[EM] Juho: Methods not subject to ridiculous bad-examples

[Michael Ossipoff]

Someone said:

> Yes, but it also helps to advocate a method that opponents can't  easily ridicule
> with bad examples.

Which methods don't have any such "bad examples" that can be used for
negative marketing? ;-)

[endquote]

Approval and Score.

Not only to they meet FBC and Later-No-Help, but they meet all of
consistency criteria that I know of. Among those that they meet are:

Participation, Consistency, Mono-Add-Top, and IIAC.

The traditional Condorcet methods being proposed fail all four of
those consistency criteria. Majority-Judgment fails Participation,
Consistency and Mono-Add-Top.

(And I might add that MJ's Participation-failure example is dramatic
and most embarrassing. It serves well to demonstrate MJ's
undesirability in comparison to Approval and Score.)

When a method fails a consistency criterion, it contradicts itself,
casting doubt on the validity and meaning of its results.

Plurality meets all four of the above-listed criteria. Proposing
something that fails consistency criteria that Plurality passes means
that we'd be _losing_ something, when changing from Pluralty to that
new method.


Changing from Plurality to Approval or Score, we wouldn't be losing
anything, but only gaining valuable properties and
criterion-compliances.

As for the Plurality Criterion, Mono-Add-Plump, and Monotonicity (and
probably others), they're met by Approval, Score, and the usual
traditional unimproved Condorcet methods. Maybe some of those criteria
can be met by MJ, if the right tiebreaker is used (a particularly
complicated one?).

It's very difficult to criticize a proposal that wouldn't lose
anything desirable, but would only bring new desirable properties.

Though Plurality fails FBC, a failure of FBC is an embarrassing
bad-example, when proposals for new methods are compared to eachother.
Voters should have a strategic need to abandon their favorite.

Michael Ossipoff


I