[EM] To Martin re: SFC

[MIKE OSSIPOFF]

Martin wrote:

MIKE OSSIPOFF wrote:

>
>With wv methods, a group consisting of a majority of all the voters
>can ensure that a lesser-evil (B) won't win, without applying any
>strategy whatsoever, without doing other than sincerely ranking
>all the candidates, provided that they prefer the sincere CW to B,
>and that falsification doesn't occur on a scale sufficient to change
>the election outcome.

"and that falsification doesn't occur on a scale sufficient to change
the election outcome"

I'm not convinced this is a useful change... in fact, it seems like a
tautology, unless I misunderstand it. It seems that it would be true if
even a single falsified vote causes SFC to fail, and it would be true if
no matter how many falsified votes there are, SFC won't fail.

Example of former: the method which discovers the sincere CW from an
oracle, and elects the sincere CW provided that nobody falsifies, and
otherwise elects the devil as a punishment.
Example of latter: the method which ignores the votes as voted, and just
elects the sincere CW.

I reply:

I must admit that I don't know what you're talking about.

I agree that a method that would elect the devil would be undesirable,
but perhaps you'll note that none of my suggested count rules contain
such a provision.

And the methods you propose aren't at all patterned on SFC,
since the goal of electing a CW isn't mentioned in SFC.

You've missed the point. SFC's guarantee applies, for complying
methods, if a few plausible conditions are met. Congratulations
for showing that you can invent an unproposable method in which
one voter falsifying can change the outcome. With proposable methods,
it usually takes lots of voters to change the outcome. Of course,
with any method, it can happen with just one or two voters too, and
so you needn't have come up with methods that _reliably_ let their
outcome be changed by one voter.

Will it be common for falsification to take place on a scale
that will change the outcome, in SFC-complying methods? No.

So that condition, like SFC's other conditions, is quite plausible.
So SFC-complyng methods will typically offer the guarantee that
SFC describes.

As I said, the wording actually stipulates that no one falsifies,
but it's put that way just for simplicity. Obviously typically one
or a few people could falsify, and it wouldn't make the slightest
difference in the outcome. Enough falsification to make a difference,
enough to cause an SFC-complying method to fail to offer the
guarantee that SFC offers (B won't win) will be rare, with the
kind of methods that meet SFC.

I'm just talking about the freedom to not bother with any strategy,
and the plausible conditions that allow that freedom, with SFC-complying
methods.

Mike Ossipoff




_________________________________________________________________
Get your FREE download of MSN Explorer at http://explorer.msn.com