simple thing: Can the reform states' favorite,
or can he/she not, win nationally. That requires being able to get
a majority of the electoral college votes. Simple as that. So that's
what Steve's proposal goes by. That has no connection, nor should it,
with the issue of what method is used when the reform states count
their rankings to determine their favorite.

You, Don, are making us say everything about 4 times. Why don't
you just read it more carefully the 1st time? Maybe it would be
a good idea to start a spinoff list from EM: em-repetition at eskimo.com
Again, that's just a suggestion from a member, and I don't claim to
be saying it from a position of authority in the management of this
list.


> Runoff uses the NATIONAL standings of the candidates as the order of
> candidates to be dropped one by one. As of now the other single winner
> methods cannot couple with the national standings.

But the other proposal also uses the NATIONAL standings of the
candidates, as measured by their ability to win, a more relevant
measure than you use. So your use of IRO (Instant Runoff) fails
on both counts: It's a very poor method for the reform states to
use to determine their favorite. And it's also not a measure of
genuine national winnability.

> 
> 
> Steve: >The states are comprised of two groups: the unreformed states and the
> >reformed states.
> >
> >1. All the reformed states would agree to pool their ranked ballots
> >together when tallying, and agree to award all their electoral
> >college delegates as a bloc to the same candidate.  If the total
> >number of delegates from the reformed states is a majority of the
> >electoral college, then the results in the unreformed states
> >wouldn't matter, and the unreformed states would have a strong
> >incentive to quickly adopt the reform to regain their lost clout.
> 
> Don: Steve, so far this is good.
> 
> >2. The candidate who will receive all the reformed states' delegates
> >is the highest finisher, according to the pooled ballots in the
> >reformed states, who can win a majority of the college.
> 
> Don: Steve, you are putting two conditions on the table. One: the candidate
> must finish the highest in the reformed states. Two: The candidate must be
> able to win a majority on the national electoral college. There is a good
> chance that no candidate will meet these two conditions.

Excuse me, Don, but paragraph #2 was right in front of your face when
you wrote that. Take another look at it. It doesn't say that there
has to be someone who a) finishes highest in the reform states; &
b) be able to win an electoral college majority.

No, it says to pick the reform states' highest finisher who is
able to win an electoral college majority. Steve demonstrated that
with an example. If Perot is the highest finisher, but can't win
an electoral majority, and if Clinton is the next highest finisher,
then, if Clinton can win an electoral majority, the reform states
give all their electoral to Clinton, who is their best-finisher among
the candidates who can win an electoral college majority. Take that
to Voltaire & your foil & ask them to explain it to you if it isn't
clear yet. 

> 
> I know what you are trying to do here - you are trying to couple your
> single winner method on the state level with the national standings. And

No you don't know what that proposal is trying to do here. There are
2 separate issues: Which candidates could win nationally. And
in what order do they finish in the reform states' vote. So pick
the highest finishing one that can win. Is that so complicated?



> you are trying to do it by passing the rule: The state winner must be able
> to win a majority on the national electoral college. It is not going to
> work. This rule comes in to effect after you have worked your pairwise
> method on the state's election numbers. The rule cancels out your pairwise
> work and cancels out the election in the reformed states.

More misquotation. Don, this is my suggestion to you,
with the understanding that I have no managerial authority on this
list & speak only for myself: Remain subscribed to ER, but seriously
ask yourself whether there's any reason to go through the motions
of participating here, when you aren't willing to genuinely participate
in the sense of at least reading what we say before you reply to it.

No one has the power to decree that the favorite of the reform
states' voters be able to win an electoral college majority. That
isn't what the proposal says. It merely says that the candidate
to whom the reform states give their electoral votes must be
someone who can win an electoral college majority, and also be the
best reform-statess finisher among those candidates who can win
an electoral college majority.

If no candidate finishing above last in the reform states' count
can win an electoral college majority, then it makes sense to
just give the electoral votes to the first finisher.
 
> 
> You have no right to deny the winner of the reformed states his due of
> electoral votes. He won using your pairwise single winner method. If your
> pairwise method cannot handle this election maybe it cannot handle any
> single winner election.

Sorry--the first finisher of the reform-states' Condorcet count
has no automatic "due". If he can't win nationally, then it wouldn't
be at all pragmatic to give him the electoral votes. And pragmatism
is what the voters want. They don't want to let a greater evil win
because we give our votes to our favorite instead of to the lesser-
evil who could have beaten the greater evil.

If you call giving the reform states' electoral votes to someone
who can't win a failure to "handle this election", then that
failure to "handle the election" has nothing to do with the 
single-winner method used to determine the reform states' favorite.
It's due to our wish to help the best person _who can win_.
You have a peculiar notion of what it means to "handle the
election", but that issue has no relevance to the choice of
single-winner method used to find the reform states' favorite.

We've discussed, at great length, why IRO does a really lousy
job at finding the group's favorite. And your use of it in
your proposal does _not_ determine national winnability.


>
 
> 
> Steve: >
> >All of the reformed states would need to use the same single-winner
> >method to tally the pooled ballots.  The choice of that method is
> >beyond the scope of this message.  However, the subscribers of the EM
> >list recently polled ourselves about which single-winner method(s) to
> >advocate, and you can expect a report on that to be posted here soon.
> >
> Don: Better hold up on that report Steve - you may need to go back to the
> drawing board - pairwise seems not to be able to cut it here in electoral
> reform.

We've already extensively discussed, and then voted on, the relative
merits of Condorcet's method, Smith-Condorcet, & Instant Runoff.
You know what the results were. Why don't you give it up like a
good sport?

The sense in which pairwise doesn't "cut it", as you put it, is
that it doesn't give the reform state's electoral votes to a
candidate who can't win nationally, in a sloppy, arbitrary
bastardization of a lousy sw method with the national electoral
college system.

> 
> 
> Steve: >Example 1:
> >Suppose the only reformed state, so far, is California.
> >Suppose the order of finish in California is:
> >   1.Nader  2.Clinton  3.Dole
> >Suppose that Nader doesn't win any other state, and that Clinton wins
> >at least 270-54 delegates in the other states.
> >Then California's 54 delegates would go to Clinton.
> 
> Don: Think - Think - Think what is happening here in this Example One.
> There is no need to have an election in California - no matter what the

That's a truly asinine statement. The fact that Clinton came in 2nd
in California, along with the fact that he can win, is the reason
why California gives its electoral votes to Clinton. 

> results would be in the state all the electoral votes go to Clinton
> regardless. The California election would have no influence in the final

Let me explain this to you very slowly & carefully: If Dole had finished
second in California instead of Clinton, and could win an electoral
college majority with California's help, then California's electoral
votes would go to Dole rather than to Clinton. It really does have
to do with the California election result. Take time to verify that
before you post those incredibly silly statements.

[To anyone who takes exception to these remarks to Don: 
I don't usually talk that way, but this situation is getting
ridiculous. What do we do? It's necessary to be frank when
a situation like this happens, it seems to me]


> use of the state's electoral votes.
> 
> Steve: >Example 2:
> >Suppose 10 states pass the reform which would bind their tallying
> >together, and that together they have 150 delegates to the electoral
> >college.
> >Suppose the order of finish when the ballots of the 10 states are
> >pooled is:
> >   1.Perot  2.Bush  3.Clinton
> >
> >2.1 Suppose Perot won at least 120 (270-150) delegates in the other
> >40 states.  Then the 150 would be awarded to Perot.
> >
> >2.2 Suppose Perot won less than 120 delegates in the other 40 states.
> >Suppose Bush won at least 120 in the other 40 states.  Then the 150
> >would be awarded to Bush.
> >
> >2.3 Suppose neither Perot nor Bush won at least 120 in the other
> >states, but Clinton did.  Then the 150 would be awarded to Clinton.
> >(I don't know if I like this, since Clinton finished last in the 10
> >states.  There may be ways to improve this.)

Steve, I disagree here. Since no one except the reform states' last
choice is winnable by electoral college majority, then the reform
states should give their electoral votes to their 1st finisher,
Perot. Pragmatism no longer applies, so just give them to the
collective favorite of the reform states' voters.

> 
> Don: Example Two abridged:
>              2.1 Then the 150 would be awarded to Perot.
>              2.2 Then the 150 would be awarded to Bush.
>              2.3 Then the 150 would be awarded to Clinton
> Again let us think: The voters in these three parts of example two would
> have no influence on the awarding of electoral votes. If there were no
> election in California this system would still award the electoral votes
> the same way.
>                Why have them vote? - ( why indeed )


Apparently even a carefully explained example isn't enough to show
you how this works. In 2.1, the reform states' electoral votes went
to Perot because he finished first in the California count. In
2.2, they went to Bush because he finished 2nd, and was the highest
finisher who was winnable. Is it popssible that you really haven't
noticed that yet? If not, then I hope you'll take my suggestion.


> 
> Part 2.4 turns out to be academic - so no need to vote in this part.(4 for 4)
> 
> Don: These are either poor examples or Pairwise cannot cut it.

Excuse me? I wasn't aware that we were talking about Pairwise. (??)


> 
> 
> Steve:>
> >                   - the algorithm which awards delegates must
> >take into account the voting in other states.

It must take into account how electoral college votes have been awarded
in other states, in order to determine whether a particular candidate
could win if he received the reform states' electoral college votes.
Yes, these non-reform states base their electoral vote awardings on
votes, and so then Steve's proposal does take into account voting
in other states. Unlike your proposal, it also takes into account
actual winnability in the national election.

> >
> Don: This is the key to your problem. You have yet to come up with a good
> way to couple state results with the national standings using pairwise

Apparently we disagree on what is "good". But, again, pairwise isn't
mentioned in Steve's proposal, so why do you keep talking about it?
Steve's proposal doesn't name a single-winner method for the 
reform states to use to determine the order in which the candidates
finish in the reform states' election. So your continued references
to pairwise make no sense at all. The proposal indeed does "couple"
national winnability to finishing order in the reform states'
election count (without saying anything about what the sw method
should be). Whether it does that in a way that is "good" depends
on what you want to accomplish. Some of us would like to give the
electoral votes to the best candidate (according to the collective
ranking resulting from the reform states' count) who can win.
I'm sorry if you don't think that's "good".


> elections methods. It is hard for you to use pairwise on the national
> standings because most of the votes in the nation only have one selection.

I don't know that it would be "hard" to do that, but it wouldn't
make any sense to attempt to use the reform state's count method
as you seem to want to, to take into account votes in all the
states.

> But - maybe you should try it anyway to see how that would go - (can't hurt
> to try)  Go back to the drawing board and rethink this.

If votes in the non-reform states were treated as 1-candidate
rankings, in 1 big national Pairwise count, you tell me what the
result would be--you're the one suggesting it. I do know that
wouldn't make any sense to do that. That procedure, like the
other one that you propose, would have no relevance to making
sure that the reform states' electoral votes aren't wasted on
someone who can't win in the national electoral college count.

> 
>              You could always use Instant Runoff - Ha Ha (don't be mean)

I suppose you could always do any of a number of silly things.
Your ignorance is matched only by your arrogance, Don. But then
those 2 things tend to occur together, don't they.

> 
> Donald
> 

Mike Ossipoff

> 
> 
> .-
> 


--