[EM] I eat my words (but not wholly)

Warren Smith wds at math.temple.edu
Fri Oct 14 21:37:21 PDT 2005


I don't get it.  Supose there are 3 candidates.  Do you really regard it
as inconceivable that all voters could, without effort, learn enough
about all 3 of them to rank them?  I think that could easily happen.

Let us suppose that they are "10000 people die", "5 die" and "eternal happiness."
Now is it really true that you are just too damn lazy to rank them?  You
insist on mentioning just one in your vote?  It seems to me there is
a substantial subclass of elections in which everybody knows their full
ranking, and everybody feels like an idiot to behave in Kislanko-laziness-style,
and that feeling of don't-want-to-be-an-utter-idiot outweight the feeling of
want-to-save-1-second-of-work.   In that case, voters will
rank everybody.   

I have no study to cite. I simply regard it as obvious.

Next, I rather dislike the idea that a method with truncation allowed
can be considered in the same breath as "condorcet methods" at all.
A lot of authors simply rfuse to even consider ballot truncation or
equal-rankings-permitted.  Maybe they are dumb to do that, but they do it,
and then when they talk about "condorcet method" they are referrring to 
ranked ballot methods.  period.  Not "partial ranking maybe if I feel
like it" ballot methods.  The opposite approach favored on EM is to
allow total truncation and total equality.  According to EM, then,
the following method:
  1. the only vote you are allowed to submit is ther maximally lazy
vote in which A=B=C=...=Z.
  2. we will pick a random winner.
Is a "condorcet method" and somebody like me has no right to object to
that phraseology.  But I do object.  I think it has to mean something,
and I think the more truncation you allow, the less the word has any meaning.

I also think, there is a reason the Australians made truncation illegal.
Maybe you disagree with it, but obviously a large number of people in the
world feel truncation is bad, even just haviing one voter truncate one vote is bad,
and if your method refuses to cater to that feeling, well, then you
are sacrificing something.
(By "refuses to cater" I mean, "maassively screws up in response to this
voter behavior" in many cases.)

wds



More information about the Election-Methods mailing list