[EM] Top-majority options for Approval? An alternative 3-slot option.

[MIKE OSSIPOFF]

I was speaking of MTAOC and MCAOC as options in an Approval election, but those names of free-standing methods probably aren't 
good names for voting-options. I'll collectively refer to those two methods as "top-majority methods".

(Of course, MTAOC and MCAOC are still good as separate free-standing methods)

That
 sort of voting works in an Approval election, more or less. The election 
could be regarded as a top-majority method election in which
the plain Approval ballots are treated as having top-rated the candidates whom they vote for. 

Of course MTA and MCA differ only when there are 2 or more top-majority candidates.

Since
 I'm talking about a voting-option, maybe the voter needn't commit 
hirself to giving or not giving votes to hir middle-rated
candidates 
who have a top-majority. Why not just say "When 2 or more candidates 
have top-majority, your ballot gives a vote to its middle-rated 
candidates only if 1 or more of your
middle-rated candidates, but none of your top-rated candidates has a top-majority."

That
 would make more sense. But, since it's an Approval election, where the 
count rule makes no mention of majority, how much sense does it make for
some particular voter to place her approvals based on majority, which isn't mentioned in the Approval count rule?

An alternative to top-majority occurred to me, for a 3-slot voting option in an Approval election.

I'd
 like to clarify that I'm speaking of something to replace the "MTA..." 
part of "MTAOC", replacing the top-majority approach. This has no effect
 on the
conditionality options, which are offerable with any of these 3-slot voting-options, or with plain Approval.

My
 alternative to top-majority, for 3-slot voting-options in an Approval 
election, might sound naive, as it would be if I claimed too much for 
it:

How about saying that your ballot gives a vote to its middle-rated candidates only if none of your top-rated candidates win?

First
 there's a count of everyone's top-ratings. Every ballot that doesn't 
top-rate the winner of that count gives a vote to its middle-rated
candidates,
 and there is a 2nd count. If that doesn't deny top-rated winners to any
 middle-rating-witholding ballots, the count is over. But if it does, 
then there is a
3rd count in which every ballot gives votes to all of
 its top-rated and middle-rated candidates, and that count's winner wins
 the election.

Of course that could, and often would, just amount 
to an Approval count. Say your favorite wins the 1st count. Everyone who
 doesn't top-rate 
that winner gives votes to their middle-rated 
candidates, automatically. Say that changes the winner, and you no 
longer have a top-rated
candidate who wins. Your ballot then gives 
votes to all of its top candidates and middle candidates, as does every 
ballot that no longer top-rates
the current winner. So it ends up as an Approval count.

But
 maybe, in the 2nd count, you still have a top-rated candidate who wins.
 Then your ballot doesn't give a vote to its middle-rated candidates.
Of course you wouldn't want it to. That possibility justifies this count suggestion.

I'm not saying that's a _big_ benefit, but it does seem desirable.

Whereas
 my previous 3-slot voting-options were top-majority voting options, 
I'll call this alternative one, described above, "Middle-When-Necessary".

It sounds a little like conditionality by top-count, 
but, for one thing, those two things act differently. For another thing,
 they're completely different _kinds_ of
things: Conditionality by 
top-count is one of two kinds of conditionality (the other being 
conditionality by mutuality--for which I've described several
implementations).
 Those conditionality options can and should be available to voters 
choosing the Middle-When-Necessary 3-slot option, just as 
it would be when choosing the majority-top 3-slot options that I formerly proposed, or for approvals on an Approval ballot too.

I'm
 not saying that Middle-When-Necessary brings a big improvement over 
just voting an Approval ballot with maybe some conditional approvals. 
But it seems to me that it a voter could sometimes thereby get a better result, and that it allows better expressivity.

By
 the way, I should probably clarify that, regardless of whether the 
voter is using, or not using, the 3-slot voting option, conditionality 
by mutuality could be
offered and implemented as shown in the MTAOC 
pseudocode;or by MMT; or by GMAT. Three different ways that the 
conditionality option could be offered.
Any one of those 3 would work
 as a conditionality option in an Approval election, and could be 
applied by the voter whether or not s/he is using the
3-slot option (for which I now propose the modest Middle-When-Necessary).

The MTAOC pseudocode kind of conditionality, and MMT, and GMAT are mutually incompatible as Approval options, but any one of them
would work well as a conditionality option in an Approval election.

It
 only recently occurred to me that Middle-When-Necessary could be 
practical and useful, and so of course it could have a problem, 
drawback,
or shortcoming that I don't know of yet. For instance, I don't know if it could spoil FBC compliance. So, at this point, it's only a tentative
possibility-suggestion, and not a proposal.

Mike Ossipoff
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20120202/e1e40024/attachment-0003.htm>