[EM] CD & truncation-resistance. Definitions people like.

[Michael Ossipoff]

(Replying farther down)

On Oct 13, 2016 9:03 PM, "C.Benham" <cbenham at adam.com.au> wrote:
>
>
> 43: A
> 03: A>B
> 44: B>C  (sincere is B or B>A)
> 10: C
>
> C>A  54-46,    A>B  46-44,   B>C 47-10.
>
> Here A is the sincere CW and supported by the largest of the three
factions of voters, but Winning Votes rewards the buriers by electing B.
>
> Benham and  LV(erw)SME   easily elect  A.    Smith//Approval (equivalent
here to Max Covered Approval) and Approval Sorted Margins elect C.
>
>
>
> On 10/14/2016 1:08 AM, Michael Ossipoff wrote:
>
>> 1. Typically that CWs is a middle CWs, as in your example.
>

(You wrote):

> C: (By "CWs" Mike means 'sincere CW').  No, in my example A  could be a
wing candidate.

(endquote)

You said that sincere for the B voters is B>A or B. If it's B>A, then B is
middle CWs.

But yes, not if B voters' sincere preference is just B.

But by the assumption of a 1- dimensional political spectrum, a CWs will
usually be middle.

The rarity of top cycles in CIVS polls suggests that the 1D assumption is
at least fairly good.

(Replying farther down)

  Say A is Centre-Left,  B is Centre-Right and  C is
> (to use a Kevin Venzke expression) a "Martian" candidate, perhaps Far
Right.  C should be irrelevant and would be without the B supporters'
> burying.
>
>
>> So the B voters in your example dislike C more than the A voters do.
>>
>> A C victory will be worse for the B voters than for the A voters.
>>
>> So the A voters have a more credible threat.
>
> C:  No.

...unless there's a 1D spectrum, as suggested by poll results.

>
>> The A voters could say to the B voters:
>>
>> "Ok, let's get this straight: You're going to bury our candidate, and
you think we should show our gratitude by supporting your candidate in 2nd
place  ".
>
>
> C: This assumes that the B faction gave notice of their intention to bury
A

Punitive strategy achieves nothing if not announced.

(Replying farther down)

(and didn't all just blindly copy their candidates' "how-to-vote card" that
they saw
> for the first time just before they voted.)
>
> Also it assumes that for all (or nearly all) the A voters the B faction's
voting strategy is the most important issue in the election.

(endquote)

Not really.

But, if burial is at all likely, CWs preferrers should plump as a matter of
course.

Yes, the CWs can't always be reliablly predicted. The advantage of wv &
MMPO over Bucklin is that, if burial is deterred, then failure to plump is
harmless.

(You wrote):

But in any case in the example
> only 6% of the A voters failed to truncate, and that was sufficient to
allow the B factions' burial to succeed under Winning Votes (and MMPO).

...and, with the methods you prefer, there's no way the CWs's voters can
protect hir (other than by an elaborate Probabilistic voting strategy).

(I'd said):

>
>> What happens when there's a CWs is more important than how a natural top
cycle is solved.
>
>
> C: In the 3-candidate (and so 3 factions) scenario, I think we can
meaningfully distinguish between a "sincere CW" that in order to be the
actual voted
> CW has to rely on rival factions not merely truncating, and one that only
needs rival factions to not actively bury (order-reverse or vote an
insincere preference).

(endquote)

Exactly. Truncation of a CWs won't take away hir win in WV & MMPO.  ...but
it will in the methods you prefer.

> I don't much care about the former (weaker) type, and in any case I think
they are hopeless cause.  Resolving a "natural top cycle" properly is more
important
> than trying to save them (or to "deter" big bad "offensive truncators")

Unless you care about what it takes to protect a CWs, as voters try to do.
Unless you care how drastic their defensive voting must be.

And big bad truncation needn't be strategic. It could be big bad
principled, hurried, careless, etc.

...but in your methods it easily takes away a CWs win. Your choice.

>
> On the other hand the largest-of-three factions definitely shouldn't be
vulnerable to burial because only 94% of them truncated!

In your methods, S/he's vulnerable to burial with nothing the CWs's voters
can do about it.  ...and vulnerable even to big bad truncation.

Michael Ossipoff
>
> Chris Benham
>
>
>>
>> Hi Chris--
>>
>> (Replying farther down)
>>
>> On Oct 11, 2016 10:55 PM, "C.Benham" <cbenham at adam.com.au> wrote:
>> >
>> > On 10/12/2016 12:12 PM, Michael Ossipoff wrote:
>> >
>> >> Regarding the CD example, Chris suggested scenarios in which one of
the other candidates, not A, is the CWs.
>> >
>> >
>> > 33: A>B
>> > 32: B (sincere may be B>A)
>> > 34: C
>> >
>> >
>> > Mike,
>> >
>> > I don't recall doing that exactly.   I like compliance with CD.   The
example I posted recently:
>> >
>> >> Given that burial vulnerability is unavoidable in Condorcet methods,
I think that is more democratic if  (in this respect) larger factions
>> >> have the advantage over smaller factions.
>> >>
>> >> 43: A
>> >> 03: A>B
>> >> 44: B>C  (sincere is B or B>A)
>> >> 10: C
>> >>
>> >> C>A  54-46,    A>B  46-44,   B>C 47-10.
>> >>
>> >> Here A is the sincere CW and supported by the largest of the three
factions of voters, but Winning Votes rewards the buriers by electing B.
>> >>
>> >> Benham and  LV(erw)SME   easily elect  A.    Smith//Approval elects C.
>> >
>> >
>> > The  "WV strategy" enthusiast might chastise the 3 A>B voters for
allowing the B supporters' burial to succeed by not truncating...
>>
>> (endquote)
>>
>> Yes. If you believe your favorite is CWs, you can protect hir win by
defensive plumping.  ...or at least not ranking someone whose voters might
bury.
>>
>> You wrote:
>>
>> but if they'd
>> > done that then WV would have elected their least-preferred candidate.
>>
>> (endquote)
>>
>> Yes, if the B voters bury, then for the A voters,  truncation will  only
worsen their result.
>>
>> But such non-risk-free punitive defensive strategies are common
throughout the animal kingdom, including in human affairs.
>>
>> 1. Typically that CWs is a middle CWs, as in your example.
>>
>> So the B voters in your example dislike C more than the A voters do.
>>
>> A C victory will be worse for the B voters than for the A voters.
>>
>> So the A voters have a more credible threat.
>>
>> 2. In such situations, the defender has a more credible threat.
>>
>> The A voters could say to the B voters:
>>
>> "Ok, let's get this straight: You're going to bury our candidate, and
you think we should show our gratitude by supporting your candidate in 2nd
place :^) ".
>>
>> But, sure, there's at least the hypothetical possibility of the
perpetual burial fiasco, if A is in the B voters' bottom set.
>>
>> But I've told why that fiasco's requirements are a bit mutually
contradictory.
>>
>> Even if A is  in the B voters' bottom-set, the B voters might not bother
to  bury (even if it's theoretically optimal), if the chance of success is
sufficiently small.
>>
>> I'm not saying that the majority are always right, but wv is popular.
>>
>> Benham & Woodall don't reliability have wv strategy.
>>
>> In Bucklin & Approval, the CWs's voters can protect hir by plumping.
Likewise methods with wv strategy.
>>
>> Voters will do what it takes to protect a perceived CWs, because that's
the best they can get.
>>
>> What happens when there's a CWs is more important than hkw a natural top
cycle is solved.
>>
>> Michael Ossipoff
>>
>> >
>> > The tops-of-the-ballots oriented Benham and  LV(erw)SME  handle the
scenario by (effectively) laughing off C (and electing A in breach of
>> > Minimal Defense), while the more bottoms-of-the-ballots oriented
Smith//Approval notes that C is the most "approved" (i.e. voted above
>> > bottom) candidate and says to the 44 B>C voters "since you voted for C
you can have C".
>> >
>> >
>> >> Chris recently mentioned an incompatibility between CD & Minimal
Defense (Eppley's votes-only version of SDSC or WDSC).
>> >
>> >
>> > C: Certainly we can't have all three of  Plurality, CD and Minimal
Defense.
>> >
>> >
>> >> When discussing Margins Sorted Approval, I objected that it penalizes
truncation instead of electing the CWs
>> >
>> >
>> > C: MSA isn't one of my very favourite Condorcet methods, but I prefer
it to Winning Votes.  In the example it elects C.
>> >
>> > Chris Benham
>> >
>> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20161014/da74125f/attachment.htm>