[EM] 05/13/02 - The Education of Poor Richard:

[Michael Rouse]

----- Original Message -----
From: "Craig Carey" <research at ijs.co.nz>
To: <election-methods-list at eskimo.com>
Sent: Tuesday, May 14, 2002 7:34 AM
Subject: Re: [EM] 05/13/02 - The Education of Poor Richard:

>
>  >                  There are other important criteria that should be
>  >considered: the Majority Criterion, the Monotonicity Criterion, and the
>  >Independence of Irrelevant Alternatives Criterion, to name a few.
>  >
>
> There are 3 there and 2 are not important I presume. The Majority Criteria
> is too weak but that you can tell us how a fix of that problem could
> proceed. The "Independence of Irrelevant Alternatives" test is not
> important at all, unless you have an uncommon definition.

It isn't important that introducing a losing candidate can affect the
winning candidate? "I prefer Joe to Bob as my representative. I hate Jack,
so if he gets in the race I would prefer Bob to Joe." Granted, it may be
impossible to fix (depending on which of Arrow's criteria you follow), but
it does not affect all voting methods equally.

> Mr Rouse might not have been reading my messages closely enough when I
> was last subscribed.

It depends on the title of the message and (if that looked interesting) what
the body of the message said. I'm probably not alone in killing certain
threads.

>  >there is a circular tie, where the "irrelevant alternative" is part of
the
>  >cycle. Unfortunately, IRV violates IIAC as well, *and* it violates
>  >monotonicity. If the possibility that your second-place vote could help
>  >defeat your primary candidate is a problem, how much worse is the
>  >possibility that greater preference for a candidate can defeat him?
(Others
>
>
> Mr Rouse might mean the "second preference" rather than "second vote".
> Anyway the meaning of the words, "primary candidate", seem to be too
> unclear.

primary (adj)
  1.. First or highest in rank, quality, or importance; principal.
  2.. Being or standing first in a list, series, or sequence.
  3.. Occurring first in time or sequence; earliest
By primary candidate I mean the candidate that you rank above all others and
place first on an ordered list of candidates. And I mentioned "second-place
vote" (second on the ordered list of candidates) rather than "second vote."
"Second preference" is certainly an alternative way of saying this, I agree.

>
> What exactly is unfortunate about the Alternative Vote violating the
> 1950s axiom of Independence of Irrelevant Alternatives ?. A first thing
> to do on the 1st night of hearing of something as implausible as IIA,
> might be to see if it takes less than the 17 papers that prove the
> Alternative Vote non-monotonic.

Since we are striving for maximum clarity, by "papers" do you mean
scientific papers, newspapers, or official documents? Because if you mean
"votes," you should be aware that "papers" and "votes" are not generally
used as synonyms.

> Where do the transient and false
> comments that IIA has some merit, come from ?. Please answer my question.
> It could be speculated that false information is taught at American
> universities.

You have heard of Kenneth Arrow, 1972 Nobel Laureate? Apparently, the Royal
Swedish Academy of Sciences was fooled by this false information. Granted,
there are many who disagree on how important it is, but I was unaware of any
proof that IIA was false.

> In the 3 candidate Alternative Vote, if equal quantities of the papers
> (A), (B), (C), are added, and if the quantity added tends towards positive
> infinity, then the percentage difference between the 1st and last
candidates
> (by a counting that considers only the 1st preferences), narrows towards
0%
> even though the central candidate of the 3, can win.

If the right-wing and left-wing candidates are able to moderate their
message enough so that the center candidate has fewer first-place votes, the
center candidate is eliminated in the first round, just as he is now in
FPTP.

>
> As soon as the uniform distribution is said to be used, then the results
> can be rejected.

As mentioned before, you can use Gaussian (bell curve) distribution, and it
only moves the left/right-wing extremes, it does not remove the effect of
squeezing out the center candidate.

> Anyway, why is any respect being held out for a test that
> is not in the list of what was said to be important at the top of the
> message I reply to. It sounds like a completely new test rather than a
> corollarly or something.

I was mentioning why I did not like IRV, and gave an example of one problem
at the end of my message.

> How is it that Mr Rouse did not mention truncation resistance?.

Because I did not see it on the "list of what was said to be important at
the top of the message I reply to." Perhaps if you gave an example of how
IRV is superior in truncation resistance I might respond.

Michael Rouse
mrouse at cdsnet.net

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em