Is Winning Votes intuiltive?

[MIKE OSSIPOFF]

Blake said (at the end of his message):

BTW, I expect that Mike will have something to say about this issue, as
is his right. I've argued this point with Mike before, however, so I
don't intend to reply to him unless someone else indicates that they
would like to see a rebuttal for a particular argument.

I reply:

Well yes, it is my right, just as it's Blake's right to repeat
his old rebutted arguments.

Blake starts out:

So, is the point of your example that the Bush voter's are dishonest
then? Because if you just judge by the votes you give, without any
preconceptions about what Bush voters would really prefer, I think it is
obvious that Bush really does deserve to win in your example.

I reply:

Using strategy isn't necessarily dishonest. I don't believe that most
of us would put it that way. In fact many would say that truncation
isn't even insincere, just incomplete.

The message that Blake is replying to here had said that the Bush
voters had strategic reason to not rank Gore, something that Blake
apparently missed. It's true that the Bush voters in that example
had strategic incentive to truncate even if they prefer Gore to Nader,
unless they believe Nader is able to get a majority against Gore and
win. And, knowing that, with Buchannan in the election, to the right
of Bush (if he actually is), Gore & Nader voters would have a
defensive strategic need to not rank Bush, even if they felt that
there was some tiny chance that Bush might need their help to prevent
a Buchannan victory. That's why I say that Ranked-Pairs(margins) creates
a need for defensive strategy, as I define the term.

Blake continued:

However, if you count ballots in a way that tends to penalize those who
vote partial rankings, which SSD does

I reply:

With 0-info, SSD can give strategic incentive to randomly order
low-preference candidates among whom a voter has no preference.
It can pay off in some natural circular ties. So if, under those
conditions, you don't insincerely order those equally-disliked
candidates, someone could say that the nonuse of that strategy
was penalized. But, with SSD or with RP(m), you could
be sorry that you ranked them. And with RP(m), that could be the regret of
a majority who allowed the election of someone they dislike, whose
election they could have prevented by not ranking him. This
strategic incentive is of a whole different order than the milder strategic
incentive to randomly rank or to insincerely equal rank top candidates
in SSD, CSSD, RP(wv), or PC.

Blake seems unaware of RP(m)'s strategic needs, and their difference
from the milder strategic incentives of the wv methods.

Blake continued:

So, if people are better off believing that ranked ballots are not
allowed, why not just ban partial rankings?

I reply:

Banning truncation would be one way to avoid RP(m)'s vulnerability
to truncation. But then wouldbe truncators would just order-reverse
instead, causing a defensive strategic need for someone else to
defensively order-reverse by ranking a compromise alone in 1st place.

An advantage of the wv methods such as SSD, BeatpathWinner, CSSD,
RP(wv), & PC is that truncation isn't a problem for them. It doesn't
cause majority rule violation if some voters truncate. It doesn't
steal the win from a sincere CW. These things are true of the wv
methods, but aren't true of the margins methods.

Truncation is a problem for the margins method, and that's a problem
because truncation has taken place in all the rank ballotings that
I've conducted or participated in. Sometimes with stated strategic
intent. ("I don't have to vote between those two", when doing so
would obviously help Middle become voted CW, and not doing so would
therefore help that person's chance of having his favorite win the
strategic circular tie, taking victory from the sincere CW).

But in public elections with many candidates, many won't
rank all the candidates. Truncation will always be present in
public rank elections, and in RP(m) elections the truncation will
often cause majority rule violations and the defeat of sincere CWs,
creating a need for defensive strategy that wouldn't be there if counting 
were by a wv method.

Blake once objected "How important is
it to elect that CW that some people didn't bother to rank?"

Blake missed the point: The fact that that candidate is sincere CW
is important because it means that there's no sincere majority against
him. That fact is part of the demonstration that the wv methods meet
SFC. Instead of calling him sincere CW, we could say "a candidate against 
whom there's no pair-majority if no one order-reversed". But
sincere CW is brifer to say. The fact that that, in RP(m), truncation
can steal the election from such a candidate, in violation of majority
rule, is worth mentioning.

RP(m)'s vulnerability to truncation is brought out by its failure
of SFC & GSFC, 2 criteria that are met by SSD, CSSD, BeatpathWinner,
and RP(wv).

Mike Ossipoff




_________________________________________________________________
MSN Photos is the easiest way to share and print your photos: 
http://photos.msn.com/support/worldwide.aspx