[EM] Minimizing need for insincerity

[MIKE OSSIPOFF]

In a reply to Markus I spoke of minimizing need for insincerity.
When I say that I want criteria that measure how well a method
minimizes need for insincerity, I mean that I'd like to
minimize the degree of insincerity needed, and it would be good
for that protection to extend to as many voters in as many
situations as possible.

FBC doesn't in any way limit who is protected. SARC only
protects people voting undominated strategies, but that
effectively includes all voters. But FBC only protects against
need for preference reversals involving one's favorite. It
could have included last choices too, but I felt that not having
to dump one's favorite is what really counts.

The adverse results that SARC protects against are those that
can be caused by reversing preferences involving 1st or last
choice. In that way SARC can be regarded as another way of
stating FBC, in terms of adverse results instead of in terms
of the strategy itself. But I consider it useful to say it
that way too, because we have traditional adverse results
criteria, and so why not have one that doesn't make the
unrealistic assumption of sincere voting?

WDSC protects against need for _any_ reversal of preference order,
not just those involving 1st or last choice. But, to get that
better coverage of strategies, it gives up coverage of some
voters, since its protection applies only to a majority who
prefer A to B.

Likewise with SDSC. For some methods it's possible to add the
additional guarantee against having to vote equal 2 candidates
for whom you vote.

SFC & GSFC offer a much more impressive guarantee about strategies:
No strategy is needed. But it further limits the situations for
which it can make that gurarantee. As with WDSC & SDSC, they
only protect a majority who prefers a certain candidate to B.
And, additionally, the protection only exists if no one votes
an unfelt preference. We wanted criteria that show where the
best voting systems can be completely free of need for
defensive strategy.

That complete freedom from strategy is the reason for SFC's name,
and we all agree that it's a desirable goal. So it's worthwhile
to tell when that gurarantee can be made, and for whom--and for
what methods.

I feel that SFC & GSFC are the most important advantages of
Condorcet's method. That's why I consider PC to be pretty good,
as it passes SFC.

BC is useful because its compliance guarantees compliance with
SFC, GSFC, WDSC, & SDSC. So it's only necessary to determine one
criterion compliance instead of 4, and only necessary to name
one compliance when describing the methods' important compliance.

When, for certain voters in certain situations, you minimize
the degree of insincere strategy that they need to defeat some
greater evil, you're also protect them from being penalized for
not using that insincere strategy.

Unlike FBC & SARC, the majority defensive strategy criteria actually
name a greater evil whom we want to be able to defeat with
a minimum of insincere voting.

Saying that we protect voters from need for insincere voting
is the same as saying that we protect them from being punished
for not voting insincerely. And so all 7 of the defensive
strategy criteria (FBC, SARC, UUCC, SFC, GSFC, WDSC, & SDSC)
do what Markus said he wants. He said that he doesn't want voters
to be punished for sincerity.

So really, Markus has no disagreement with me.

And, to underscore that, note that Markus's favorite method
is one of the very few (three) methods that meet all 4 majority
defensive strategy criteria, without having a Monotonicity badexample.

Those methods are BeatpathWinner, Tideman(wv), & SSD.

And please note also that those 3 methods are the favorite methods
on EM. The 3 favorite methods on EM are the 3 methods that
meet all 4 majority defensive strategy criteria without having
a Monotonicity badexample. So don't let Markus tell you that
those criteria lack support or acceptance.

Mike Ossipoff






_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Share information about yourself, create your own public profile at 
http://profiles.msn.com.