[EM] Range Voting vs Condorcet (Greg Nisbet)

[Greg Nisbet]

> Reasons why Range is better and always will be.

> I would like to end the truce.





That won't work I guess. Using the term "better" alone is a major flaw of
many discussions here. Obviously, it all depends on what goals a method is
expected to achieve.

Ok, using the term "better" is biased, you got me. I will proceed to defend
my position. I merely mean that it satisfies criteria I think are important
and will proceed to attempt to convince why they are important. I do admit
the flaws of Range Voting and would be happy to explore them, but I think
that including my judgment of which is better is not evil. I too love
property based discussions. The definition of the "better"-ness criterion is
as follows:

"Better"-ness Criterion:

Arbitrary pronounced by the writer to be more suitable than some other
method for elections.

:-)

I'm not trying to impose my will on anyone; I would just like to see the
matter discussed.





> I'll be generous to the Condorcet camp and assume they suggest

> something reasonable like RP, Schulze or River.





As you might guess, I appreciate this, of course :-)





> Property Related:

> favorite betrayal, participation and consistency.

> Implications:

> 1) It is always good to vote and it is always good to rate your

> favorite candidate 100. The only Condorcet method to satisfy favorite

> betrayal is an obscure variant of Minmax which I'll ignore because of

> its glaring flaws (clone dependence *cough*)

> 2) How does it make sense to be able to divide a region into two

> constituencies each electing A if B is the actual winner? Condorcet

> methods are not additive, this calls into question the actual meaning

> of being elected by a Condorcet method.





No, it calls into question the actual meaning of being elected in a region.
The misunderstanding arises only when you interpret the election of A in a
region as meaning that A is "best" in some sense. But Condorcet methods are
based on a different logic than measuring "goodness" of candidates. They
have more to do with stability, for example: Electing a candidate other than
the Condorcet Winner always faces immediate opposition by some majority who
prefers the Condorcet Winner. So, if you consider majorities significant
(which you seem to judging from your reasoning further down), you should
consequently not accept different winner when a Condorcet Winner is
available.

I see your point. Try thinking about it from a different perspective:

The regions are the "status quo", not the conglomerate super-region.

I cite this example: http://www.rangevoting.org/CondNonAdd.html

Picture it this way, the region has chosen their representative, but
suddenly the introduction of a region that chose exactly the same way has
altered the result.

This example should justify my original point. The opinions of these two
groups are not in conflict with each other per se, yet their aggregate
decision is. If you claim that a region is a fictional entity with no real
meaning, I say look at the problem from the regions' perspective and
consider the super-region a fictional entity. (This might make it seem odd
why they are participating in the exact same election, but it seems less so
if they are voting on say, the region's favorite color.)





> answers to potentital majority rule counterarguments:

> 1) Range voting isn't a majority method.

> answer: any majority can impose their will if they choose to exercise

> it.





I greatly appreciate your making this clear! Warren has often argued that
Range is *not* really majoritarian when I pointed this out. The main point
is of course the question of whether one assumes intelligent voters who vote
strategically or dumb voters who vote "honestly" against their own interest
(whatever "honestly" means with a ratings-based method - more on this
below). When we assume intelligent voters, Range is clearly a majoritarian
method.

I was attempting to make a distinction between an active and a passive
majority. Any active majority (one unwilling to make any compromises
whatsoever, voting every non-them candidate the lowest possible score) will
win. A passive majority (clear majority opinion, but makes compromises) will
not necessarily win. I argue this isn't a fault because if a majority is
passive then they can arguably be considered to support another candidate
the percentage that they voted for him. E.g. their partial vote could be
used to form a majority per se.



That leads me to the main problem with Range (as with any other majoritarian
method): It is simply not democratic. It cannot be because every
majoritarian method gives 100% of the power to less than 100% of the people
(the "demos" in greek). Often, about 60% of the people can consistently
impose their will on the other 40% without the latter being given any means
at all by the majoritarian method to influence the decision. Of course, this
is a problem of most popular election methods, but that does not mean the
problem cannot be solved. Democratic decisions are possible but not with
majoritarian methods.

Interesting point. I would argue that a compromise candidate is better than
a polarizing but barely passing candidate (like FPTP with primaries tends to
produce). I'd say this isn't a voting-issues question, but a civil rights
question. A nice constitution will help protect you from tyranny of the
majority.

"Advocates of majoritarianism argue that
majority<http://en.wikipedia.org/wiki/Majority>decision making is
intrinsically democratic and that any restriction on
majority decision making is intrinsically undemocratic. If democracy is
restricted by a constitution
<http://en.wikipedia.org/wiki/Constitution>which cannot be changed by
a simple
majority <http://en.wikipedia.org/wiki/Simple_majority> decision then
yesterday's majority is being given more weight than today's; if it is
restricted by some small group, such as
aristocrats<http://en.wikipedia.org/wiki/Aristocrat>,
judges <http://en.wikipedia.org/wiki/Judge>,
priests<http://en.wikipedia.org/wiki/Priest>,
soldiers <http://en.wikipedia.org/wiki/Soldier> or
philosophers<http://en.wikipedia.org/wiki/Philosopher>,
then society becomes an oligarchy <http://en.wikipedia.org/wiki/Oligarchy>.
The only restriction acceptable in a majoritarian system is that a current
majority has no right to prevent a different majority emerging in the future
(this could happen, for example, if a minority persuades enough of the
majority to change its position). In particular, a majority cannot exclude a
minority from future participation in the democratic process. It should be
noted, as it's often a subject of misunderstanding, that majoritarianism
does not prohibit a decision being made by representatives as long as this
decision is made via majority rule, as it can be altered at any time by any
different majority emerging in the future."

>From http://en.wikipedia.org/wiki/Majoritarianism

The above excerpt embodies the spirit of majoritarianism fairly well. It
basically says, if not the majority, then who decides? If you delegate the
responsibility to some group (even yourself) to judge what is best for
society, then you are imposing your will on people.

Arguments both for and against majoritarianism both tend to boil down to
rights. Do you have the right to non-interference from the majority? Does
the majority have the right to non-interference from you?

I'd argue that the opinion of a passive majority shouldn't be given the same
weight as that of an active majority. Nevermind the fact that the two are
indistinguishable under Condorcet methods (how do you give people MINIMUM
scores?), if a majority attempts to make compromises, can their opinion
really be considered to be absolute and inviolable?





> Voter Experience:

>

> Range Voting (based on the existence of Amazon product ratings,

> youtube video ratings, hotornot.com, the number of movies rated out

> of stars.) I cannot find a single instance of Condorcet methods

> besides elections in various open source communities. It doesn't

> qualify as mainstream.





That may be right but is irrelevant for the question of what is a good
method and what is not.

If you are attempting to get it used in any situation with non-experts, you
better believe it's relevant.



> Understandability:

>

> Range Voting (I dare anyone to challenge me on this)





Then let me challenge you right away: I don't understand at all what those
numbers a range-ballot asks me for are supposed to mean. They are not
explained but instead it is simply assumed naively that each voter will be
able to assign meaningful numbers to options. Some even suggest that voters
should apply their "gut feelings" to derive the numbers - as if voting were
about diffuse emotions and not hard facts. In real world situations it is
difficult enough to decide whether I *prefer* A to B or B to A or neither to
the other. Often enough it turns out that A is preferable in some aspects
and B is preferable in other aspects. Suggesting to weigh the aspects first
is of no help since it lifts the problem of coming up with meaningful
numbers only to a higher, more abstract level. Also, Range advocates tell us
that the numbers are not simply supposed to be monetary values but rather
such things as degrees of "utility" or perhaps "happiness". Simple question:
Can you enumerate y

 our happiness with, say, having saved a species from extinction? And even
if some people can, it is obviously not justified to simply assume that
every voter should be able to do such magic without at least providing some
serious scientific evidence for such a bold claim.

I now realize the mistake of labeling how easy it is to teach someone to
emulate the method 'Understandability' and for that I apologize. : - )

What I meant is that if you attempt to teach anyone how to turn Range Voting
ballots into a winner, it will take a lot less time than the same feat in
Schulze.

I do not want your well thought out arguments to go to waste though : -)

I'll respond.
The arguments you are referring to about "I *prefer* A to B or B to A or
neither to the other" was covered in the apathy section. In short, I agree
with this claim that you can only be 100% certain of the directionality of a
comparison A > B, A< B, A ~ B (~ is the apathy comparison, a notation I made
up but seems to work so far). I mentioned this. Now, I would argue that
utility and money are intuitive concepts. The comparison of $45 to $1 to
$.99 is really, really simple. I understand that this is a far-fetched
example (by which I mean far removed from the realm of voting), but I
suggest that this is how voters actually see the world. Asking them to
quantify how much they like a candidate is not an impossible feat. Now,
range voting only deals with a single dimension of comparison, not the
aggregate impact of O(n^2) individual comparisons. I'd argue this is easier
for most voters to deal with and closer to how they actually see the world.
Also, your response comes dangerously close to refuting the concept of
decision entirely, in which case all voting is useless. "Often enough it
turns out that A is preferable in some aspects and B is preferable in other
aspects." You're walking on glass if you attempt to use this argument to
refute Range Voting but not binary Condorcet-level decisions.


=========================================================================


Greg,



If one limits criteria to the few you selected, Range "looks" good. I will
set aside for the moment the fundamental issue of whether mixing different
individuals' different scoring standards can convey any real meaning ( the
difference in candidate "quality" between candidates you score 8 and 10 may
be ten times greater than the difference your neighbor assigns to candidates
scored 2 and 10), and just speak about other criteria.



I am not attempting to limit the debate to a few criteria. I welcome all
criteria anyone deems important. I only mentioned a few to begin with, and
they didn't all support Range Voting e.g. the apathetic voter's dilemma
argument. Go ahead and accuse my coverage of being biased; it probably is, I
am human after all.

If you wish to continue setting the issue of combining different scales
together aside, just say so, but I will respond to your argument
nonetheless.

The reason they can be treated equally is their impact on the mean. Each
voter's decision might be using different scales but there is certain
behavior that all rational voters' votes will follow (such as transitivity
and no wasted space). In addition to this, the votes' impact on the mean is
identical and it is fundamentally the IMPACT OF A VOTE THAT SHAPES ITS USE,
given enough time and a ration population. Still, a very good argument,
thanks. : - )



Personally, I agree that the Condorcet-winner criterion is not as
significant as it sounds (we may not be looking at a single solid majority,
but rather antagonistic divergent majorities). I think the mutual-majority
criterion is more significant (and Range violates it).  However, I think the
Condorcet-loser criterion is a show stopper. If a candidate would lose in
every single one-on-one match up, then that candidate should not win. But
such a Condorcet loser can indeed win under Range voting.

 I see! Here is my question though, how is the mutual majority criterion
relevant when the members of this "majority" specifically chose to
compromise with other candidates? My point is essentially this. Range voting
rejects the ranking mentality. I support A, failing A, I support B, failing
B, I support C etc. With Range it's a bit more abstract and less procedural.
Because nothing resembling an iteration (one could view comparing pairwise
contests sequentially as a type of iteration) ever occurs under Range
Voting, none of the candidates ever conceptually fail such that one's full
support is deemed to fall with a different candidate. I know this is a bit
complicated and I might be explaining things badly, but the bottom line is
this: Range Voting treats a vote of 100% for Bob and 60% for Alice as
exactly that. This voter is not determined to support Bob 100% over Alice
for the purposes of forming a majority. Thus a majority in the sense of the
mutual majority criterion doesn't really carry any weight. Thus it does obey
the mutual majority criterion if you define "belonging" to a particular
majority as 100% of this voter is going to the Bob camp and 60% to the Alice
camp.

If that answer is unsatisfactory, this is the simplified version "You are
ignoring the magnitude of preferences, which ARE knowable in Range Voting
hence the criticism doesn't apply."

I also believe the later-no-harm criterion is of crucial importance, which
Range fails.

IRV, FPTP, and some obscure variant of Minmax are the only methods to
satisfy later-no-harm.

FPTP is just pathetic.

IRV is non-monotone and leads to perpetual two-party domination.

I do not know enough about the obscure variant of Minmax to criticize it
specifically.

Range is more prone to strategic voting manipulation than either IRV or
Condorcet (see analysis by James Green-Armytage in his doctoral paper  linked
on this list a couple of months ago).

Once I find the paper I will have some better arguments, but in the
meantime, here is this.

Range Voting manipulation is straightforward, exaggeration. Manipulation in
Condorcet and IRV involve intricate support of other candidates to lead to
favorable match-ups. Strategic Range is just equivalent to Approval, which
fares better than Condorcet and IRV in Bayesian Regret simulations. In Range
Voting, the exaggerations are at least partway honest, reducing the chance
of a vastly inferior dark horse from winning.

Essentially, even if Range Voting is more prone to this, that doesn't prove
the result of strategy in fact makes Range Voting an inferior method to
Condorcet and IRV.  It also doesn't address the damage when strategic voting
is attempted. I argue that Range will at least lead to semi-plausible
winners under this system.

Range is also more prone to spoiler scenarios than IRV or
Condorcet-compliant methods, because the score a voter assigns is dependent
on what other candidates are, or are not in the race to compare with.

First, um, independence of irrelevant alternatives…

Next, IRV spoiler scenario: http://rangevoting.org/IRV1519.html

Third, Condorcet methods fail IIA

Now to defend the behavior that has been called the spoiler effect. First of
all, the rough magnitudes of voter comparisons are stable, thus an
additional candidate should not alter the placement of current candidates
unless they are the new best or new worst. You exaggerate the additional
impact of new relative comparisons of old ones.



Beyond the realm of standard criteria, I am also concerned about the effect
different voting methods have on candidate campaign behavior, and resulting
voter information. Some voting methods discourage candidates from revealing
their true positions on controversial issues, if avoiding voter alienation
is more crucial than earning first-preference support (this can be true of
both Range and Condorcet).

Let's look to Australia and Ireland for evidence of the impact of IRV,
particularly Ireland. You might have seen pictures of campaign posters
advocating a certain person for spot #1 and others for #2 or #3. Or one
could observe the preference-swapping agreements in Australia. This is a
direct consequence of later no harm. The parties THEMSELVES share power with
each other. (In spite of this, both countries have two party systems).

My best guess regarding Condorcet/Range's impact is this: the parties will
be campaigning in the same areas more and more, trying to improve their own
position in the same demographic. Campaigns will be less negative because
you have more than one opponent. Attacking that opponent will help your
rivals as well, so it's a waste of your effort. Explicit preference swapping
agreements will be rarer under Condorcet and nonexistent under Range, but
they will be campaigning in other parties' turf when they think they can
improve that group's opinion of them just a little.



Thanks,

Greg Nisbet
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20081011/9adc0607/attachment-0002.htm>