[EM] Abd: Condorcet vs RV

[Michael Ossipoff]

Abd said:

Warren has a tendency to hyperbole, but his point should not be missed. 
Unless and until it is confirmed, his work must be taken with some salt, but 
it nevertheless does show what he claims. It appears that some are 
criticizing it without being familiar with it, perhaps on the basis that 
they find the results unpalatable.

I reply:

I’m not familiar with the details of Warren’s simulations, but he told what 
his assumptions were, and that was enough.

Abd continued:

But the *huge* problem with a great deal of EM evaluation is that it is so 
often based on election criteria, presumed to be desirable, rather than on 
some kind of measurement of success in *results*.

I reply:

That’s a common objection, as if the results of a simulation can be taken as 
experimental proof of what would happen in real life.

Abd, criteria are about results. Ok, there’s such a thing as a 
rule-criterion, but I have no use for those. I only value results-criteria. 
They’re about things that will never happen, or will always happen.

Presumed desirable? You can decide for yourself what is desirable to you. 
That’s what determines what criteria you choose, and, thereby, what 
method(s) you prefer.

Abd continues:

Warren's work is not based on any specific assumption of how voters vote. 
Rather, it applies various stated assumptions, showing the effect of sincere 
voting, strategic voting, and mixes, to a large number of sample elections.

I reply:

>From what he said, Warren ignored the fact that people can vote more 
sincerely, without regretting it, in Condorcet than they can in RV.

In some of his work he assumed that no one ranks anyone equally.

In some of his work he assumed that voters have such good predictive 
knowledge that RV elects the CW every time.
Abd continues:

As he noted, the result that Range is more successful at picking the sincere 
Condorcet winner than Condorcet methods is, to say the least, 
counterintuitive. The problem, of course, is strategic voting, where voters 
do not express their sincere preferences, fearing that if they do, the 
result will be less satisfactory than if they shift them. One of the 
characteristics of Range is that there is never an incentive to reverse 
preferences. Strategic or insincere voting in Range is limited to equating 
candidates at the extremes when, in fact, you do have a preference between 
them. Warren calls this partially sincere. I've called it "magnification;' 
the gain on an amplifier has been increased until the output pegs the meter. 
The contrast on an image has been increased until some detail is lost in 
total black or white. It isn't insincere in the same way as preference 
reversal is. Warren likes Range with unlimited resolution because it becomes 
possible to express fine differences in preference while retaining 
effectively full voting power. (It is true that if you downrate B from A by 
a thousandth of a point, you could cause B to lose to C. But the possibility 
becomes vanishingly small. One way to look at a range vote of 0.999 -- 
compared to 1.0 -- is that it is one-thousandth of a vote different. 
Warren's results though, were with limited Range resolution. I think he has 
done work with Approval, with Range 10 and with Range 100. But the Condorcet 
Criterion itself is severely limited.

I reply:

The Condorcet Criterion isn’t one of my favorites, for reasons that I’ve 
discussed on EM before. But it’s still a good thing to have, as has been 
demonstrated on EM when we did polls using various voting systems.

Abd continues:

Once one realizes how it is quite easy for the best candidate (from a social 
utility point of view, which is really the only criterion that has been 
suggested that actually measures election performance)

I reply:

Sorry, Abd, but that’s incorrect. All criteria measure some aspect of 
election performance, though you might prefer some to others.

Abd continues:

to not be a Condorcet winner, the Condorcet Criterion can be seen as less 
than ideal.

I reply:

Abd, I’ve repeatly pointed out on EM that the CW tends to be the SU 
maximizer, suggesting that Condorcet is likely to do better than RV with 
regard to SU.

If voting is spatial, and we measure distances by city-block distance (as 
I’ve argued that we should), then the CW is always the SU maximizer. If we 
measure distance by Pythagorean distance, often called Euclidean distance, 
the CW is the SU maximizer under the conditions commonly assumed in 
simulations (and I know you have great respect for simulations). For 
instance, multidimensional normal distribution of voters in issue-space, or 
uniform distribution.

There can be non-spatial disutility that can cause a CW to not be SU 
maximizer. But the CW position would be a popular one, and there would 
surely be candidates there without non-spatial disutility.

Abd continues:

Generally, the neglect of preference strength is the elephant in the living 
room of ranked methods.

I reply:

It’s the advantage of rank balloting. You prefer rating because it lets you 
give more information. Is that always good? Having to give information isn’t 
always a good thing. Do you want to fully rank X over Y, so that you’re 
fully voting X over Y? If so, then you can’t rate Y over anyone in RV. Do 
you really want to give that information? In rank balloting, you indicate 
that you like X better than Y, and Y better than B. Condorcet fully counts 
both preferences that you voted, as you would want it to. I don’t want my 
X>Y preference weakened by the fact that I want to vote Y over Z.

As another example of it not always being a good thing to have to give 
information, how good is it if a mugger asks you for information about 
whether you’d rather give him your wallet or take the consequences that he 
offers you? Having to give information can give you a choice that you’d 
rather not have to make, whether the information is required by RV or by a 
mugger.

Abd continues:

It is trivial to give simple examples that show, beyond doubt, how Condorcet 
can choose the wrong winner.

I reply:

…wrong according to another method. I presume that you mean a low-SU winner. 
I’ve just finished answering that objection in the posting to which you’re 
replying. If someone at the CW spatial position has non-spatial disutility, 
then surely there will be others at or near that position who don’t.

Abd continues:

What is really odd to me is that it seems to have taken so long to recognize 
this. Does anyone know when this began to show up in the literature?

I reply:

That objection has been quite common for as long as I’ve participated in 
voting system discussion. It isn’t new.

Look, if you prefer what RV offers, then RV is for you. If you don’t want to 
have to choose between fully voting X over Y, and fully voting Y over Z, 
then RV is not for you. Make your choice based on what you want, but don’t 
criticize others’ choice when it isn’t the same as yours.

Abd continues:

Once a method allows preference strength to enter into consideration in 
determining the winner, the method *must* fail Condorcet. (With the limited 
exception that if preference strength is only used to choose between members 
of a Condorcet cycle, there need be no Condorcet violation.) Likewise the 
Majority Criterion, as generally stated, will not be satisfied in all 
elections. However, quite arguably (I'd say *clearly*), Range fails 
Condorcet in order to choose a better winner.

I reply:

Newbies commonly say that what they like best is “better”. RV is better if 
you prefer what it offers. Prefer what you want, but don’t claim that it’s 
better in some absolute sense. RV’s results lose their meaning if voting 
isn’t sincere. And don’t say that strategic voting in RV will easily elect 
the CW, because that depends on how well-informed the strategic voting is. 
As I said, Riker showed that _any_ method will always elect the CW if the 
electorate is sufficiently well-informed and votes strategically. So that 
hardly counts as an RV advantage. A method’s merit can be judged by how it 
is when voters don’t have complete information.

Abd continues:

This is absolutely clear if all voters vote sincerely and accurately.

I reply:

For one thing, not all can be expected to vote sincerely. For another thing, 
if you’ve ever taken part in an EM poll, you know that “accurate” rating in 
RV is easier said than done.

Abd continues:


It is muddier in the presence of strategic voting (the definition of which 
becomes somewhat problematic with Range), but Warren's results show what 
some of us anticipated: strategic voting doesn't do nearly as much harm as 
some Range critics have claimed it would.

I reply:

…under Warren’s assumptions.

Besides, though there’s a tendency to speak of how well a method resists 
being adversely affected by strategy, you might want to consider to what 
extent a method forces strategy on a voter, making him/her regret sincere 
voting.

Mike Ossipoff