Briefly why I prefer Reg-Champ to Smith//Con

[Mike Ossipoff]

Bruce Anderson writes:
> 
> On May 27,  5:00pm, Steve Eppley wrote:
> > Subject: Re: Attachment CRITTBL1
> > Bruce, you've provided a table of some ranked ballot voting methods
> > which meet or fail to meet 7 criteria.  Quite a few of the methods
> > meet all 7 criteria, including Smith//Condorcet.  
> > 
> > You've also written that you prefer Regular-Champion to
> > Smith//Condorcet, but I don't recall you explaining why. 
> > 
> > Since we're probably going to poll ourselves soon on methods, I'd
> > like to hear your reasoning; otherwise I feel compelled based on 
> > the discussion so far to rank Smith//Condorcet higher than
> > Regular-Champion. 
> > 
> > I presume that you're planning to post definitions of the methods 
> > mentioned in the table.  I'm looking forward to that, because many of 
> > the methods' names are unfamiliar to me and because I'm curious about 
> > how Condorcet is defined (and by whom).
> > 
> > ---Steve     (Steve Eppley    seppley at alumni.caltech.edu)
> > 
> >-- End of excerpt from Steve Eppley
> 
> If you, or others, would like more details on any of the voting methods I 
> recently described, please let me know which ones, and what kind of additional 
> details you would like to see.  Otherwise, I will assume that the descriptions 
> are adequate for your purposes.
> 
> I have not get posted why I prefer Regular-Champion to Smith//Condorcet.  
> Briefly, the reasons are as follows.
> 
> First, from what I have seen so far, it seems to me that the strategy-protection 
> capabilities of Condorcet and Smith//Condorcet have been vastly oversold.  I 
> have seen no precise statements of such strategy-protection criteria on this 

N> list (maybe they were posted before I joined), and have seen no proofs of such 
> criteria anywhere.  When I tried to formulate and prove similar criteria, I 
> found severe weaknesses in this supposed protection.  I have not saying that 
> such criteria could not be precisely stated and formally proven.  I'm just 
> saying them I have seen no proofs at all, and that the statements of such 
> criteria that I have seen had what I considered to be severe flaws.  I will 
> attempt to post an example of such a flaw for the "lesser of two evils" 
> criterion in the near future.
>
That's curious, because I haven't mentioned a "lesser-of-2-evils criterion"
to this list. I've defined a Generalized Majority Criterion. 
So if you're going to find flaws, hopefully you'll do it with something

Bruce has said that he doesn't consider the lesser-of-2-evils
problem to be important. That represents a wide divergence
between Bruce & the electoral reform community. The lesser-of-
2-evils problem is widely agreed among electoral reformers
to be Plurality's worst problem, when comparing it to the
single winner methods that we'd like to replace it with.
that was actually posted to this list. And while you're finding the
flaws in my Generalized Majority Criterion, don't forget to find
the ones in the Invulnerability to Mis-Estimate property.

By the way, my Generalized Majority Criterion is different from
the criterion that Bruce calls by that name, but whose area of
application is much less general. I call that other one the
"Mutual Majority Criterion".

[Some lines in this reply are being duplicated & displaced,
due to a line length mismatch that happens only when I
attempt to reply to Bruce's postings via the "edited-copy"
method of replying]

esides, some of us have already agreed that standards can be
Buseful even if they aren't expressed as criteria, y/n tests.

 
> Second, but less important, Condorcet and Smith//Condorcet can be vulnerable to 
> strategy manipulation in some cases when the methods I prefer (Regular-Champion, 
> Consensus-Champion, Complete-Champion, and Qualified-Champion) are not.  This 
> certainly does not mean that the methods I prefer are necessarily better, but it 
> raises questions concerning exactly what are the meaningful properties that 
> Condorcet and Smith//Condorcet really do possess.  I posted one example of this 
> vulnerability already.  I'll post a second, closely related, example later 

Actually, you didn't. You've posted, at this time of writing, your
second letter about that (which I haven't checked yet), but no
1st letter about that from you has been posted.

But I'll reply to your 2nd one, even though your 1st one hasn't
been posted yet. As I said, I'm replying to this letter first.


> tonight.
> 
> Third, and most important, Condorcet and Smith//Condorcet can select as winners 
> what I believe the vast majority of voters (in economically developed countries) 


As I've repeatedly explained to you, Bruce, any method is going
to sometimes dramatically violate some steandard other than the one
on which it is based. Your attack-strategy of finding examples
of that therefore doesn't make any sense.

It merely depends on what standards we consider important. To
you, "candidate-counting" is the important thing. How many
alternatives does a particular alternative beat, and how many
is it beaten by. Fine, that's what's important to you. But
it isn't the problem whose solution is important to electoral
reformers.

As you know, Fishburn wrote a "bad-example" for Copeland too.


> would think are incredibly ridiculous choices.  Such a selection may rarely 

In fact, you never said how your examples could ever happen, or why
they ever should. Your examples require an alternative to be the
most popular in terms of how many people think something else
is better, but also that it be beaten by more alternatives than
is any other alternative, sometimes, in some of your examples,
beaten hy every one of the other alternatives. As I've said,
Smith//Condorcet meets the criteria that you've posted here.

> happen, but if it happens even once in an election of any significance, it could 
> set the single-winner reform movement back for decades, maybe even a century or 


Y> two.  I don't think it's worth the risk for the minimal strategy protection that 
> has been meaningfully demonstrated so far for Condorcet and Smith//Condorcet.  
>

You keep making statements like that, but you don't reply to specific
statements of mine about the strategy differences between Condorcet
& Copland (including Regular-Champion).


 Personally, I would feel quite comfortable defending any Regular-Champion (or 
> Consensus-Champion, Complete-Champion, or Qualified-Champion) winner before a 
> group of intelligent but hostile critics.  Conversely, I see no way of defending 
> some of the winners that Condorcet and Smith//Condorcet can produce.  I will 
> attempt to post relevant examples shortly.

Yes, I've replied to those examples. You say there's no way of defending
the results that you describe, but I did, and you didn't reply.

> 
> Given the blatant losers that Condorcet and Smith//Condorcet can select as 

Translation: Blatant losers, for Bruce, means alternatives that
don't do well according to _his_ candidate-counting standard--
how many alternatives a particular alternative beats or ils
beaten by.


> "winners," the choice between Condorcet and Smith//Condorcet and methods that 
> cannot select such intuitive losers seems easy to me.  The harder question is to 
> choose among those other methods.  In particular, choosing among  
> Regular-Champion Consensus-Champion, Complete-Champion, Qualified-Champion, 
> Qualified-Kemeny, Kemeny//Random, Niemi-Riker//Random, Smith//Borda//Random, and 
> Smith//Bucklin-ext//Random is not so easy.  I based my choice on my which I felt 
> was the simplest.


So the hard thing for you, then, is choosing from among your methods
that can elect someone with a majority against him when no one
else has a majority against him.

Your methods that are vulnerable to truncation, and which will
often require voters to insincerely vote a lower choice equal
to or over their favorite.

Methods in which a mis-perception among voters that a certain alternative
is a middle Condorcet winner can give the election away to that
alternative.

Methods whose outcome, in terms of which party or faction wins,
will depend on how many candidates the various parties or
factions can afford to run.


Mike

> 
> Bruce
> .-
> 


--