Schulze strictly meets strong 1st Choice Criterion

[Mike Ositoff]

> 
> On Mon, 9 Nov 1998 00:24:44    Mike Ositoff wrote:
> >
> >Schule's method strictly meets the strong 1st Choice Criterion,
> >because the way that VA methods can fail it involve subcycle
> >fratricide, and Schulze's method isn't subject to subcycle
> >fratricide.
> >
> >I must add, though, that, since that criterion, as I defined it,
> 
> It's becoming pretty clear that the 1st Choice Criterion will
> never be defined.  Have you considered instead defining a

You must have missed the message in which I defined it. Maybe,
then, I should repeat the definition that I posted:

Weak 1st Choice Criterion:

It should never be necessary to vote a less-liked alternative
over one's favorite in order to prevent offensive strategy
from succeeding.

(Offensive strategy I define as doing other than sincerely ranking
all of the alternatives sincerely. That concrete definition
is useful even though much truncation won't be strategically
intended. Offensive strategy succeeds if it gains the election
of someone the stratgy-user prefers to the CW, when the CW would
have won had the strategy not been used).

Strong 1st Choice Criterion:

It should never be necessary to vote a less-liked alternative
over one's favorite to prevent offensive strategy from succeeding,
or to vote a less-liked alternative equal to one's favorite in
order to gain the election of the CW.


***

That's the most recent form that I'd posted for those definitions
at the time of Blake's posting.

Might it be preferable to limit it to protection of the CW,
and make a separate criterion about preventing offensive
strategy? Or to simplify by speaking only about preventing
the success of offensive strategy? Maybe, but the criterion
isn't without a definition--the one I wrote above.
Maybe it could be said to be more useful & simple to 
talk of defensive strategic need & define that in terms
of a majority voting insincerely to gain election for something
they prefer to what would have otherwise won. There are changes
that could be made, but so far I'm defining it as stated
above.

***

Instead of saying "No one should have defensive strategic
need to...", it's probably clearer to say, as I did here,
"It should never be necessary to _______ in order to ______",
or, better yet, that _________ should never 
be attainable only by ________. Excuse the blanks, but I'd
like to avoid repeating.


These criteria could obviously be worded in various ways,
and maybe some of the other wordings would improve clarity,
but it seems to be going too far to say that the 1st Choice
Criterion doesn't have a definition.

Anyone who agrees with the basic idea of the criterion
can define it differently, or suggest that anyone using
it should define it differently. But it's a little suspect
when the proponent of a method that's a flagrant violator
of any reasonably-defined 1st Choice Criterion makes the
claim that the criterion can't or won't be defined.

> "retaliatory strategy" criterion?  I think that would be
> definable and do what you're trying to do with 1CC.

A separately-named criterion for thwarting offensive strategyk
might be desirable, but I like the idea of naming it all under
the 1st Choice Criterion, since we're talking about the need
to sell-out your 1st choice to different degrees.

One desirable change would seem to be that, instead of
just speaking about _preventing the success_ of offensive
strategy by voting a less-liked alternative over your favorite,
it seems better to speak of _deterring_ offensive strategy,
which would mean preventing its success and electing someone
that the strategists have voted lower than the CW.

Blake, that Criterion is newly proposed, so don't be in such
a hurry to declare that it's definition will never be finalized.
It does have a definition so far, written earlier in
this message.

Mike Ossipoff



> 
> ---
> Blake
> 
> 
> -----== Sent via Deja News, The Discussion Network ==-----
> http://www.dejanews.com/  Easy access to 50,000+ discussion forums
> 
>