[Election-Methods] RE : Re: Re: rcv ala tournament

[rob brown]

On Dec 30, 2007 3:57 PM, Kevin Venzke <stepjak at yahoo.fr> wrote:

> Rob,
>
> Thanks for answering my questions here. Comments below.
>
> Unfortunately your method didn't do what I was hoping here, which was to
> elect B. Suppose we try this one instead:
>
> 46 A
> 24 B
> 30 C>B


> Would I be correct in guessing that your method picks B now? Because in
> such a case, the 24 B voters actually should not tell the system that they
> have a second preference for C, because the system will just elect C with
> that vote.
>

Actually it picks A, by a thin margin. :)

a 35.15
b 33.21
c 31.64

Also, this would make my point that Condorcet methods have similar
> defection issues to those possessed by Approval.
>

Well, I don't argue that there aren't contrived cases that can allow someone
to vote strategically and gain a slight advantage.  But I really think they
are highly contrived and not so relevant in the real world.

Here's my big unanswered question.  As I mentioned, my favorite explanatory
device is a DSV system that includes "a software agent operating solely on
the voter's behalf, which takes your actual preferences as input and
produces the most strategic approval or range ballot".

Given that, why would a voter have an incentive to mislead the agent?  The
agent is simply casting an approval ballot on the voter's behalf, it is not
revealing the voter's "true preferences" to the agents operating on the
behalf of other voters.

Correct me if I am missing something, but it seems to me that the only way
you would need to mislead the agent regarding your true preferences is if it
is not really operating in your interest.


> You say you don't see much point in discussing various Condorcet methods.


Well, what I meant is that I don't want to get into that debate myself.  I'm
not suggesting others don't.

If I had my way, all the condorcet advocates would hold a vote, and then all
get behind the method that wins.  Well, let's say the first one that is a
condorcet winner.  :)

Regardless, my main issue is being able to present scores, which I feel is
important if we are going to sell it to the masses.  (I am well aware that
this isn't such an important issue to others)  I think that all condorcet
methods can produce reasonable scores, it's just a lot harder to keep the
scores stable on some methods than others.

It is my opinion that all condorcet methods are imperfect, but that
plurality is so much worse, and so much more used, that a better use of my
own time is debating condorcet vs. plurality rather than condorcet vs.
condorcet.

If I have to debate against the range people, so be it, but that is mostly
because because I think the effect of the range advocates is actually to
keep the status quo.

The ones that I don't like have the quality that sometimes when the
> quantity of voters who rank candidate A, and don't rank candidate B at
> all,
> is larger than the quantity of voters who rank B at all, B can still win.


I'm a little confused by that.  To me, they always rank B....even if
implicitly.  If there are 4 candidates, a vote for
 A
is identical to a vote for
 A>B=C=D

Here is a simple example:
> 7 B>C
> 5 C
> 8 A
>
> What do you think? Is there good evidence and logic available for a method
> to decide that B is the best candidate to win?
>

Hmmm.  I'm not claiming there is 100% solid logic one way or another, it is
certainly a cycle.  For what it's worth, my scoring system places B out
front, but it is a close race.

B 35.08
C 33.39
A 31.53
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20071230/539049ee/attachment-0003.htm>