[EM] Dave reply, 4/13/12

[Michael Ossipoff]

Dave:

You wrote:

On Apr 12, 2012, at 6:47 PM, Michael Ossipoff wrote:

> I said that Plurality only lets you rate one candidate. That isn't  
> true. You're still rating all of the
> candidates in Plurality, but you're required to bottom-rate all but  
> one of them.

Looking ahead, Plurality lets the voter present a small amount of  
information; Approval a bit more; and Condorcet additional - each such  
as the previous methods do not permit.

[endquote]

As I was saying, it isn't that Plurality allows less information than does Approval.
It's just that Plurality requires a lot of _false_ information. Most of the information
on many or most Plurality ballots is false. You could say, however, that Plurality allows
less _genuine_ information.

Forced falsity has no place in a democracy's voting system.


You continued:

Agreed that Approval allows approving more than one, and that each  
approved is preferred over each unapproved, just as the one approved  
in Plurality is preferred over all others.

[endquote]

That's the myth. The myth that voters in Plurality are voting for their favorite.
Millions aren't, of course. If Plurality's justification is based on the assumption that
votes are for genuine favorites, then Plurality is based on a false assumption. An assumption
as false as most of the information on a Plurality ballot.


I'd said:

> That can't be said for Condorcet or Kemeny, or any other rank method  
> or complex method.

You replied:

Now it is time to be more careful.

In Condorcet if I give one rank to all I prefer I have given the same  
preference to those ranked over those unranked as I could do with  
Approval's approving.

But ability to use multiple ranks in Condorcet or Kemeny gives me  
additional power - among the ranked candidates my preferences can be  
unequal and I show this by ranking higher each that I prefer over  
other ranked candidates.

[endquote]

When "That can't be said for Condorcet or Kemmeny, or any other rank method
or complex method", I was referring to my statement that Approval is quite
obviously an improvement on Plurality, and only an improvement. That can be said
for Approval, due to Approval's extreme elegant simplicity, and the fact that it's
a minimal change from Plurality, one small freedom-modification.

I stand by my statement that that can't be said for Condorcet, Kemmeny, or any other
rank method or complex method. 

No doubt, Condorcet, Kemmeny, MJ, etc., are improvements on Plurality. You know that. 
I know that. Nearly no one knows that.

An elaborate contraption like Condorcet or Kemmeny will be viewed as likely to have
unforseen consequences--as, in fact, rank methods do tend to have. People won't know
if it's really an improvement on Plurality, or whether, instead, it will bring some
dreadful problem that will create disaster. 

Media, opponents and corrupt politicians will, of course pick that up and run with it.
They're sure to say, "That will require a lot more study".  Translation: It will never
be enacted.

You wrote:

Condorcet perhaps should be described as a family of election methods,  
usually agreeing as to details such as winner chosen - such as Kemeny.

[endquote]

No perhaps about it: Condorcet is a family of methods. Condorcet(wv), too is a family, or
maybe a subfamily, of methods. When I introduced the Condorcet(wv) subfamily of methods,
it didn't occur to me to call it "Ossipoff's Method", unlike some other method-introducers.

>
>
> I don't  know anything about Kemeny's properties, and I was just  
> asking what it does with the
> 2nd set of rankings in my previous posting, and whether or not it  
> passes FBC. I don't claim to
> know Kemeny's properties.

You continued:

"2nd set" implies misunderstanding - in the Condorcet family voters  
are normally permitted  to use more than two rank values.

[endquote]

I said "the 2nd set of rankings in my previous posting" (or something close to that and meaning
the same thing).

I was referring to this set of rankings:

27: A>B
24: B
49: C


You continued:

"FBC" is simply one of many acronyms for which definitions are hard to  
find (and to verify having found correctly).

[endquote]

FBC is very well-known. It stands for "Favorite-Betrayal Criterion", though
it could just as well be regarded as "Favorite-Burial Criterion.

FBC says that there should never be incentive to bury one's favorite (vote someone
else over hir).

In a little more detail, it says that under no circumstances should burying one's favorite
give a result such that you can't get as good a result without burying your favorite. (given
some particular configuration of candidates and other voters' votes)

For the purposes in the above definition, the one result 1 is more good than result 2 if 
you prefer result 1 to result 2. In other words, you prefer the candidate elected in result 1
to the candidate elected in result 2.

[end of FBC definition]

FBC has been used for a long time, and not just by me. It's common knowledge here what FBC
is.

To give more precisely-worded statement of the more detailed definition given above:


A method meets FBC if it isn't possible to contrive a configuration of candidates, voters,
and the votes of voters other than you, such that you, by burying your favorite, can get
a result better than any that you could get without burying your favorite.

For the purposes of the above definition, saying that result 1 is better (to you) than result 2
means that you prefer the candidate elected in result 1 to the candidate elected in result 2.

[end of definition of FBC)

Mike Ossipoff





  		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20120413/9474ca17/attachment-0003.htm>