[EM] SU - some understanding

[Dave Ketchum]

Warren's post seems to bring some understanding - but NOT much happiness:

I comment on this post below, but my serious topic is his claim that 
Condorcet rates below Approval as to SU.

Not interested in debating SU numbers, I question why anyone should care:

Provided Condorcet voting restricts itself to '=' and omits '>'  and '<', 
it can vote for EXACTLY the same candidates with EXACTLY the same effort 
and result in EXACTLY the same winner as Approval.

For this, seems like it deserves EXACTLY the same rating as to ability.

But Condorcet offers voters a chance to vote such as Nader>Gore, something 
that is OFTEN desired and can SOMETIMES make a difference - reasons for 
rating it above Approval in whatever measures usefully compare election 
methods.

On Fri, 23 Feb 2007 14:44:39 -0500 Warren Smith wrote:
      Re: [EM] CW & SU (response to Ossipoff false claims)
> I suggest to  Mike Ossipoff that after I point out to him that
> I refuted his false claims in a web page
>     http://rangevoting.org/BlackSingle.html#condmyth
> he at least consult that web page BEFORE making claims that
> have already been disproved there.

First topic I see there is some voters whose median location defines how 
they would have to vote to maximize SU - BUT - there is nothing 
controlling from their locations that defines how they WOULD CHOOSE to vote.
> 
> I already gave examples in 1D where the Condorcet Winner - who in that
> example is the median voter - is not best in terms of social utility.
> 
> I repeat.  Essentially every claim Ossipoff ever made about
> Condorcet winners vis a vis social utility and median voters, is false.
> Now while that was bad, it is worse to continue to make
> already-refuted claims.
> 
> Just to make it ultra-easy for Ossipoff, I have now
> made a concise "executive summary" webpage since I realize actually reading 2 full pages is
> beyond him:
>    http://rangevoting.org/BlackSExec.html

Topic of the first page here is impossibility of Condorcet cycles.

Having been around long enough to know that cycles need responses more 
serious than pretending they cannot exist, it is time to look closer:
      The theorem starts with "If" so, if its conditions are not all true, 
there is still room to admit possible existence of cycles.

DWK
> 
> All those things are proved on the web page I cited that Ossipoff keeps not reading.
> 
> And no *I* did NOT say academic authors had their "heads up their ass."  Ossipoff did.
> I merely pointed out that they usually publish theorems that are correct,
> unlike Ossipoff.  Try not to ascribe your opinions to me.
> 
> I see he now has retreated to the vague stance that
> "At least one can say  that, in spatial voting,
> the wv Condorcet methods will tend to do well by
> SU, even if one can't make stronger guarantees."
> 
> "Tend to do well"?  Pretty pathetic, isn't it.     Guess we could
> say that about every voting method, couldn't we.  But anyhow, rather than relying on
> vague assurances, we can actually MEASURE how well they do.  That is what my
> simulator program IEVS does.  Although rank-equalities in votes have
> not yet been implemented in IEVS, that makes no difference if we are only
> discussing honest voters.  IEVS usually finds that Condorcet methods are midpack
> in terms of SU (social utility).
> 
> The high-SU methods include range voting, Brian Olson's IRNR system, and range+top2runoff,
> and Benham's MCA system (a 3-slot approval-like method).
> 
> Approval Voting using "zero info strategy" is found in net in current
> IEVS scenarios to have better SU than
> every honest-voter-Condorcet method tried (and IEVS has about 20 of them).
> 
> Warren D. Smith
> http://rangevoting.org
-- 
  davek at clarityconnect.com    people.clarityconnect.com/webpages3/davek
  Dave Ketchum   108 Halstead Ave, Owego, NY  13827-1708   607-687-5026
            Do to no one what you would not want done to you.
                  If you want peace, work for justice.