Academic Criteria & Condorcet's Method (fwd)

[Mike Ossipoff]

Mike Ossipoff writes:
> From mail.eskimo.com!eskimo.com!election-methods-list Mon Apr 15 05:42:44 1996
> Date: Mon, 15 Apr 1996 05:41:36 -0700 (PDT)
Now that Bruce has posted his "faults of Condorcet" article to
EM, it seems necessary to re-post some letters I posted in reply.
These letters discuss Bruce's claim that it's serious that plain
Condorcet fails the criteria that he names. As has been said,
Smith//Condorcet meets those criteria.

Again, the fact that this letter has 2 headers is the result
of the fact that I don't have a way to delete lines or blocs
of text. Or if I do have a way to, I haven't found it yet.

> Message-Id:  <9604150008.aa26831 at bbs.cruzio.com>
> Reply-To: election-methods-list at eskimo.com
> Originator: election-methods-list at eskimo.com
> Sender: election-methods-list at eskimo.com
> Precedence: bulk
> From: Mike Ossipoff <dfb at bbs.cruzio.com>
> To: Multiple recipients of list <election-methods-list at eskimo.com>
> Subject: Academic Criteria & Condorcet's Method
> X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
> 
> The criticisms against Condorcet's method are mostly the criticisms about
> how plain Condorcet doesn't meet the academic candidate-counting criteria.
> 
> Of course Smith//Condorcet meets those criteria, & every objective
> criterion that I've heard of. It will be interesting to hear what
> desperate arguments anti-reform types will come up with to criticize
> Smith//Condorcet.
> 
> And Demorep demonstrated how natural a Condorcet Loser Criterion rule
> fits into a set of Condorcet's method rules--right after where it says
> that any alternative that beats each of the others is elected, it could
> say that any alternative beaten by each of the others is disqualified.
> 
> ***
> 
> But this message will just be about plain Condorcet & the criticisms
> against it, based on the academic candidate-counting criteria.
> 
> Actually, this message will just discuss the Condorcet Loser Criterion,
> and the other candidate-counting criteria will be discussed in subsequent
> messages, each having a title based on the criterion that it discusses.
> For example, the next message from me about these criteria will be
> entitled "The Smith Criterion & Condorcet's method". It will probably
> be quite short, since this letter will cover what there is to be
> said about the Smith Criterion, which is really like a milder version
> of the Condorcet Loser Crierion--milder in the sense that violating it
> is a less dramatic violation. Of course, in another sense, it's a stronger
> criterion, since any method meeting it automatically meets Condorcet Loser,
> and since Smith demands more than Condorcet Loser does.
> 
> ***
> 
> Though I don't object to complying with the Condorcet Loser Criterion,
> either by a Condorcet Loser Criterion rule, or by a Smith Criterion rule,
> or a Smith set rule, I don't consider it important.
> 
> I've talked about the widely-held standards met only by Condorcet's
> method. Compared to that, the CL Criterion is just cosmetic--a rule
> to avoid an embarrassment in a hopeless bottom-end situation where
> there's no right answer anyway, and where, if plain Condorcet picks the
> CL, it's because everything other than the CL has more people voting 
> for something else against it than the CL does. Does that sound like
> a situation where the result matters, or a situation where there's
> a compelling case for not picking, as winner, the CL? Can anyone who
> uses the CL Criterion against plain Condorcet seriously claim that that's
> important?
> 
> When plain Condorcet picks as winner the CL, that's a peculiarly, 
> suspiciously, un-disliked Condorcet Loser. I mean, if it's a really
> unpopular alternative, then how come it has the fewest voters saying
> that a particular other alternative is better than it?? Bruce's
> bad-examples, and the text that accompanies them seem to be saying
> that the CL that plain Condorcet could pick is a really bad alternative.
> So bad that it has fewest people saying something else is better?
> 
> Such bad-examples are contrived, and make no sense, because they
> show you a self-contradictory kind of voting, an electorate that
> contradicts itself, and they give no explanation for how or why
> that would happen.
> 
> Summary:
> 
> The CL Criterion is unimportant for at least 3 reasons:
> 
> 1. It's nothing but a rule to avoid embarrassment in a hopeless
> bottom-end situation where there's no satisfactory outcome anyway,
> & the result doesn't really matter. And if everything other than
> the CL has a bigger vote against it than the CL does, then it matters
> even less.
> 
> 2. How bad can a CL be if it has fewest people saying that something
> else is better?
> 
> 3. The plain Condorcet CL bad-examples are contrived self-contradictory
> voting examples, where voters are saying 2 contradictory things, and
> for which the author of the bad-examples has given no explanation.
> 
> ***
> 
> More to follow about other academic criteria.
> 
> ***
> 
> Mike
> 
> 
> 
> 
> 
> 
> 
> -- 
> .-
> 


--