STV truncation. SW vs PR as issues for EM

[Mike Ossipoff]

Steve Eppley writes:
> 
> Rob L wrote:
> [snip]
> >do you see my point?  The 10 votes could get a long way from home
> >unless we stay true to many voters' intention *not* to have their
> >vote transferred. 
> 
> Yes, I see your point but I'm not sure if it's more important than
> allowing votes to be wasted.
> 
> Question: Does STV permit a voter to rank two candidates equally?  If 

The Santa Cruz STV proposal does, and several others do.

> so, how do the votes transfer?  (The reason I ask is because unranked 

One's remaining vote value is divided equally between the 
continuing candidates occupying one's highest rank-position containing
continuing candidates.

> candidates are essentially ranked *equally* last.  Do they share a 
> fraction of the ballot if it's "transferred" to the unranked?  I 
> assume they don't.)

Nothing of your vote should be transferred to candidates whom you
haven't ranked, unless you've explicitly chosen an option to do
that, as I described in a previous letter.

There have been proposals to automatically divide the value of
an exhausted ballot among all the candidates, but that goes
contrary to what the voter expressed.

> 
> Maybe you're right.  Maybe there should always be an implied NOTB
> following the last candidate ranked.  But if severe truncation is a
> problem, and if there's a way to *infer* rankings from *similar*
> ballots (like saying that {A,B,C} is similar enough to {B,A,C,D} and
> {A,B,C,E} so that D and E should share its transfer), perhaps some
> of the otherwise wasted votes can be salvaged.

Yes, that can be arranged by giving voters an option to punch a box
on their ballot to indicate that, if their ballot is exhausted, then
its value shall be re-assigned to the transferable ballots of
the candidate with whom the exhausted ballot resides. Or perhaps
the option to have their exhausted ballot's value divided among
the transferable ballos sharing its 1st choice. The 1st of these
2 possibilities is far simpler & easier to implement, requiring only
a different formula for the candidate's transfer fraction, so as
to take into account the combined value of the exhausted ballots
of that candidate that are using that option.

> 
> I don't know which violates the 1-person 1-vote principle more.


Either is ok, but F = min (s/t, 1) doesn't violate 1 person 1 vote,
in my opinion.

I should emphasize that all PR is ok, and that I don't have a strong
opinion about 1 PR system vs another. But, since this issue is
being discussed, I felt I might as well state my position on it.
But if you enact, in your city an STV system that uses, instead,

F = s/v, where s = the candidate's surplus, & v = his vote total,
I won't object. You can't really go wrong with PR. All PR systems
approximate proportionality pretty well, & it doesn't make much
difference which is used.

I feel that right now the real issue for EM is single-winner methods,
therefore.

Let's establish which single-winner method meets the standards that
we consider most important. I personally believe that we should
devote our effort to single-winner methods now, because single-winner
methods differ radically from eachother when it comes to merit.
And, especially now, when, among those of us who are actually
discussing the subject, Condorcet seems to be advocated by more
people than MPV, or any other method proposed in these lists.
And this is a time when Copeland, for some time proposed, or
at least spoken well of, by some academics, is now being proposed
in an open forum (EM). Right now is the time to have this out.
The fact that this is a Presidential election year adds to the
importance of this time for single-winner discussion & debate.

Single-winner methods is a completely un-resolved consensus-less
subject in the overall electoral reform scene. We can change that
now. PR is already being competently dealt with by electoral reformers.
Single-winner reform is where we're needed now. Our discussion of 
standards, and the recommendations that result from it, can give
some order & some answers to a subject that so far, in the overall
electoral reform scene, has no answers & no order. I say that's
how we maximize the value of what this list can do.


Mike Ossipoff


 
> --Steve
> .-
> 


--