[EM] Meta-criteria 6 of 9: Heuristics. #1, simplicity

[Kevin Venzke]

Hi Juho,

--- En date de : Dim 9.5.10, Juho <juho4880 at yahoo.co.uk> a écrit :
> >> It's just me and some others that have said that
> margins
> >> seems at least more natural than winning votes
> (not
> >> necessarily ideal).
> > 
> > Ok, good. I just want to be clear on who is saying
> this.
> 
> Are you saying that winning votes are more natural? (or
> better for some other reasons)

What I'm saying is that nobody but margins advocates concede that margins
is more natural. As far as I know.

But yes, when you have an understanding of sincerity that truncation
means "no comment" rather than "exactly equal" then WV (and other methods
that consider it damning when a candidate receives too few explicit
votes) then I regard WV as more natural. I don't consider any scenarios
where another treatment seems realistic.

> >>> What *does* mean that WV is considered to be
> better
> >> with sincere votes
> >>> is the perspective that when voters truncate
> they mean
> >> to abstain, and
> >>> not lend weight to the ignored contests.
> >> 
> >> What does "abstain" mean? Margins thinks that
> 40-30 and
> >> 30-20 opinions are of same strength. Should we
> consider the
> >> strengths of those opinions to be (40-30)/70 and
> (30-20)/50?
> >> Now we treat the truncated and equal rankings in
> some sense
> >> as abstentions. This approach has also been
> discussed but it
> >> is not winning votes. With this approach opinion
> 30-20 is
> >> stronger than 40-30 (unlike in winning votes). It
> differs
> >> from margins in that it counts how much stronger
> the winning
> >> side is in percentage (not in number of votes as
> in
> >> margins).
> > 
> > I don't remember this approach being discussed but
> it's not how I would
> > do it. You wouldn't start with the margin: margin is a
> hybrid between
> > the ratio and the absolute vote count.
> 
> Also margins counts the absolute vote count (but not those
> on the winning side but the difference between the A>B
> and B>A votes). Technically margins may be somewhere
> between two other approaches but historically it was
> probably not generated and seldom described this way.

Yes but who cares. You have to give me a reason why I should start with
margins or else it won't occur to me.

> > What you would do is say that
> > 40:30 is 40+30 = 70, and 30:20 = 50. That, perhaps, is
> what WV wants to
> > be.
> 
> Hmm, if the number of voters with (non-equal) opinion on
> the A vs. B comparison would add the preference strength
> then those votes that support the losing side would
> contribute negatively to the strength of the defeat of their
> favourite (in a non-monotonic way).

Yes, I've actually told you that before.

> > On the other hand, maybe WV just wants to respect
> majorities expressed
> > over majorities that can only be calculated to exist
> due to vote
> > splitting. The "abstain" could mean "don't count my
> vote when looking for
> > majorities."
> > 
> > "Abstain" could mean "detract from the importance of
> the contest."
> > This is important when the race is over when voters
> decide whether the
> > method respected the most important contests.
> 
> Winning votes have some characteristics in this direction
> but I don't have any short nice description of what it does.
> If there are many ways to interpret equal rankings and
> truncations maybe one could have methods where the voter can
> explicitly indicate in the ballot if then intention is to
> say "I don't know", "absolutely equal", "I don't care",
> "abstain", "not an important defeat" etc.

I'm not sure what the point is. There is no reason to vote "absolutely
equal" unless that's just how you sincerely feel. Strategically it's
like voting 5/10 in Range. Your vote isn't important enough for you to
nail the value precisely at 5/10. You figure out which way the ocean is
flowing and you push with it or push against it with all the might that
you have.

> >> I don't have any better explanations to winning
> votes as
> >> the ideal way to measure preference strengths
> (better than
> >> "number of votes on the winning side against some
> other
> >> candidate").
> > 
> > Suggest another rule that respects majorities.
> 
> Doesn't margins do so?

Margins sees majorities everywhere; I'm talking about actually voted
majorities.

> >>> My point is more that I don't know of any
> coherent
> >> explanation of why
> >>> margins is "ideal" with sincere votes, and WV
> is not.
> >> But that explanation
> >>> must be out there if the only problem WV
> advocates
> >> have with margins is
> >>> its strategic incentives. There must be some
> >> 10+-year-old posts from WV
> >>> advocates unhappy that they can't just use
> margins.
> >> 
> >> There are many mails and web articles. But maybe
> less on
> >> the performance with sincere votes than about
> strategies. At
> >> least according to my experience it is difficult
> to raise
> >> good quality discussion on that very central
> topic.
> > 
> > I would say that's because it's difficult to
> understand what it refers to.
> > Sometimes people write criteria to say how a method
> should behave,
> > without direct reference to strategy, such as
> Condorcet, Plurality,
> > Smith, or CDTT. I don't know what the margins standard
> is other than
> > that its resolution rule can be explained in few
> words. Even the notion
> > that an equal ranking should be treated as a split,
> self-opposing vote
> > seems to be based in strategic concerns.
> 
> I often divide the discussion about various criteria in two
> parts. Those criteria that deal with which candidate should
> be elected with sincere votes and those that address the
> problems with strategies.
> 
> To me the description of minmax(margins) (that uses
> directly the concept of margins) as "elect the candidate
> that needs least additional votes to win all others" is at
> the same time both an exact definition of the method and a
> description of who would be the ideal winner. According to
> this philosophy one should elect the candidate that would
> face least opposition after being elected (by people that
> would have preferred some candidate X to the winner and
> could join forces to oppose try to harm the elected winner).
> I think there are many kind of elections where the used
> criteria to determine the optimal winner may be different.
> This philosophy is thus not the only possible philosophy but
> it is certainly one that could be considered even ideal in
> some/many single-winner elections.

That's fine. Feel free to use margins where it makes sense.

> >> I wonder if someone has somewhere written a
> description on
> >> how winning votes is intended to measure sincere
> opinions.
> > 
> > You already did. You observed that WV advocates tend
> to see a different
> > meaning to truncation. This is a different concept of
> sincerity that WV
> > accomodates.
> 
> I think I mentioned all the components that you refer to
> but I didn't find yet a description and understanding that I
> would call natural. I'm also always a bit confused about the
> truncation related rules and criteria since the calculation
> rules treat truncations and other equal rankings exactly the
> same way (i.e. technically there is no difference between
> these two).

They are both important but the truncation issue is far more important.
And I say again that there is no point to deliberately voting in the
margins style. It's like flipping a coin in Approval to simulate half-
approval.

For example WV has much better FBC-efficiency than margins. I tested this
some time ago and would have to find it, but it's easy to see why: In
margins if you need to weaken a defeat against your compromise, you can
do it *faster* by just ranking him ahead of your real favorite.

So yes, sometimes you want to "abstain" even at the top of the ballot.

> > When I run simulations and I see a bullet vote "A" to
> me this does not
> > mean "A>B=C" due to the fact that in a public
> election that interpretation
> > would be presumptuous and unrealistic.
> 
> This is confusing to me in the winning votes argumentation
> since technically equal rankings and truncations are handled
> the same way.

My simulations don't usually allow above-bottom equal rankings. The
behavior isn't well-defined for all methods.

> > If the setting were different, where voters are
> supposed to be
> > knowledgeable about all the candidates and vote the
> whole ranking whether
> > they want to or not, and it could be expected that the
> voters would
> > never be torn on whether to use strategy or risk
> giving the election
> > away (I say "strategy": in the form of burial this
> could have a defensive
> > or offensive intent in margins, and that's why I think
> it's a big problem
> > to remove truncation as an option) then I might have a
> use for the
> > margins standard of sincerity.
> 
> As long as we are talking about pure ranked ballots I tend
> to see all equal rankings as equal rankings. They can be
> allowed in all forms (truncation and other). If we should
> read some additional meanings from the ballots that should
> be clearly defined.

Have I done this? You seem to get confused about the WV interpretation
of truncation at the same time that you claim that WV wants to read
approval. Well, go ahead and look at plenty of other methods that do
explicitly interpret this as approval: Does it become clear what they
are doing?

The point is not that WV uses approval, it's that it works well with
a voter expectation that truncation has a certain significance. Maybe
the Plurality criterion explains this well enough. Maybe the principle
that voters would understand actually is approval.

> >> P.S. My biggest fears with winning votes is that
> it might
> >> in some real elections (with sincere votes)
> produce a result
> >> that people do not find natural.
> >> 
> >> 10: A>B=C
> >> 20: A>B>C
> >> 16: A>C>B
> >> 01: B>A=C
> >> 01: B>A>C
> >> 26: B>C>A
> >> 03: C>A=B
> >> 03: C>A>B
> >> 20: C>B>A
> >> 
> >> This set of votes is cyclic. B and C are from the
> same wing
> >> (they support each others). But should C win (as
> in most
> >> methods with winning votes) although B has more
> first
> >> preferences than C and also A supporters like B
> more than
> >> C?
> > 
> > As long as A loses I don't care that much. If voters
> care so much about
> > a 2% difference in FPs then we are generally in
> trouble. (What principle
> > does margins adhere to that prevents this from being a
> problem in
> > general?) And not even half of the A voters supported
> B.
> 
> If A represents the left wing then the left wing voters
> said "we prefer B to C if right wing gets majority and one
> of their candidates wins". In the left wing there were more
> voters that didn't maybe care or found all the right wing
> candidates to be equally bad than in the right wing. That is
> natural. Within the right wing B seems to be clearly more
> popular than C, so the right wing agrees with the left wing
> that B is better than C. The only remaining tricky part is
> that the votes are cyclic. Can we derive some such logic
> from that cycle that C should win instead of B?

Why all this talk about B:C, why don't you talk about A:B?

If this scenario is important I can't imagine what else is too. I would
be more worried about what the public will think when they see that
margins can overturn a sole majority decision, to respect a win by a
candidate who doesn't even matter. Blame it on the voters if you want but
I don't think they will be impressed.

> > The public needs to ask themselves what A>B=C
> means. They should have a
> > look at those ballots. (Really I would say it's only
> fair for truncation
> > to be completely banned in margins, except that I
> think that people
> > truncating without meaning to say "all the worst
> candidates are exactly
> > equal" would actually be causing less damage than if
> we forced them to
> > come up with something.)
> 
> I'm ok with margins and equal rankings (also other than
> truncation). Truncation due to not knowing enough or due to
> considering all the remaining candidates to be more or less
> equal are quite good and sincere reasons to truncate.
> Truncation due to laziness may lead to distorted results
> (since all opinions are not measured). Truncation due to
> fear of supporting the competitors of one's own favourites
> by ranking them in one's ballot is usually a
> misunderstanding. 

Completely disagree with that. A voter can typically gain absolutely
nothing by (sincerely) ranking candidates who are too poor given their
expectation for the result. Any risk or perception of risk (and margins is
not a LNHarm method) isn't mitigated by anything.

> There are also some cases where truncation
> could be related to strategic behaviour but in large real
> life elections this is hopefully marginal and mostly not
> rational.

Highly unlikely if "strategic behavior" includes not voting for the
worse of two frontrunners. That is a defensive strategy, so if you're
not including that I'm not sure what you mean.

In Burlington I believe 39% of the Republican voters declined to vote
for one of the major candidates of the other wing, and among that wing,
something like 18% of the voters defected (didn't support the other
major candidate of that wing). And that was an IRV election!

Kevin Venzke