[EM] Re: Falsification & "completion"

MIKE OSSIPOFF nkklrp at hotmail.com
Fri Jan 23 00:37:01 PST 2004


Kevin--

You said:

I think you may be confused about the process of Symmetric Completion
as different from the criterion.

I reply:

I'm not interested in what you think. It would be more worthwhile if you 
could tell us why you think what you think.

Did I say something that suggested that I believed that  the process and the 
criterion were the same, or that "symmetrical completion" was the name of a 
process but not a criterion, or that "symmetrical completion was the name of 
a criterion but not a process?

One thing that I said was: I pointed out that if SCCIRVE meets the Symmetic 
Completion Criterion, that's because it does symmetric completion. In that 
sentence I spoke separately of the process and the criterion, speaking of 
them as 2 separate different things. Though it isn't necessary to do the 
symmetrical completion process to meet the Symmetric Completion Criterion, 
it's obvious that that is one way to meet that criterion. I was pointing out 
that a method that does that process will meet that criterion.

Is that clear to you now? For you to make your your above-quoted statement,  
after reading my sentence that mentions separately the process and the 
criterion--your statement apparently resulted from confusion on your part. 
More carefully reading what you're replying to would help you to avoid that 
confusion in the future.

You know, if you want to reply to a message, it would really be great if 
you'd check that your intuitive or emotional feelings about it are justified 
by things that are actually said in the message that you're replying to. One 
thing that would help you to do that would be if you would state, in your 
repy, your reasons for what you say. In that way, you could be sure that 
there _are_ reasons for what you say. You could thereby avoid posting 
statements for which you can't give a reason.

You continued:


The criterion says that the method
gives the same results whether or not the ballots are completed
symmetrically.

I reply:

Thank you Dave :-)  In the message that you were replying to, I'd just 
stated that guess about what the criterion says.

You continued:

This is met by FPP, IRV, and Margins, none of which
employ the *process* of Symmetric Completion.

I reply:

...and so?  Did I say that SCC is met only by methods that employ the 
"process" of symmetric completion? Apparently you're all confused about this 
too: Saying that doing the process causes a method to meet the criterion is 
_not_ saying that doing the process is the only way to meet the criterion. 
Does that help you?

You continued:

So if we are talking about the merits of Symmetric Completion as a
criterion, it is irrelevant whether Symmetric Completion is objectionable
as a process performed on the ballots, because the criterion doesn't
depend on the process.

I reply:

But is that relevant to a statement of mine? What I was saying that the 
"process" is a falsification of preferences. Here we've defined 
falsification of a precerence as the voting of a preference that that one 
doesn't have. Your "process" modifies a ballot so that it falsifies a 
preference by that plausible and reasonable definition. I was saying that 
it's silly to claim that, in general, it makes sense to do that to ballots, 
though I don't care what you do to them in your method.

But I did NOT say that the falsity of your "process" is the reason why the 
criterion is objectionable. I noted that Chris said that Woodall doesn't 
justify the criterion. I asked Chris if he could tell us why it's important 
or even desirable for a method to meet the criteriion. We haven't heard from 
him on that, so apparently he doesn't know why he believes that it's 
important for a method to meet that criterion.

But, that being said, you're mistaken when you say that the wrongness of the 
falsification doesn't say anything about the merits of the criterion: A 
criterion that says that a method should give the same result after you 
falsify some preferences, falsify some of the rankings,  is a criterion that 
lacks merit if we agree that falsification isn't a good thing.

You continued:

According to the mentality of Symmetric Completion (criterion),
two voters voting A=B>C>D should have exactly the same effect as one of
the two voting A>B>C>D and the other voting B>A>C>D.

If Mike thinks this is flawed thinking, it would be interesting to hear
what he thinks the difference in effect should be.

I reply:

For determining whether or not A pairwise-beats B, or whether B 
pairwise-beats A, there should be no difference.

But where you're confused is: You believe that if those 2 scenarios should 
have the same effect in determining a pairwise defeat between A & B, that 
means that they should have the same effect for all purposes.  And no, 
determining pairwise defeats is not the only purpose for which we use the 
voters' pairwise preferences. Apparently you haven't been paying any 
attention to any of the Condorcet discussion on this list. It would be much 
better if you'd lurk for a while, to get an idea of what we're talking 
about, before you start expounding, and explaining the subject to us.

A clue for you, Kevin: Pairwise-count methods don't just look at who 
pairwise-beats whom. They also look at what we call the "magnitude" of the 
defeat. We also call that the "strength" of the defeat. The Condorcet 
proposals that our Condorcet versions are based on involve sequentially 
dropping weakest defeats or sequentially keeping strongest defeats. Does 
that help your confusion?

Maybe not, and so maybe I'd better spell it out better for you, Kevin:

We mostly discuss 2 kinds of Condorcet versions: Winning-votes, which we 
abbreviate "wv", and margins.

Winning-votes says:

If X beats Y, then the strength of that defeat is the number of people who 
ranked X over Y.

Margins says:

If X beats Y, then the strength of that defeat is the number of people who 
ranked X over Y minus the number of people who ranked Y over X.

As you pointed out, or suggested, with margins a voter who votes X over Y 
cancels out a voter who voted Y over X, in calculating the margin, as well 
as in determining which candidate pairwise beats the other.

But, that isn't true of wv. So: You asked me how those 2 voting scenarios 
you described could give different results. The answer is that they give 
different results in wv. Incidentally they also give different results in 
some other methods too.

To you, what that means is that wv fails your criterion (You know, the 
critrerion that neither you nor anyone else has justified).

Unless you can show why it's important that a method meets SCC, we can 
assume that it doesn't matter whether or not a method meets SCC.

There are reasons why many of us prefer wv to margins. Those reasons have to 
do with the results of the methods (for some keys the upper case only is 
printing so i have no period comma or upper case letters ">" now means 
period and "<" now means comma>

these reasons have to do with the methods" results and their properties< and 
with criteria that they meet>

and they have to do with majority rule: some of us< you see< believe that it 
means something if a majority of all the voters have indicated that they 
prefer x to y. apparently you and your woodall don"t consider that 
important>

you"d say that if we take a vote between x and y< and if there are a hundred 
voters< and if two of them say x is better than y< and if one of them says 
that y is better than x< you"d say that we should count fifty and a half 
people as saying that x is better than y< and so x has a majority win over 
y> (fifty and a half is two plus half of ninety seven)> those ninety seven 
voters who said that they don"t have a preference between x and y are 
counted by woodall< and therefore by you< as saying that they prefer x to y< 
and that they prefer y to x> so< by "symmetrically completing" their refusal 
to express a preference< we get half of ninety seven votes for x over y in 
addition to the two who actually voted x over y>

you ask me how "symmetrical completion" gives a different result? that"s how 
it gives a different result>

i hope that i"ve answered your question< and showed you the consequences of 
that criterion that you copied from woodall> uncritical copying can result 
in adoption of some pretty silly things>

now if i move the cursor< this computer wants to highlight everything over 
which i move the cursor>hopefully that won"t prevent me from sending this 
message>

you said:

my interpretation is that a faction shouldn't spoil the election for 
themselves
because they opt to express strict preferences among their favorites, 
instead of
using approval strategy.

I reply:

That sounds like a desirable goal< though you have yet to show that it"s 
more important than other desirable goals> you see< kevin< many criteria are 
possible> many of them are mutually incompatible< and so it"s necessary to 
choose which we consider more important>

what you said in the paragraph before last is similar to the "acceptibles 
equal ranking criterion (aerc) that i was going to define< and which i"m 
still going to define after answering these messages> maybe tonight< maybe 
tomorrow>

but what a ridiculous way to try to gain that result< by falsifying 
preferences> do you want to meet your goal via the scc criterion or via the 
symmetric completion process< or are you confused about the difference 
between them?

i"m going to send this now< if that"s possible< and continue on a different 
computer< because this computer has done some unwanted deletion>

moving the cursor highlights text< and then using the return key deletes the 
highlighted text>

Kevin Venzke

_________________________________________________________________
Find high-speed ‘net deals — comparison-shop your local providers here. 
https://broadband.msn.com




More information about the Election-Methods mailing list