[EM] SARA voting: easier-to-describe MAS
Michael Ossipoff
email9648742 at gmail.com
Wed Oct 26 11:29:11 PDT 2016
On Wed, Oct 26, 2016 at 1:07 PM, Jameson Quinn <jameson.quinn at gmail.com>
wrote:
> My overall point is that, while VSE does not directly and precisely
> measure the true "goodness" of the election outcome, if done carefully and
> if assumptions are varied, it gives us the best and least-biased measure we
> have of that goodness.
>
It isn't a measure of goodness at all, as my homeless-man/billionaire
example shows.
.
Yes, D2, D3, Dexp & Dhyp are arbitrary.
Well, maybe D2 isn't so arbitrary, when it says that the importance of
changing a disutility is proportional to the magnitude of the disutility.
But D2, D3, Dexp & Dhyp all resemble what we know to be right. ...what we
intuitively and subjectively know to be right. And the basis of
ethics/morality _is_ intuitive and subjective.
> If you want to criticize it, propose an alternative framework for
> evaluating voting systems.
>
That's what I've been doing.
For one thing, I used D1, D2, D3, Dexp & Dhyp to evaluate the
election-result in Forest's example. They all choose candidate D in the
example., suggesting that D is the right choice by the best ethical/moral
standards.
Of course I claim that D1 isn't the good measure that the others are. If
the others agreed with eachother but not with D1, then I'd say that their
conclusion is the one that is right.
I've also been saying that voting systems should be evaluated according to
how little a voter needs insincere strategy to make hir best effort to
achieve what s/he wants to achieve.
I suggest that, if you have a top-set, then electing from that top-set is,
by the definition of a top-set, the most important voting-goal for you.
Though it does't match the "sincerity" definition that I use with criteria,
I suggest that top-voting one's top-set is sincere in a meaningful sense.
A lot of people want to vote _among_ their top-set too. I've been arguing
that, in general, doing so lowers their Pt (probability of electing from
their top-set), and is therefore suboptimal.
But I recognize that people might want rankings, or need them for reasons
that I've described in previous posts. In a method that strictly or
effectively meets MMC, a majority-favored (MF) voter can safely rank
sincerely, choosing among hir top-set, if the other members of hir
mutual-majority (MM) do so too.
But, for one thing, you can't always be sure if you're MF, and guessing
that you are must result in a suboptimal ballot.
For another thing, not everyone will be MF, and so that takes some of the
importance away from MMC.
These considerations take away the advantages of rank methods over Approval.
...except that there remains the fact that MMPO meets Weak CD, in addition
to FBC, and its Strong CD failure is mitigated by its deterrence of burial.
If you aren't MF, then ranking is suboptimal, so, in Bucklin or MMPO it
would be better to approval-vote, to equal-top-rank your top-set.
That's why I claim that FBC, so that top-voting one candidate won't undo
the benefit of top-voting another candidate, is necessary, based on the
above standards for evaluating voting-systems.
But, if some people will want to rank anyway, then it becomes desirable to
minimize the insincere strategy that they need, to elect from their
top-set, or just to elect the best individual candidate they can (which is
why they want to rank).
The CWs is the best individual candidate that they one can get. Also, the
CWs is maybe most likely (but certainly not necessarily) the candidate who
is in the most people's top-set. So, if you aren't MF, then at least it's
in your interest to elect the CWs.
So I've been evaluating methods according to how much insincere strategy is
needed to protect the CWs's win.
So yes, I've been proposing an alternative framework for evaluating
voting-systems.
Michael Ossipoff
(I don't have a way to delete the text below)
On Wed, Oct 26, 2016 at 1:07 PM, Jameson Quinn <jameson.quinn at gmail.com>
wrote:
> My overall point is that, while VSE does not directly and precisely
> measure the true "goodness" of the election outcome, if done carefully and
> if assumptions are varied, it gives us the best and least-biased measure we
> have of that goodness.
>
It isn't a measure of goodness at all, as my homeless-man/billionaire
example shows.
.
Yes, D2, D3, Dexp & Dhyp are arbitrary.
Well, maybe D2 isn't so arbitrary, when it says that the importance of
changing a disutility is proportional to the magnitude of the disutility.
But D2, D3, Dexp & Dhyp all resemble what we know to be right. ...what we
intuitively and subjectively know to be right. And the basis of
ethics/morality _is_ intuitive and subjective.
> If you want to criticize it, propose an alternative framework for
> evaluating voting systems.
>
That's what I've been doing.
>
> 2016-10-26 13:01 GMT-04:00 Michael Ossipoff <email9648742 at gmail.com>:
>
>>
>>
>> On Wed, Oct 26, 2016 at 8:31 AM, Jameson Quinn <jameson.quinn at gmail.com>
>> wrote:
>>
>>> Michael, you're arguing that certain kinds of utility or disutility are
>>> more important than others.
>>>
>>
>> A change in a large disutility is more important than the same change in
>> a small disutility.
>>
>>
>>
>>> As a human being with my own judgment and morals, I'd agree with you.
>>> But it's not just impossible, but actively counterproductive, to try to
>>> build that kind of judgment and morals into a voting system. If a voting
>>> system weights certain kinds of ballots more, sophisticated voters will
>>> strategically cast that kind of ballots, and unsophisticated voters will be
>>> ignored.
>>>
>>
>> Of course. That's why I'd only suggest D2, D3, Dexp or Dhyp for a
>> completely honest and altruistic ideal utopian electorate. .
>>
>> I like Score-Voting, and I completely agree with CES's advocacy of it,
>> along with Approval. But minimizing BR definitely isn't what justifies
>> Score.
>>
>> As soon as you speak of the social rightness of BR minimization, then
>> you're already entering the realm of ideal morality & ethics. In that
>> realm, BR definititely isn't it, because, obviously, changes to a large
>> disutility are more important than changes to a small disutility.
>>
>> Approval has a social optimization when people approve only their
>> top-set. Not utopian, but attainable with any electorate.
>>
>> In Utopia, for specific action decisions, or maybe for some
>> policy-choices, we could do better with D2, D3, Dexp or Dhyp. Approval
>> would be fine for electing a candidate or party.
>>
>> But, because Utopia isn't here (and legitimate elections aren't here
>> either), then of course it isn't an essential thing to reach agreement on
>> now.
>>
>>
>>
>>> , but, by an argument similar to the Condorcet Jury Theorem, in the long
>>> run it's got the best chance of agreeing with a system with did.
>>>
>>
>> Could you reword that?
>>
>>
>>
>>> To take your specific example: there are a lot more homeless people than
>>> billionaires, so in general a democratic election system will (correctly)
>>> weight the preferences of homeless people above those of billionaires.
>>>
>>
>> Yes, so democracy, if we had it, would give a good result, even without
>> utopian altruism.
>>
>>
>>> (And if the billionaires can successfully trick all the homeless people
>>> into thinking they prefer a candidate who will actually serve the
>>> billionaires, there's nothing the voting system per se can do about that.)
>>>
>>
>> There's a lot that honest, open media can do about that. Billionaires'
>> media can & do convince people that Hillary, her corruption, her bad
>> policies, is the best they can get, but, that position wouldn't hold up at
>> all in a society in which the media are honest, open, participatory, &
>> agenda-free.
>>
>> Anyway, when the elections are illegitimate due to unverifiable
>> vote-counting, there's no particular reason to believe that the voters and
>> how they vote have any role in public policy..
>>
>> Michael Ossipoff
>>
>> (I don't have a way to delete the text below)
>>
>>
>>
>> 's a negative change, because changes in greater disutilities are more
>> important
>>
>> Michael Ossipoff.
>>>>
>>>> On Tue, Oct 25, 2016 at 4:07 PM, Jameson Quinn <jameson.quinn at gmail.com
>>>> > wrote:
>>>>
>>>>> I've tweaked the wording for SARA again. The only substantive changes
>>>>> in outcome from this new wording is the change from "50 points or more" to
>>>>> "more than 50 points".
>>>>>
>>>>> Here's the latest wording, in 3 steps:
>>>>>
>>>>> Support Accept Reject Abstain (SARA) works as follows:
>>>>>
>>>>> 1. *Voters can support, accept, reject, or abstain on each
>>>>> candidate. Default is abstain. Candidates get 2 points for each percent of
>>>>> "support" and 1 point for each percent of "accept", for a total of 0-200
>>>>> points.*
>>>>> - *"Support" the best candidates (perhaps a quarter of them),
>>>>> "reject" the worst (perhaps half of them). "Accept" and "abstain" are for
>>>>> the ones in the high middle range. For those, "accept" if you want to help
>>>>> them beat somebody worse, and "abstain" if you could live with them but are
>>>>> hoping for somebody better.*
>>>>> 2. *Eliminate any candidates rejected by over 50%, unless that
>>>>> leaves no candidates with over 50 points.*
>>>>> - *If possible, the winner shouldn't be somebody opposed by a
>>>>> majority. But this shouldn't end up defaulting to a candidate who couldn't
>>>>> at least get accepted by over 1/2 or supported by over 1/4 (as in, a
>>>>> majority subfaction of a divided majority, such as Nashville voters in the
>>>>> example below).*
>>>>> 3. *Highest points wins. In case of a tie, fewest rejections wins.*
>>>>> - *This finds the candidate with the widest and deepest
>>>>> support.*
>>>>>
>>>>>
>>>>> 2016-10-22 18:24 GMT-04:00 Jameson Quinn <jameson.quinn at gmail.com>:
>>>>>
>>>>>> Support Accept Reject Abstain voting works as follows:
>>>>>>
>>>>>>
>>>>>> - Voters can support, accept, reject, or abstain on each
>>>>>> candidate. Default is abstain.
>>>>>> - Call a candidate "acceptable" if they are rejected by 50% or
>>>>>> less and supported or accepted by over 25%. If any candidates are
>>>>>> acceptable, eliminate all who aren't.
>>>>>> - Give remaining candidates 2 points for each "support", 1 point
>>>>>> for each "accept", and half a point for each "abstain". Highest points wins.
>>>>>>
>>>>>>
>>>>>> This moves a bit away from the Bucklin roots of MAS, but it further
>>>>>> reduces the instability of cooperation in a CD scenario.
>>>>>>
>>>>>
>>>>>
>>>>> ----
>>>>> Election-Methods mailing list - see http://electorama.com/em for list
>>>>> info
>>>>>
>>>>>
>>>>
>>>
>>
>
take away some of the advantage of rank methods over Approval.
>
> 2016-10-26 13:01 GMT-04:00 Michael Ossipoff <email9648742 at gmail.com>:
>
>>
>>
>> On Wed, Oct 26, 2016 at 8:31 AM, Jameson Quinn <jameson.quinn at gmail.com>
>> wrote:
>>
>>> Michael, you're arguing that certain kinds of utility or disutility are
>>> more important than others.
>>>
>>
>> A change in a large disutility is more important than the same change in
>> a small disutility.
>>
>>
>>
>>> As a human being with my own judgment and morals, I'd agree with you.
>>> But it's not just impossible, but actively counterproductive, to try to
>>> build that kind of judgment and morals into a voting system. If a voting
>>> system weights certain kinds of ballots more, sophisticated voters will
>>> strategically cast that kind of ballots, and unsophisticated voters will be
>>> ignored.
>>>
>>
>> Of course. That's why I'd only suggest D2, D3, Dexp or Dhyp for a
>> completely honest and altruistic ideal utopian electorate. .
>>
>> I like Score-Voting, and I completely agree with CES's advocacy of it,
>> along with Approval. But minimizing BR definitely isn't what justifies
>> Score.
>>
>> As soon as you speak of the social rightness of BR minimization, then
>> you're already entering the realm of ideal morality & ethics. In that
>> realm, BR definititely isn't it, because, obviously, changes to a large
>> disutility are more important than changes to a small disutility.
>>
>> Approval has a social optimization when people approve only their
>> top-set. Not utopian, but attainable with any electorate.
>>
>> In Utopia, for specific action decisions, or maybe for some
>> policy-choices, we could do better with D2, D3, Dexp or Dhyp. Approval
>> would be fine for electing a candidate or party.
>>
>> But, because Utopia isn't here (and legitimate elections aren't here
>> either), then of course it isn't an essential thing to reach agreement on
>> now.
>>
>>
>>
>>> , but, by an argument similar to the Condorcet Jury Theorem, in the long
>>> run it's got the best chance of agreeing with a system with did.
>>>
>>
>> Could you reword that?
>>
>>
>>
>>> To take your specific example: there are a lot more homeless people than
>>> billionaires, so in general a democratic election system will (correctly)
>>> weight the preferences of homeless people above those of billionaires.
>>>
>>
>> Yes, so democracy, if we had it, would give a good result, even without
>> utopian altruism.
>>
>>
>>> (And if the billionaires can successfully trick all the homeless people
>>> into thinking they prefer a candidate who will actually serve the
>>> billionaires, there's nothing the voting system per se can do about that.)
>>>
>>
>> There's a lot that honest, open media can do about that. Billionaires'
>> media can & do convince people that Hillary, her corruption, her bad
>> policies, is the best they can get, but, that position wouldn't hold up at
>> all in a society in which the media are honest, open, participatory, &
>> agenda-free.
>>
>> Anyway, when the elections are illegitimate due to unverifiable
>> vote-counting, there's no particular reason to believe that the voters and
>> how they vote have any role in public policy..
>>
>> Michael Ossipoff
>>
>> (I don't have a way to delete the text below)
>>
>>
>>
>> 's a negative change, because changes in greater disutilities are more
>> important
>>
>> Michael Ossipoff.
>>>>
>>>> On Tue, Oct 25, 2016 at 4:07 PM, Jameson Quinn <jameson.quinn at gmail.com
>>>> > wrote:
>>>>
>>>>> I've tweaked the wording for SARA again. The only substantive changes
>>>>> in outcome from this new wording is the change from "50 points or more" to
>>>>> "more than 50 points".
>>>>>
>>>>> Here's the latest wording, in 3 steps:
>>>>>
>>>>> Support Accept Reject Abstain (SARA) works as follows:
>>>>>
>>>>> 1. *Voters can support, accept, reject, or abstain on each
>>>>> candidate. Default is abstain. Candidates get 2 points for each percent of
>>>>> "support" and 1 point for each percent of "accept", for a total of 0-200
>>>>> points.*
>>>>> - *"Support" the best candidates (perhaps a quarter of them),
>>>>> "reject" the worst (perhaps half of them). "Accept" and "abstain" are for
>>>>> the ones in the high middle range. For those, "accept" if you want to help
>>>>> them beat somebody worse, and "abstain" if you could live with them but are
>>>>> hoping for somebody better.*
>>>>> 2. *Eliminate any candidates rejected by over 50%, unless that
>>>>> leaves no candidates with over 50 points.*
>>>>> - *If possible, the winner shouldn't be somebody opposed by a
>>>>> majority. But this shouldn't end up defaulting to a candidate who couldn't
>>>>> at least get accepted by over 1/2 or supported by over 1/4 (as in, a
>>>>> majority subfaction of a divided majority, such as Nashville voters in the
>>>>> example below).*
>>>>> 3. *Highest points wins. In case of a tie, fewest rejections wins.*
>>>>> - *This finds the candidate with the widest and deepest
>>>>> support.*
>>>>>
>>>>>
>>>>> 2016-10-22 18:24 GMT-04:00 Jameson Quinn <jameson.quinn at gmail.com>:
>>>>>
>>>>>> Support Accept Reject Abstain voting works as follows:
>>>>>>
>>>>>>
>>>>>> - Voters can support, accept, reject, or abstain on each
>>>>>> candidate. Default is abstain.
>>>>>> - Call a candidate "acceptable" if they are rejected by 50% or
>>>>>> less and supported or accepted by over 25%. If any candidates are
>>>>>> acceptable, eliminate all who aren't.
>>>>>> - Give remaining candidates 2 points for each "support", 1 point
>>>>>> for each "accept", and half a point for each "abstain". Highest points wins.
>>>>>>
>>>>>>
>>>>>> This moves a bit away from the Bucklin roots of MAS, but it further
>>>>>> reduces the instability of cooperation in a CD scenario.
>>>>>>
>>>>>
>>>>>
>>>>> ----
>>>>> Election-Methods mailing list - see http://electorama.com/em for list
>>>>> info
>>>>>
>>>>>
>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20161026/ed8b3d42/attachment-0001.htm>
More information about the Election-Methods
mailing list