[EM] Voting-System Choice for Polls (Just one more thing I want to say)

[Michael Ossipoff]

On Sun, Dec 18, 2016 at 12:46 PM, Toby Pereira <tdp201b at yahoo.co.uk> wrote:

> It depends on what you mean by polls. If it's just an opinion poll to see
> what the likely result will be in an election (so something that isn't in
> any way binding itself), it's about more than finding a method that will
> produce a winner.
>

Yes. This isn't a political poll at all. It's an online poll on a
non-political question, to find out what the CWs is, and which alternatives
have majority approval.



> You want to be able to see the support of all the candidates, and by
> having a method that's just set up for a winner (your suggestion of the
> pairwise winner from two different counting methods), you're not going to
> achieve what you want to achieve. For this sort of poll, you can use a
> variety of different methods and publish all the results
>

Yes, when the result won't be implemented, then there's no reason why it's
necessary to name the count-method in advance, because the results by
various methods could be given. If there's participation in that poll, I'll
report results by MAM, Approval, &  probably Pairwise-Winner (MAM,
Smith,MMPO).

 Most likely there won't be strategy, and the CWs will win as CWv.

But sometimes, when there are only a few voters, other methods can be
needed for tiebreaking.

I recently did a poll on voting-systems, and MAM returned a tie between
Approval & Score.

The alternatives were:

Approval
Score
Bucklin
MDDA
MDDAsc
IRV
Benham

But I noticed that Approval did better than Score in its pairwise
comparisons.

1. Approval pairbeat more alternatives than Score did.

2. The sum of Approval's pairwise votes against the other alternatives was
greater than that of Score.

#1 means that Approval beats Score by Copeland.

#2 means that Approval beats Score by a version of Borda.

But #1 & #2 seem to compellingly indicate that Approval beats Score.

Of course there are good reasons to not announce Copeland or Borda as the
count-rule. But, when there's a tie, they point to a winner in a meaningful
way.

Somewhere else, at PoliticalForums, I'm conducting a presidential poll
between Hillary, Jill, Donald, & Gary.

I don't know the winner so far, because someone additional has just voted,
and i haven't yet updated the count.

Though PoliticalForums' Opinion-Polls forum supports Approval, with
balloting, count, & count-display fully automated (voting is by
click-bubble), I instead invited rankings.

Because I wanted people to vote, and because IRV is so popular and
relatively well-known, I designated IRV as the count-rule.

But I'll announce the winner by Approval,  Benham & by
Pairwise-Winner (MAM, Smith,MMPO) too.

(I told people to rank only the candidates they approve of.)

I conducted a voting-systems poll there too, and Approval is the winner
there as well.

I'm conducting two voting-system polls: One at PoliticalForums,and one at
CIVS (Condorcet Internet Voting Service).

I emphasize that those two polls, at PoliticalForums & at CIVS, are to find
out how _the general public_ feels about voting-systems. Obviously the
people here at EM, who are not representative of the general public on that
subject, shouldn't vote in those general-public-opinion polls.

But of course feel free to check out the results.

As I said, Approval is the consistent winner, at both polls. Score is
2nd-best at CIVS. It seems to me that, at PoliticalForums, Score is tied
with Plurality. No one but me approved MDDA.

.
>
> But if you're just talking about elections that aren't for public office,
> then things are different. Some of these elections can be done online.
>

Yes, I've got several online polls going, on presidential candidates,
voting-systems, & nonpolitical reform questions.

In the CIVS presidential poll, with 72 votes in, Jill Stein is the CWv
(with Bernie removed from the count).

I invite people at EM to vote in the PoliticalForums presidential poll, if
they want to. Go to PoliticalForums (You can find a link to it via google),
and then go to its Opinion Polls forum. Among the polls there is my
presidential poll.

As I said, of course the voting-system polls are only for the general
public, not for people who are familiar with voting-systems.



> And for those that are likely to have an involved electorate that are
> likely to be knowledgeable about the system, I would suggest score voting
> but with live totals published and changeable votes. So people can enter
> their scores, but if the current result suggested they will need to adopt a
> more strategic approach, they can change their vote accordingly. The only
> other thing I would add is that the end time should probably be in some way
> non-deterministic. Otherwise the live updates are likely to be less
> effective. People might withhold their vote until the last minute, or have
> a completely false vote that they change at the last minute. So you might
> have 24 hours guaranteed (or however long is deemed appropriate), and then
> it might randomly end with a half life of an hour or something (which could
> be longer if the initial guaranteed time is longer).
>

Yes, that would be a good reliable way to find the CWs. It could be
feasible in a meeting-room, but, for most online polls, it isn't feasible.
For one thing, it's difficult enough to get people to vote once.

Rankings is usually a good way to find out the CWs, and MAM always works
fine at CIVS. But, at a forum where the people are very familiar with the
matter being voted on, and highly committed to some alternative(s), I feel
that it might be better to add Smith,MMPO's defection-proofness, via

Pairwise-Winner (MAM,  Smith,MMPO)

or

Pairwise-Winner (MAM,  Smith//MMPO)

Any opinions on which would be better?


Mike





------------------------------
*From:* Michael Ossipoff <email9648742 at gmail.com>
*To:* EM <election-methods at lists.electorama.com>
*Sent:* Sunday, 18 December 2016, 6:05
*Subject:* [EM] Voting-System Choice for Polls (Just one more thing I want
to say)

Before quitting EM & retiring from voting-systems, there's one more thing
I'd like to say:

What's the best voting-system for polls?

Though FBC is important for official political elections, I don't think it
serves a purpose in polls, where the purpose is to get sincere rankings,
and hope that people vote sincere rankings. In polls, the compulsion to
favorite-bury is much less likely.

As I've said, I haven't noticed any sign (top-cycles for 1st place) of
strategy in polls.

I suggest that, for sincere electorates, MAM is the ideal best. That means
it's best for polls at the Condorcet Internet Voting Service, where there's
been no sign of strategy.

But what if you're doing a poll among people who are highly involved in the
subject that you're polling about, and have strong committment to some
alternatives? Like, for example, suppose you're pollng at EM about
voting-systems?

Maybe some members of the electorate will resort to strategy. Especially if
the electorate are a voting-system mailing-lislt.

So you can't really be sure that there'll be no chicken-dilemma defection.

So maybe, instead of MAM, Smith//MMPO should be used.

It is automatically resistant to chicken-dilemma defection.

But, for burial, it isn't as good as MAM. With MAM, a candidate that you, &
sufficiently-many others, don't rank can't beat the CWs by burial.

...but it can in MMPO, though there's a lot of uncertainty & risk in trying
burial in MMPO.

Maybe Smith//MMPO's reliable automatic chicken-dilemma protection is more
important, because defection is easier & less drastic a strategy than
burial.

But maybe MAM's better burial protection is more important, because burial
temptation & opportunity is a lot more common than a chicken-dilemma
situation.

My suggestion: Use both.

Do the count by Smith//MMPO, & by MAM. Of the winners by those 2 methods,
the final winner is the one that pairwise-beats the other.

That's a solid good solution, because:

In the chicken-dilemma example, and also in a burial example, the intended
victim of the offensive strategy pairbeats the perps' candidate.

So, declaring, as winner, the one of those 2 winners that pairbeats the
other is definitely the best solution, if MAM & Smith//MMPO are the best
choices, each of which offers better protection in different ways.

Of course, it's been pointed out that methods that elect the pair-winner,
among the winners by 2 different methods, tend to fail FBC.

But FBC isn't needed in polls, where you want sincere ranking, not
equal-top-ranking.

Michael Ossipoff

----
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/20161218/271271b8/attachment-0001.htm>