[EM] How much does ABucklin improve on Approval?

MIKE OSSIPOFF nkklrp at hotmail.com
Sat Mar 24 12:49:55 PDT 2012

I like ABucklin (an easier name by which to refer to ER-Bucklin, as defined in the electowiki), because, when done
right, it meets the Mutual Majority Criterion (MMC), and, thereby, even gives co-operation-defection protection under certain
conditions (lots of parties; the two parties in question don't add up to a majority).

In ABucklin, maybe there won't be a 1st-choice majority, an immediate initial majority. But, even if not, then there is obviously
a very good chance that the next stage, wherein ballots give to their next choice, will result in a majority.

If so, then, if you've ranked someone below 2nd choice, then you haven't helped hir. S/he might lose, where s/he would
have won, had you ranked hir higher. There's good strategic reason to rank important compromises at least in 1st or 2nd
place. So, ABucklin's unlimited number of rank positions might not really be a good idea to use.

ABucklin retains the unmatched simple and clear strategies of Approval. If it's an option in an Approval election,
the Approval strategies are still valid if you're voting an ordinary Approval ballot. If the method is ABucklin, then
the Approval strategies are valid for you if you vote Approval, by only using ABucklin's top rank position.

But one thing that is lost if Approval isn't used is Approval's unique optimization: No matter what kind of election it is,
and no matter whether or not it's 0-info, Approval maximizes the number of voters who are pleasantly surprised by
the outcome. The use of options for other ways of voting would lose that optimization.

And it (obviously) elects the candidate marked as approved by the most voters. That counts for a lot, if public approval
and satisfaction with the govt matters.

Here's something that can be said for ABucklin: Under the condtions outlined above, ABucklin gives co-operation/defection
protection. But, because that's only under specified conditions, that doesn't change the conclusion that I stated earlier:

Approval can't be improved upon, other than questionably and doubtfully.

(Jameson: I'm referring to a comparison of Approval with other ballots-only methods. Of course delegation is one
of several options that can be offered in an Approval election. Though SODA's FBC violation isn't the worst kind, it still
represents a departure from full, ideal compliance with FBC. Tolerable, but not complete. I have no objection to adding
the delegation option to an Approval election. I consider it a valid and maybe desirable option, as are ABucklin,
MTA and MCA. As I've said, objections would be difficult to oppose (I don't oppose the above-named options. 

But "SODA" means Approval with the delegation option, and not any other options, as I understand the definition of SODA.
The delegation option could be offered, but can you really argue that the ABucklin option shouldn't be allowed? Or MTA or


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20120324/51158d45/attachment-0003.htm>

More information about the Election-Methods mailing list