[EM] Kristofer: Defection-resistance. Strong FBC.

[Michael Ossipoff]

Kristofer:

You asked:
**
* * What's the advantage of ICT over ICA?

[endquote]

ICT is defection-resistant. That's because, in dealing with a situation
where there isn't exactly one unbeaten candidate, ICT ignores co-operation.
If you defect, then ICD disregards the cooperation that you hoped to take
advantage of.

Maybe someone could propose other methods based on that principle.

You continued:

* * A tangential question: is Condorcet and defection resistance
incompatible?

[endquote]

No. CT meets both of those desiderata.

I suggested CT for some future electorate that won't be scared into
favorite burial
by Condorcet FBC failure, and who want Condorcet's Criterion. CT is
defection-resistant.

Even if voters weren't timid enough to favorite-bury in Condorcet, the
co-operation/defection
problem (the chicken dilemma) would still be there.

Of course, if everyone is completely honest, unselfish, and only wants
what's best for society,
then use Range.

You asked:

Another question: do you know of any method meeting strong FBC and
defection resistance?

[endquote]

No, but that doesn't mean much, because I haven't looked at Strong FBC. I
didn't even know what
it was until you mentioned it in a recent posting.

But it has occurred to me that the FBC that I use is weak. All lthe
better,because that says more about
the methods that fail it: There are situations where favorite-burial can
bring a result such that no result
that good can be gotten without favorite-buial.

The things that you said about that weak FBC had occurred to me too: In
principle, strictly speaking, a method
could pass it, but the only way to equal the good result of favorite-burial
might be some difficult or comlicated
thing, maybe something requiring impossible predictive informtion.

So, though weak FBC (the kind I use) is definitely useful for
distinguishing the worst degree of FBC failure, something
intermediate between that and Strong FBC would be useful too. I haven't
defined such a criterion, but I certainly should,
since I'm the one who is emphasizing FBC.

In MinMax pairwise-count methods, like MMPO and MDDTR, you can better help
Compromise against Worse if you make sure to
rank other candidates over Worst, to try to get a bigger pairwise defeat
against him.

Maybe it can be said that, in all of the mehods I've named as FBC-passing
methods, you can maximally help
Compromise against Worst merely by voting Compromise at top, and Worse
alone at bottom.

By "maximally help Compromise against Worse", I mean doing the best that
can be done to defeat Worse, where it is
assumed that is most likely to be done by electing Compromise instead.
Maybe it needs work. I'm new to the subject of stronger
FBCs.

Should I call that "Intermediate FBC"? I suppose it would need a more
rigorous definition. Or a whole different approach.

Maybe Weak, Intermediate and Strong FBC should be abbreviated WFBC, IFBC,
and SFBC.   ...with the understanding
that "FBC" means "WFBC".

This subject of Strong FBC is new to me, and so I don't know how strong an
FBC ICT meets. Certainly it meets
WFBC. I don't know if any defection-resistant method meets SFBC. It seems
certain that ICT meets IFBC, if my IFBC criterion
definition defines a useful criterion.

Weaker crirteria are more lenient. Failing a weaker critrerion makes a
stronger negative statement. Condorcet fails WFBC, the weakest FBC.

Mike Ossipoff
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20120423/1e6fbdca/attachment-0003.htm>