Rated voting puzzle question

Saari at aol.com Saari at aol.com
Tue Apr 22 07:55:35 PDT 1997


Donald writes:
>     In an Apporval Voting election a voter's lower selections receive
>support while the first selection is still a contender. Savvy voters will
>see this in Approval Voting and will refuse to make more than one
>selection.

Not all savvy voters will follow your particular sentiments.  I for one,
after due consideration, will likely vote FOR several candidates (all of whom
I approve of) when the circumstances justify same.  Voting for only one is
putting all of my eggs in one basket - I am more interested in seeing that
somebody reasonable wins rather than just going for my top choice.  I suggest
you not be so quick to decide how others will behave.  See also the next
section for another justification as well.

John writes:
>If what Donalds says is true, his post herein makes such eminent sense that
>discussion of  Approval methods should be immediately dropped as being
>worthless.

Another consideration involves the scoring method.  While Approval uses
simple addition and "best score wins" there are other alternatives.  For
example, most experts assume that nobody would vote support,support or
oppose,oppose in a two-candidate race - claiming that such a vote has no
effect and is thus irrational.  This implies that such votes (support,support
vs. oppose,oppose) have an identical effect on the outcome, i.e. none.

My desired "rated voting" system would distinguish such votes, thus giving
them meaning.  One way is to require a "minimum score" for ANY candidate to
win - otherwise a new election is held or the spot is filled by appointment
or the spot is left vacant or some other predefined fallback.  Therefore, a
vote of support,support means "I think both candidates are fine." whereas a
vote of oppose,oppose means "I reject both candidates (and prefer the
fallback)."  These would each be plausible votes under some circumstances (in
addition to the more typical support,oppose or oppose,support).

I also advocate a 3-level vote at a minimum, i.e. allowed votes are
"Support", "Oppose" or "Neither".  This allows much more expression than
basic Approval voting (which does not distinguish disinterest from active
opposition).  This to me is a serious failing of Approval voting - lack of
sufficient resolution.

Finally, I advocate the allowance of fractional votes, i.e. 1/2 support or
2/3 oppose, etc.  This allows adequate "fine-tuning" to distinguish two
candidates, both of whom are liked but one of whom is more preferred.

>Voters have absolutely NO interest in making a SECOND selection unles they
>can be assured that it doesn't color the count relative to their FIRST
>selection.

Speak for yourself but please don't speak for the rest of us.  Perhaps some
voters can only think clearly about their top choice - others of us like to
think about all of the candidates.

Here's a puzzle question.  Assume you are voting on which movie to go see
with 2 other friends.  There are a dozen choices.  You are allowed to vote
support, oppose or neither on each one, or fractional votes - simple addition
is used to decide the winner.  If no choice gets more support than opposition
then no movie at all will be seen.  6 of the 12 movies (A,B,C,D,E,F) appear
likeable to you - and A looks like the best overall.  2 (G,H) look so-so, and
4 (I,J,K,L) you consider unwatchable to varying degrees.

Leaving aside the question of whether this is a sensible method or not, I
simply ask what voting "strategy" would you use if these were in fact the
rules.

I am curious in particular if the two previous commenters would still follow
the strategy which they claim is the only "savvy" one.

Thanks,
Mike S





More information about the Election-Methods mailing list