[EM] The path to election reform, was Re:

[Abd ul-Rahman Lomax]

At 08:32 PM 12/2/2008, Kevin Venzke wrote:
>Ok, so sincerity doesn't matter. It's a red herring, it's something not
>to be dwelled upon.

My, my, is this an appeal to the common meaning of "sincerity"? After 
all we've done to point out the technical meanings and to respond to 
others who have called these words "terms of art," i.e., not to be 
taken in their common meanings?

Voting is a choice, not a sentiment. Consider a voting method as a 
black box. It has some controls on the outside. You manipulate the 
controls and you influence the result. How the controls translate to 
influence is well known, with simple methods, not necessarily with 
complex ones.

You have knowledge of the world which will be affected by your 
action. That includes knowledge of the candidates and knowledge of 
the practical consequences of your actions due to the impact of the 
actions of others.

Certainly a voter may push the buttons "insincerely." But that would 
mean, to me, that the voter does not seek to maximize the outcome. A 
vote for a frontrunner in a Plurality election, if the voter would 
prefer someone else, is that "sincere"? Does your answer change if 
write-in votes are allowed? (Usually, voters have someone else that 
they could write in whom they would prefer.)

My contention is that the frontrunner choice is a sincere expression 
of preference within the narrow confines of casting an effective 
vote, making the real choice. It is *not* a sentiment, and, in fact, 
it does not even express any kind of absolute approval. It is a 
choice constrained by practical reality, generally.

The expression of sincere preference is a fairly *easy* method of 
voting, and a system which responds to this well is clearly better 
than one which does not. The question is "responds to this well."

Open Voting (Approval) responds well to the expression of sincere 
preference; however, it is like Plurality in that one cannot express 
sincere preference for one candidate alone without, sometimes, 
concealing a preference. However, the Approval expression is always 
sincere, or may be seen to be so, once it is understood as a set 
preference rather than a single preference. Yes, it takes more 
judgment to make a choice of a set rather than of a single 
individual. However, isn't it a good idea to empower voters who will 
exercise judgment?

In Approval, these voters who take advantage of strategic knowledge 
never, as a result, vote insincerely. (At least not to maximize 
results). Their vote is always a sincere preference for a set over 
the remaining set.) But it does conceal an exclusive preference, 
sometimes. Rarely, actually. (By definition, most voters prefer one 
of the frontrunners and therefore vote for one of them sincerely.)

We can call a vote concealing preference "insincere," but that does, 
indeed, take the meaning quite away from the common meaning. Note 
that if the voter bullet votes for the favorite, the voter is 
similarly concealing other preferences within the rejected category. 
If equal ranking in the presence of a preference is insincere, then 
*every Open Voting vote is insincere.* It's *impossible* to vote 
sincerely, under that view. Which is why the view is actually preposterous.

The question before us is really this: is it better that voters vote 
"strategy free," which is equivalent to voting "zero knowledge," or 
is it better that they adjust their votes according to their 
understanding of the context, i.e., what other voters think.

To me, the gold standard is the degree to which a voting system 
mimics good deliberative, unlimited-ballot, process, where any 
result, however we got there, is validated by a majority before being 
effective.

And in such a process, the voters have interacted with the rest of 
the electorate, in order to discover some compromise. A good voting 
system extracts the information from the votes that allows it to 
imitate this process. It assumes fixed preferences, those expressed 
at the time of voting, which is, indeed, an inherent defect in 
single-ballot deterministic systems. In a deliberative process, the 
electorate learns, it does not merely express preferences. Myths are 
dispelled, facts come out that were hidden or unnoticed before.

In any case, it's obvious that preference strength matters. If there 
is a way to determine absolute preference strength, it would produce 
superior results, with proper use, than any method which doesn't use 
this. Such a method then allows a rough prediction of how the 
negotiations of deliberative process would go. We can predict that 
those with a weak preference will be more willing to compromise than 
those with a strong preference.

Any method which satisfies Later No Harm must not provide this 
information, or, at least, must not use it. This is why Later No Harm 
is actually, itself a poor election criterion. Voters may desire to 
vote Later No Harm, but this is the kind of concealment of preference 
that reduces election efficiency. It is a concealment of true 
preference strengths, so, again, any method which follows Later No 
Harm has *required* the concealment of preferences. We avoid some 
level of strategic voting by *requiring* all voters to vote that way.

But if some voters conceal preference, and others do not, with a good 
method, the results are better than if *all* voters conceal 
preference. Further, a good method allows the expression of all 
significant preferences without harm to the voter. "Harm" must mean, 
not that the voter fails to get their first preference, but that the 
preference strength between that candidate and the winner is not 
minimized. Since IRV is not monotonic, as an example, it can 
seriously harm the voter's interest if that is expressed sincerely. 
The only reasonable defense against this criticism of IRV is that 
this is rare. But, in fact, since Center Squeeze is an example, and 
Center Squeeze in nonpartisan elections, or in elections with many 
parties involved and seriously contending, is not rare, monotonicity 
failure in IRV is not rare. It is only rare in a strong two-party 
system, and *all* the political implementations have been with such 
systems, outside the recent U.S. nonpartisan ones.

Hence the insistence of IRV advocates that its criterion failures are 
rare technical problems is deceptive. It depends on context, and the 
simultaneous claim that IRV will help third parties conflicts with 
it. If it helps third parties, it will create the conditions where 
monotonicity failure is not rare. (Actual failure may still not be 
common, because of *strategic voting.* I.e., we should look at 
strategic voting as a method which voters use to improve the results 
of a poor method.)

Strategic voting, if we allow it to include the deliberate 
concealment of a preference (or minimization or maximization of 
preference strength -- they go together), as distinct from the old 
meaning, reversal of preference, in Range, harms the outcome. It does 
not harm the voter, if done wisely. For the voter, the expected 
outcome is improved. Now, this seems, at first glance, to be a 
problem. It rewards, on the face, "selfishness."

Yet that this does harm can only even be perceived if we have sincere 
Range votes, for all voters, to compare the results with. Range 
strategic voting does not harm results to the point where they become 
as inefficient as those of less accurate methods. Indeed, since a 
maximized vote in Range is an Approval style vote with an accurate 
prediction of who the frontrunners are (or, alternatively, a vote 
which is maximized with the frontrunners, and certain other votes may 
be fully sincere -- accurate -- and may still indicate, in high-res 
Range, preferences, but not necessarily disclosing accurate 
preference strengths), the efficiency of Range must vary between that 
of Approval and all-voters-fully-sincere-and-accurate Range, and the 
mix is monotonic, i.e., the results will improve with the degree of 
accurate expression.

So the "harm" of strategic voting in Range must be seen within the 
context. It's generally minor, I've seen no examples of serious harm. 
We can see this with the example FairVote uses to display Range's 
violation of the Majority Criterion: 99% of voters have, for two 
candidates, ratings of 100 and 99. 1% of voters express ratings of 0 
and 100. Thus the second candidate wins, which FairVote wants us to 
think preposterous: 99% of voters prefer a result, and a measly 1% 
outvote them?

But this makes the point. The vote of the 99% is an almost complete 
abstention. They have said, each one of them, "we don't really care 
which of these candidates is elected." The 1% took them up on this. 
Was the expression of that 1% "sincere"? Perhaps, perhaps not. We 
must treat it as if it is! And that is what deliberative process 
does. It takes the votes as writ.

In an election like this, I'd always have a runoff. This allows the 
majority to really consider, is this what we meant to do? And it 
gives that 1% an opportunity to make their case. "Sure, I know you 
all prefer pepperoni pizza, but I'm Jewish (or Muslim). Sorry, I 
really can't accept the pepperoni." Frankly, the problem here may be 
that it takes 1% of voters to allow this point to be made! Translate 
this to Range 1000. 99.9% of voters vote 1000, 999, and 0.1% vote 0, 1000.

The point is that, yes, if the isolated voters are exaggerating their 
expressed preference, the overall result has not been optimized. But 
by how much did it fall short? Very, very little. In the Range 100 
example, overall satisfaction, assuming that the isolated voter 
really had the same preferences as the others, only reversed, has 
been damaged by 9999 minus 9801 = 198, which is 2.0%. To me, eating a 
pizza, the difference between 100% and 99% satisfaction would be 
undetectable. And even if the 1% voter was insincere, my strong 
tendency would be to say, "If you want it that much, go ahead. Enjoy it."

Would I prefer everyone at a level of happiness indistinguishable, in 
practice, from the idea, or only 99% happy? Given that choice, I'd 
surely choose full consensus. It can be quite hard to find, the 
FairVote example is an easy one! (Note that the Range Votes are very 
close to a tossup overall. Even if they are all sincere, this is a 
difficult choice, and I'd really consider the desirability of 
unanimity to be what makes it easy.)



> > Under many voting systems, a voter may vote a preference
> > that is not the voter's true preference, specifically,
> > that *reverses* preference. The most common form of this is,
> > of course, voting for a frontrunner in Plurality when one
> > would prefer a different candidate *if that option were
> > considered possible.* In the U.S., where write-in votes are
> > generally allowed, this means that *almost always,* voters
> > are voting *insincerely*
>
>But this is subject to a certain interpretation of what sincerity means
>under Plurality.

Indeed. What I've been pushing for is precision in the definition, 
not just the sloppy use of a very hot-button word without careful 
definition and application. Many writers, early on, noted that 
"sincere vote" with Approval did not have a clear meaning, that there 
wasn't just one "sincere vote," there were many. Within that many, 
though, most didn't define some as "sincere" and others not. That 
came, I think, later, where writers started to assume that the voter 
had some absolute approval cutoff, and was concealing that by voting 
differently. This is underneath most of what I've seen as criticism 
of Approval on the basis of "vulnerability to strategy."

We need to define "sincere vote," as it applies to voting systems, 
using a uniform definition that applies to all methods, or we cannot 
determine whether a method encourages "sincerity," or treats sincere 
votes properly. What are sincere votes?

What is clear to me is that Approval encourages *more* sincerity than 
does Plurality. What about IRV? IRV probably encourages sincere 
voting, with some exceptions. However, the problem is how it treats 
these votes. Center Squeeze is a prominent example of how sincere 
voting harms the voter who votes sincerely. Sincere voting, we may 
still need to discuss, but I think we all agree that reversed 
preference is insincere. Open Voting never rewards reversed 
preference. IRV does. Clearly.

>I'll say I agree with you, that Plurality creates incentive to reverse
>preferences, and that this destroys information, all things being equal.

Sure.

>Ok, but you can interpret voters to be voting for lotteries no matter
>what the method is or even if it isn't a method but some other framework
>for a decision. If the action is consistent with what you prefer, then
>it is sincere.

That's correct. The vote is a sincere expression of preference among 
lotteries. So all voting systems, seen in this way, reward sincerity.

Which points out, really, that sincerity is a red herring. What we 
need to do is look at how systems treat (1) preference reversal, 
i.e., clear insincerity, (2) concealment of preference, and (3) fully 
sincere votes. Note that "full sincerity" can be defined with ranked 
methods, easily, because they do not allow the expression of 
preference strength. This places an additional constraint on Range 
methods, since they are the only methods that allow this expression 
-- except for Borda, to some degree, and unreliably. "Fully sincere," 
with Range, means that preferences have been expressed accurately. 
And that, itself leads to a problem.

It can be difficult to express preferences with full accuracy. 
Insisting on lack of consideration of the context greatly increases 
the difficulty, by requiring consideration of all candidates or 
candidate pairs. In ranked methods, especially if equal preference is 
allowed, sincere voting is the easiest voting. Practically trivial. 
But in Range, to vote *accurately*, i.e, zero-knowledge, is quite 
difficult. And I certainly don't expect the vast majority of voters 
to vote that way. And they don't need to.

"strategic voting" in Range, is really easy. In the vast majority of 
cases, there are only two frontrunners, so one starts with Approval 
strategy, the same strategy one uses in Plurality: vote fully for 
your favorite frontrunner, fully against the worst, and you can do 
whatever you feel like with the rest! It simply doesn't matter much! 
So if you like a third party candidate, you can decide how much you 
want to risk the election of that candidate, and if you wouldn't mind 
terribly, give him or her some votes. This is why Warren thinks that 
Range will have an incubator effect. We will see some kind of sincere 
ratings of third party candidates. And a centrist might actually win, 
if most voters prefer that centrist to either extreme among the major 
parties. Rationally, that would mean that they would rate the 
centrist higher than 50%. But "strategic voting" may cause them to 
rate the centrist lower. This is another example of how strategic 
voting can harm results. It is a continuation of the Later No Harm thinking.

>However, this is a completely different notion of "sincerity" from what
>we mean when we say that FPP encourages insincere voting in the sense
>of "order reversal" in the sense of "voting for a candidate who isn't
>your favorite in the abstract sense."

Yes. But what I'm doing here is noting that the common meaning of 
sincerity does not automatically apply to voting systems. The real 
argument is that votes are choices, not sentiments.

If we want optimal results, we need to find ways to encourage sincere 
voting, that's true. But we need that preference strength information 
to optimize results. A runoff, with Range, encourages a shift toward 
sincere voting, since a voter will then attempt to reveal all 
significant preferences. This effect gets stronger as the Range 
resolution decreases -- but, then, runoffs will become more common as well.

> > With ranked
> > methods, an "insincere vote" has a clear meaning,
>
>No, as you just showed, there is no necessity to say that there are any
>insincere votes at all, depending on how you define what people are doing
>when they vote.

That's correct. From the choosing among lotteries perspective, there 
is no question of sincerity. However, here, apparently, I was turning 
to more common usage.

I'd say, bottom line, that what people are doing when they are voting 
is voting. They are making choices and putting them into the voting system.

There are two ways to do this, from the voter's perspective: 
actively, i.e., understand the voting system and the context and act 
to be effective.

And passively: i.e., answer questions.

Except that ballots generally don't ask questions. The Plurality 
ballot does *not* say, "Vote for your favorite." And it would be 
offensive if it did. The Approval ballot *should* not say, "Vote for 
all candidates you approve of." It would be just as offensive! "Vote 
for as many candidates as you choose," would be fine.

Now, what does IRV do?

RCV in Pierce county, from a web site giving instructions:

>You may rank up to three choices.
>    * Mark your first choice in the first column by completing the 
> arrow pointing to your choice.
>    * Mark your second choice, selecting a different candidate in 
> the second column.
>    * Mark your third choice, selecting a different candidate in the 
> third column.
Good. Choices. It does not say "first preference."

If a question is asked or instructions are given, "What is your first 
preference?" Or, "Mark your first preference in the first column, 
your second preference in the next column," etc., the voter would be 
answering the question insincerely if the voter does other than that.

If a Range ballot asks voters to "Show your relative rating for each 
candidate," that question could be considered to be answered 
insincerely if those ratings are distorted.

Now, something cultural. In some cultures, exaggeration is not 
considered lying. In western common law, exaggeration in sales is 
"puffery" and is not considered lying. It's expected. Now, a good 
salesperson, in the long run, is one who gives customers trustworthy 
information. But in the short run, it can be different, and caveat 
emptor. When I ask a salesperson for information, I take into account 
that this person's information might be distored by a desire to sell 
me something. I don't take it as some kind of moral offense, even 
though I'd certainly consider accurate information to be morally 
superior. In fact, when I see that a salesperson is giving my advice 
against their own personal interest, I'll tend to go back to that 
place and that person. That's why it pays off in the long run.

Considering Range, where it really matters (I consider the concept of 
insincere voting in Approval to be preposterous, the method is 
designed not to reward clear insincerely -- which means preference reversal.)

Only with Range, though, does it even become possible to exaggerate 
preferences. And the meaning of "exaggeration" is unclear, we have to 
define what a "fully sincere" range vote is, and that is not 
particularly easy. But we can say, as with Approval, that any vote 
that does not reverse preference is not insincere.

And then we say that "fully sincere" means that preference strength 
is expressed with reasonable accuracy, fidelity to true underlying 
utilities, and thus useful for finding compromises assuming that 
further information cannot be obtained (as with a runoff).

Range makes it *possible* to move beyond the assumption of fixed 
preferences. It does not guarantee that we will, but I don't see 
that, when it does not, it has therefore been a net harm over fixed 
preference systems.



> > it is one which, on its face, supports one candidate over
> > another while, in fact, the voter prefers a different
> > candidate. Voters vote this way in order, we can assume, to
> > the extent that they do, in order to improve the outcome; a
> > good method will not require or reward this. However, there
> > is a class of methods which do not provide any incentive to
> > reverse preference.
>
>But it is only your choice to say that preference reversal is the
>definition of sincerity.

That's the classic usage. I did not make it up. What happened is that 
strategic voting was discussed extensively, with respect to ranked 
methods alone (including Plurality as a ranked method with two ranks 
and only one candidate allowed in first rank, all others are in 
second.) Then, when Approval was proposed as a strategy-free method, 
there was a reaction that involved creating a new definition of 
strategy. This depended on assuming some kind of absolute approval 
cutoff, which the voter then "lies about."

But we all agree that preference reversal is "insincere." The 
question is if there is another kind. With Range, we can define one, 
but it's not at all clear that the voter would insincerely vote 
intentionally. It all depends on the process that the voter goes 
through to decide votes. If the voter starts by determining some kind 
of relative value of the candidates, and then modifies this to 
"exaggerate" some preferences -- which requires minimizing others -- 
then, yes, we could call this "not fully sincere." "Insincere?" 
Depends. And my real point is that by depending on criteria that 
depend on "sincere voting" -- I'll get to that -- we have introduced 
an element of subjectivity.

As Warren Smith has pointed out, whether or not Approval and Range 
satisfy the Majority or Condorcet criteria depends on how you define 
the terms. Approval satisfies some old statements of the Majority 
Criterion, reasonably understood; others have decided that the 
criterion *does not apply* to Approval, i.e., the method is outside 
the definitions, and still others have concluded that it fails. What 
is interesting is to see James A-G struggling to find a definition 
that causes Approval to fail.

See, the context is that *supporters* of Approval often think that it 
is a *good thing* that the criterion fails. So they aren't exercised 
to look at problems with the whole discussion. But James A-G is quite 
aware that there is a problem.


>You can cite an Approval advocate and also James Green-Armytage as
>evidence that this could be the definition of sincerity, but it really
>isn't helpful, if I'm trying to understand why this is a good and
>useful definition.

Essentially, in order to apply a voting criterion based on "voter 
preferences," we have to define the connection between the 
preferences and the votes. If a preference exists, and can be 
expressed, it's not the method's fault if it is not expressed and the 
method therefore fails to satisfy the criterion.

Plurality is generally considered to satisfy the Majority Criterion. 
(Actually, I think Woodall may have concluded differently, but I 
don't have the reference handy.) Yet a majority may have a preference 
and fail to express it, with Plurality, due to, perhaps, a mistaken 
understanding of the election probabilities.

Clearly, to apply the Majority Criterion, we must require that the 
voters *express* the relevant preference. Plurality allows that. 
Therefore it is reasonable to conclude that Plurality satisfies the MC.

But Approval allows the same expression. If a majority of voters vote 
for their favorite in Approval, preferring that favorite to all other 
candidates, and do not confuse or conceal this by voting also for 
another, the majority choice cannot fail to prevail.

So why is Approval considered to fail the M.C.? Believe me, I've been 
around and around this in other discussion. It's because multiple 
approvals are considered sincere votes, so the voter has voted 
sincerely. And voting sincerely has been considered the necessity in 
deciding MC compliance. Why was that the definition?

It wasn't mentioned in the early definitions. In fact, those 
definitions did not even mention the possible gap between voter 
preferences and voter votes.

Why is all this important? Because these terms are bandied about as 
if they condemn a method. "Rewards insincere voting!" "Fails the 
Majority Criterion!" "Vulnerable to strategic voting." I just saw 
someone write that, sure, with "sincere votes," range is an ideal 
method, but "because it is vulnerable to strategic voting, I cannot 
support it for public elections."

> > But there still is an incentive to equate preferences when
> > the voter actually does have a preference. This has been
> > called "strategic voting," in the basic meaning,
> > confusing the hell out of the field, because originally
> > "strategic voting" was used with respect to
> > preferential voting systems and meant preference reversal,
> > and thus strategic voting implied insincere voting. But
> > setting an approval cutoff and voting according to that
> > isn't *insincere*. It simply does not disclose a
> > preference that the voter considers less important than
> > voting effectively.
>
>Ok, so we go back to the other definition. True, there is no need to say
>that any kind of vote is "insincere" as long as the vote is considered
>effective.

Yes. Though that really goes to far. Preference reversal just grates 
against our sensibilities, even though that is not the question asks. 
Methods vulnerable to preference reversal require a kind of cunning 
insincerity.

In fact, of course, this is how voters can improve on a poor method, 
which is what they do with Plurality. Plurality is *awful* if voters 
simply vote sincerely. But most voters don't (though it may be 
unconscious, hence when they encounter a discussion of Range and hear 
that it rewards "strategic voting," they get the idea that this is 
some weird and perverted action. But it's just what they've been 
doing all along.)


> > Much of the fluff about Approval Voting has been based on
> > the concept that there is some absolute Approval cutoff,
> > that it can be assumed that voters "approve" one
> > set of candidates and not another set, and that if the voter
> > votes differently, then the voter is voting
> > "insincerely" and the method can encourage this,
> > and that therefore "Approval is vulnerable to strategic
> > voting." But there is no absolute; what outcomes we
> > "approve" is not a quality of the outcomes alone,
> > but of our assessment of the probabilities of each.
>
>Right. Depending upon what definition of "sincerity" you are using with
>Approval, this is a problem.

Agreed. Spin doctors manipulate concepts to generate knee-jerk 
responses. "Sincerity" Good. "Insincerity" Bad.

> > However, if we accept that restriction, and likewise accept
> > distortion through round-off error caused by practical
> > limitations, only a Range ballot collects the information
> > needed to determine an optimized outcome, given
> > "accurate" ballots, what others have called
> > "sincere" ones.
>
>Ok, we will really need to stick to these terms "sincere" and "accurate."
>I will use "sincere" and "strategic" interchangeably because there is
>really no difference. We must also have the concept of "strategically
>inaccurate."

I use "fully sincere" to refer to accurate Range votes. It's not easy 
to vote that way, even if you want to! But it's only necessary to 
approximate, and voting Range approval style, done with some 
reasonably sincerity, i.e., not hiding a second choice when the 
preference strength is actually small, you would be happy with either 
result compared to your election expectation, isn't even harmful, 
because it will average out over many voters.

> > However, voters won't vote accurate ballots, for a
> > number of reasons. I'd contend that we don't even
> > know how to do it. Rather, we are *instinctively* programmed
> > to consider probabilities. It's relatively easy for me
> > to determine, of A and B, that I prefer one to the other.
> > But then, introduce C. If A and B remain the favorite and
> > least-favored, where do I rate C? When we try to think of
> > Range Voting as involving "sincere ratings," then,
> > we see the purported difficult of Range. There is no
> > specific meaning to those intermediate ratings.
>
>Ok. So Range ballots could permit the collection of information needed
>to provide an "optimized outcome," if the voters are accurate, which
>they won't be, because there is no specific meaning to the ratings they
>can give.

That's not true. It's just that there is not *one* specific meaning. 
It's like Approval in that way. Range *permits* the collection of 
accurate data, but by no means does it guarantee it. Nevertheless, 
because some voters will be able to vote with reasonable accuracy -- 
and we can, if needed, give fairly precise definitions to this, just 
as with Range 2: Preferred, Accepted, Rejected, or Range 3: 
Preferred, Accepted, Disliked, Worst. Or something like that -- Range 
improves results over Approval. If everyone votes in the easiest 
manner (in most elections), it turns into Approval. It doesn't turn 
into a monster. It turns into a very good method, simply not an ideal one.

What Range 2 does is to allow voters to vote based on affinity vs 
aversion, which we are pretty good at detecting. When we aren't able 
to detect the difference, the candidate goes in the middle. So a mid 
rating is an expression of relative indifference. (There is a 
question of whether or not this should be included in a majority 
approval. I'd say it could, because the mid rating is essentially not 
voting on the question of accepting the candidate, it isn't the same 
as voting No on that question. But I'd rather see an explicit 
acceptance or rejection. For example, the mid rating could be divided 
into two classes: I'd accept this candidate, and I'd need more 
information or would reject, perhaps based on lack of information or 
on some mild actual aversion that I decided not to express with a 
negative vote.)

 From my examination of Range 2 polls conducted last year or early 
this year, it was amazingly informative compared to raw preference 
polls or even approval polls. And accurate, comparatively.

> > But when we simply think about Range Votes as fractional
> > votes in an Approval election, that they are weights we are
> > tossing in baskets, and the heaviest basket will win, we are
> > instinctively able to do this kind of analysis. We do it all
> > the time with any goal-seeking behavior. We don't
> > necessarily put our efforts toward the ideal outcome, when
> > we don't think it reasonably possible; rather, we devote
> > our limited resources to an outcome that is an optimal
> > combination of desirability and probability of success.
>
>But that doesn't help us vote accurately, it only helps us vote sincerely/
>strategically.

The point is that, in Range, a strategically maximized vote is a 
reasonably accurate one. Just not perfectly accurate. We are 
efficient by not considering, in our real decision-making process, 
irrelevant alternatives. In fact, this is why we like IIA as a 
criterion. Add an irrelevant vote to a Range ballot, it does not 
change the outcome from one candidate to another. It's only when one 
votes "sincerely" with respect to such a candidate, and therefore 
adjusts all other ratings (but perhaps one),  -- to normalize the 
ballot -- that it can change the result. In other words, strategic 
voting makes Range immune to IIA. And this is, I'm sure, why Relative 
Utilitariasm can satisfy Arrow's theorem, because the shift from 
relative (normalized) absolute utilities ("fully sincere") to 
normalized von Neumann-Morganstern utilities causes the preference 
strength to go to zero as the uncertainty in the preference pair goes 
to zero. I.e., we knew who would win among Gore/Nader. So no voting 
strength (other than an infinitesimal one) would go into that pair.

But we *instinctively* vote this way. We do not necessarily 
instinctively Rate this way, hence a voter, if they vote without any 
thought of voting power, might rate all the candidates low to 
indicate a general disapproval. That's not harmful to overall 
results, probably. But I think most voters, with good ballot 
instructions and public education, will do this. Instructions might 
state that if you do not rate on candidate at max and one at min, at 
least, your vote will have less influence on the outcome, which is 
clearly true and which might help a few voters.

We will do better if Approval is implemented first, because it leads 
voters into thinking about how the system works. It's Range 1. The 
first step toward Range from Plurality. Just fix this one little 
problem with Plurality. Then we will want more, we will want to be 
able to express preferences in a more refined way.

Or we could just start with Bucklin. Simple. Allows the expression of 
preferences (up to three in Duluth Bucklin). Phases into Approval as 
a majority winner is not found with a canvassed rank. Add in the next 
rank votes. Preferential voting method that incorporates Range-like 
characteristics and is not vulnerable to Center Squeeze. At least not 
as seriously as IRV. We still see Center Squeeze if all the voters 
really want Later-no-Harm and insist on it by not ranking anyone else.

(I've proposed, yesterday, a hybrid IRV/Bucklin method that allows 
voters to insist on LNH compliance with their votes....)

> > Because we can conceive of Approval Voting very simply, as
> > representing a decision to support a set of candidates, all
> > of whom are preferred over all non-supported candidates, and
> > because Approval never rewards insincerity in this (i.e.,
>
>Well, it never rewards *inaccuracy* in that respect, no. When you define
>"sincerity" in contrast to "accuracy" then "insincerity" is actually
>never rewarded in any context (unless the actor is mistaken in his
>calculations).

That's right. There is, however, no way to define "full sincerity" 
with Range votes that doesn't involve accuracy of preference strength 
expression. Otherwise they are just Approval votes, which are fully 
sincere with respect to a voter-chosen approval cutoff.

> > including a candidate in the Approved set, when there is
> > another candidate preferred to that one who is left in the
> > unapproved set, or vice-versa), Approval is strategy-free,
> > in the old sense, and this is why Brams introduced it as
> > such.
>
>In this discussion we probably should not use the term "strategy-free"
>except in cases where there are no meaningful decisions.

My point is that it was used that way, in peer-reviewed publications, 
with a lot of agreement, re Approval.

> > The voter sets the Approval cutoff at will, based on
> > election probabilities, presumably, or just on pure personal
> > preference (reasons other than affecting the outcome), and
> > then will vote, we may assume, sincerely with respect to
> > this.
>
>Again, "accurately."

Sure. But in this case, "accurate" and "sincere" are identical. 
(Excepting anomalous votes based on considerations other than 
affecting the outcome, and which are harmless, in general).

I'm claiming, to repeat it, that Approval votes may be presumed to be 
fully-sincere expressions of a division of the candidates based on 
preference and a voter-set approval cutoff, such that we can assume 
that the voter prefers every member of one set to every member of the 
other. If the voter has equal preference, almost certainly, or within 
some small margin, they will place the clones in the same set.

In other words, any error or deviation from full sincerity of 
preference expression is small, down to the point where the voter has 
difficulty discerning the preference.

In Open Voting.

> > This created chaos in the voting systems world; it offended
> > many authors because there was a family of sincere votes,
> > not just one. Suddenly there was no way to take a preference
> > profile and determine from it, alone, a "sincere
> > vote." One needed to make some other assumptions.
> > Messy. But real.
>
>Well, none of this matters much as long as we use consistent terms
>when having a discussion.

Perhaps. But a discussion will also be read by many others. It's 
tricky to use words with a specialized meaning when a larger audience 
will read the text with generalized meanings. Hence "accurate," which 
itself needs definition, is far less a loaded term than "sincere." 
Accurate raises the immediate question "accurate to what?" And that's 
what a reader would need to know.

Example of term used in a specialized way, but presented to create a 
generalized impression that is deceptive: "majority." As in the RCV 
voter information pamphlet for San Francisco: "The winner will still 
be required to gain a majority of votes." It's defensible by giving 
special meaning -- entirely outside the normal meaning -- to "a 
majority of votes." That was really good spin, if it was deliberately 
inserted into the process by IRV advocates. Or did it merely mean 
that the committee that put together the information was itself 
confused? Certainly there was a lot of propaganda that implied just 
that, standard IRV promotional material. It would be interesting to know.

> > Range Voting creates "problems," because it
> > allows the expression of critical information, preference
> > strength.
>
>This is actually a problematic claim. It's trivially true, that Range
>allows this expression. But so does a typewriter or a microphone. To
>simply allow something, is not very valuable.

If the information allowed to be expressed is necessary information 
for making the best choice, it is certainly valuable, potentially. 
Even if the information is not actually used. (As with voters all 
voting Approval style, or so voting in all frontrunner races.) But 
there will be intermediate votes, and, if nothing else, they make 
ties far less likely. Range is a more sensitive method.

There is some reason to think that Range raises the expected utility 
of all voters, whether they vote Approval style or not. This may be why:

Expected utility can be examined by only looking at the races where 
the voter can influence the outcome. Zero-knowledge, the voter may, 
for example, choose to vote Approval style. However, in pure 
Approval, no single vote can change the outcome; rather it can only 
convert a tie to a win or a loss to a tie. In Range, a single vote 
can -- and will, as long as *one* voter has voted an intermediate 
rating -- convert a loss to a win (or a tie to a win). (It can 
convert a loss to a tie as well, but in higher resolution range, we 
could neglect this). Hence it is about twice as likely that a vote 
will change the outcome, hence the voter's expected utility increases.

Indeed, this is consistent with a criticism of Range: strategic 
voting increases the expected utility of the strategic voter, but 
only if other voters vote sincerely. But what was new about my result 
was that even one other voter, voting sincerely, is sufficient to 
make this happen.

> > If voters choose not to express this, they may,
> > under realistic conditions, find some advantage. But what is
> > the alternative? The alternative is much worse: don't
> > allow that expression.
> > Range, when voted
> > "strategically," "degrades" to Approval.
>
>And actually we can substitute the word "sincerely" with no problem.
>
> > Which when voted "strategically,"
> > "degrades" to Plurality
>
>I don't agree with that. I've rarely heard that claimed. You can as
>easily say that it degrades to Antiplurality.

Sure. But antiplurality voting is highly unlikely except in some 
bizarre races. Remember, strategic voting in Plurality almost always 
means "Vote for your favorite," and, in similar situations, that's 
the maximized Approval voting strategy. It's only in real three-way 
races, or with even more frontrunners than three, that different 
votes are strategically maximizing. If a majority is required, there 
is a better way to deal with a really bad candidate who might win. 
Just don't vote for that candidate! If a majority agree, the 
candidate will not be elected. There may be vote splitting on the 
part of other candidates, so we don't know which one of them will 
win, but that presumably doesn't matter to someone tempted to vote 
for all but one.

> > And Plurality is actually a much better method than
> > we've given it credit for,
>
>I agree with this statement without even the qualifications that follow.
>
> > when it is used within a
> > generally functional political system.
>
> > In any case, isn't it suspicious that voting for the
> > favorite is considered "sincere" with Plurality,
> > and not with Approval?

>Maybe. As I tried to argue in a previous post, it requires assumptions
>to say that a vote or specific manner of making a decision can be
>"accurate" in some sense in the first place.

Sure.

> > What critics of Range would have us do is to continue to
> > forbid the expression of preference strength, on the
> > argument that some of us won't do it
> > "sincerely," and thereby gain some advantage.
> > Simulations show that, as I'd expect, Range continues to
> > perform well with various levels of "incomplete
> > expression," or "strategic voting."
> >
> > Because we can't be sure that all voters will vote
> > "sincerely," something which is never,
> > conveniently, defined, we should prevent all voters from
> > voting accurately?

When I've made claims on this, I've generally been careful to define 
"sincerely," I will usually say, "fully sincere," which only takes on 
meaning with Range, where preference strength is expressed. In higher 
resolution Range, yes, there is no specific meaning. In reality, 
these are just votes. You have up to ten votes, Range 10, to give to 
each candidate. The candidate with the most votes wins.

They are *not* ratings, and they are *not* utilities. But we can 
derive them from utilities, if we have the information. And we can 
derive, as with Approval, preferences from them, but we cannot derive 
*accurate* preference strengths for each voter, because the votes 
will be distorted toward Approval style, in general. However, those 
distortions will tend to average out, and the degree to which voters 
vote toward the accurate end will likewise shift the result toward 
the "fully sincere" result.

Not allowing the information does not improve the overall results, it 
simply makes it impossible to even consider the differences. Borda, 
by the way, gets around this, using a pure preferential ballot. But 
that has well-known problems, which are resolved by allowing equal 
ranking and creating empty ranks whenever a voter ranks equally. 
I.e., it becomes Range Voting. Which is better, to assume equal 
preference from rank to rank, or to allow voters to vary this.

In reality, preferences vary, often greatly. An optimal system needs 
this information. Or garbage in, garbage out. People will still use 
distorted ratings, but Approval style isn't necessarily insincerely 
distorted. It's merely that the voter didn't take the trouble to 
express more accurate ratings -- it's *harder* -- or, alternatively, 
is well-informed and chose to vote Approval style with the 
frontrunners. (There is no particular reason to vote that way with 
moot candidates.)

The point is that the additional preference information available in 
Range does not harm the outcome over not allowing that information. 
To think otherwise, we have to assume ignorant voters who don't know 
the difference between a strong and effective vote and just vote 
"sincerely," however they define that. But these voters help the 
outcome, it's just that they don't maximally help themselves. If they 
care enough about the outcome, they will vote with more strategy, 
i.e., in general, give strong votes to frontrunner pairs.

>The answer probably depends on which critics you're talking about.

Sure.


> > Because some voters may "get what they want" by
> > voting with full strength -- I thought that seeking to get
> > what they want is what voters are supposed to do! --
>
>Yes, if they're being sincere/strategic they are supposed to do that.

It's what we are totally accustomed to, and it's why Plurality works 
as well as it does.


> > we
> > should require all voters to accept results that are,
> > overall, inferior to Range Voting?
>
>Well, you argued earlier that voters will only vote sincerely, and don't
>know how to vote accurately.

Not with *full accuracy.* But, I'll contend, they will know how to 
vote with reasonable accuracy. And voting Approval style is 
reasonable accuracy, for the most important candidates! They can vote 
more accurately than that, to be sure. And in some cases, a mid 
rating would be more appropriate, and easily discovered by the voter 
and sincere, than an extreme rating. In this sense, Range 2 makes it 
easier to vote. With improved accuracy.

>  You stated that the various ratings don't
>have inherent meaning. That's why accuracy is a red herring, surely.
>This makes it confusing that you want voters to be allowed to do something
>that they 1) cannot do and 2) ought not to do if I take literally your
>last comment that voters are supposed to be trying to vote effectively.

But voters can rate with improved accuracy, the Range 2 example shows 
it. It's actually instinctive.

Watch out about debate tactics. Aha! Got you! You were inconsistent. 
I'm not here to debate, I'm here to explore, on behalf of myself and others.

*Some* voters can vote with improved accuracy. Others might just vote 
100% for their favorite and leave it at that. Depends on the voters 
and depends on the context. For the large majority of voters, under 
present conditions, voting that way is actually quite effective 
strategically. But there is a significant minority of voters who need 
more; in particular, those who prefer a minor party candidate to any 
of the major party candidates, but who have a significant preference 
between the major party candidates. It is these voter who need 
additional options, might be as high as ten percent or so.

Ironically, FairVote has criticized Bucklin because in some 
elections, the use of additional preferences was confined to about 
ten percent. But that's quite enough to fix the spoiler effect! They 
also criticised Bucklin because these elections were not finding a 
majority for the winner. But, of course, Bucklin comes closer to 
finding a majority than IRV, almost certainly, and in those 
elections, it's quite likely that the only majority IRV would have 
found was the false "last-round" majority. As is almost always 
happening whenever "instant runoffs" are needed, no majority is found 
after vote transfers.

And, if I'm correct, IRV was discarded, also in primary elections 
where used, because there was too little use of the additional ranks. 
FairVote simply casts about for arguments to use and it does not care 
about fairness and equal application. It's UnFairVote.

> > (1) Start counting all the votes, and the candidate with
> > the most votes wins. I have yet to see anyone knowledgeable
> > about voting systems who thinks that this would be a step
> > backwards!
>
>While I don't think it is a step backwards, I can imagine some arguments
>against it. I think of the reasons why FPP isn't as bad as it ought to
>be in theory, and consider which of those reasons might not apply under
>Approval.

Great. I still don't see any objection there, merely the claim that 
there might be one.

Okay, Approval may encourage third parties. This is a real objection 
which has been used, in the past, against both Bucklin and IRV. It's 
false with IRV, we are pretty sure about that. Bucklin might be the 
same, actually. (All this is irrelevant with nonpartisan elections, 
by the way.) Approval, I'd guess the same. Not much of a real help, 
because it takes away the stinger that third parties can threaten 
major parties with. Listen to us, or we will run a spoiler. Range, 
though, does something else, a little better than Approval. It allows 
a measurement, not perfect, but better than what we have, of real 
support for a third party, because the expression of this is safe. It 
would allow major parties to detect a third party sneaking up on 
them, and to become more responsive to the positions of this party 
before it whacks them upside the head in a surprise election. They 
can tell by the votes how many of *their* voters are impressed by the 
third party and might switch. It may not eliminate the party system, 
but it will surely make it more responsive.

Want to undermine the party system, making it unnecessary? Asset Voting.

It undermines it without attacking it. It simply makes it unnecessary 
to gain a party endorsement to exercise real political power. It 
could *create* new parties, ad hoc, as needed. Etc.