Choices

[Mike Ossipoff]

Those are probably the only balloting methods.

[I wasn't going to send this letter to the list, because it's off of
our subject, since it's about features that I don't include in my
proposal. Still, I consider these things worth mentioning at some
time, so I've now decided to Cc this reply to the list. Since the
"devious electorate" features described here aren't relevant to
what the SWC is doing, you won't miss anything immediately relevant
if you delete this letter, or set it aside for some other time when
there's nothing to do]

There are small features that could be added, of course. For instance,
a Condorcet election, with a devious electorate, could add an option
for the voter to indicate that, if every alternative from which Condocet
is to choose has another alternative ranked over it by a majority, and
if the circular tie includes alternatives over a certain point in that 
voter's ranking, and alternatives below that point, then the voter
wants everything below that point dropped from his ranking. I call
that the "circular tie option". I don't include it in my proposal, since
it's only useful with devious electorates, and because it's an added rule.

Similar to the circular tie option is the "consistency option". If
a voter punches the box that chooses that option, then in the event
that every alternative in the set from which Condorcet's method is to
choose has another alternative ranked over it by a majority, and
if there's inconsistency between some rankings (meaning that they
disagree on what alternative is between other alternatives, merit-wise),

and if something higher in that voter's ranking would have won if
some of the rankings were changed so as get rid of the inconsistency,
then that voter wishes to drop from his ranking the alternatives 
that are the 1st choice of the people who voted those rankings.

There are other options like that, some based on unanimity vs
non-unanimity as a way of detecting order-reversal, some, as in the
above paragraph, using inconsistent rankings as an order-reversal
detection device.

So far as I'm aware, other than these options, there aren't any
proposals for balloting other than the 3 that you mentioned.
Of course the y/n balloting could be considered a special case
of rankings, and it could also be considered a special case of
point-balloting. So really, you could reduce your list of balloting
methods to 2.


Mike



--