[EM] Methods reduce to Approval?

[MIKE OSSIPOFF]

Someone said that the methods that I propose reduce to
Approval. Ok, but what nonprobabilistic method doesn’t, when there are entirely
unacceptable candidates who could win?

 

That’s one reason why experience in voting systems will make
you an Approvalist.

 

But 3-slot methods and full-ranking methods can improve on
Approval, in the sense of offering some kind of majority rule guarantees even
when a voter votes more than two rating-levels. Such guarantees can never match
the security of Approval, or of voting Approval-style,  but, in practice, they’re often secure
enough.

 

To illustrate what I’m talking about, consider MTA, and the
timid progressive (and, then, the no-Republocrats progressive):

 

The Timid progressive:

 

S/he prefers the progressives’ policies, but feels that it’s
necessary to protect the lesser-of-2-evils democrat against the Republican. 

 

In MTA, s/he can top-rate the progressive candidates,
ensuring that if at least one of them is top-rated by a majority, and no
Republocrats are, then a progressive will win. That depends only on there being
a progressive majority who can agree on top-supporting one progressive
candidate. It doesn’t depend on a mutual majority among progressives.

 

And s/he can feel safe by merely middle-rating the
Democrats. If there’s a progressive+Democrat majority, and if the Democrat
voters know to not mark (approve) Republicans, then a progressive or a Democrat
will win. Surely the Democrat-preferrers know to not approve Republicans.

 

The No-Republocrats progressive:

 

S/he can top-rate the best progressives, hir favorites, and
middle rate the other progressives.

 

Middle and top rating all the progressives ensures that if
there is a progressive majority, a progressive will win. Again, it only
requires that majority agree on one progressive candidate to approve. It
doesn’t require a mutual majority among progressives.

 

So some of the progressives are hir favorites, and the
others are hir “lesser-evils”.  The
assurance that a progressive will win, if there’s a progressive majority requires
that the lesser-evil progressives favorite-voters won’t approve any
Republocrats. But, of one of the lesser-evil progressives encourages hir voters
to approve Republocrats, I don’t want to help hir anyway, because s/he’s not
sufficiently different from the Republocrats.

 

So, for both kinds of progressives, there’s a reliable
majority rule guarantee. More so than with full-ranking methods, in fact:

 

A 3-slot method makes a clear distinction, draws a clear
line between favorites, acceptable, and not acceptable candidates.

 

And you can be confident that the Democrat voters will
recognize that distinction and not approve Republicans. And you can be
confident that a progressive candidate who is any good will, likewise, not
encourage hir voters to approve Republocrats.

 

The distinctions between rank positions in full-rank methods
are more vague.

 

The only lack in MTA is the inability to distinguish between
your favorites. Maybe you like some more than others and would like to help
some over others.  But that’s a small
thing to give up, considering what you get with a good 3-slot method.

 

With full ranking methods meeting SFC and UP, you have two
different kinds of protection, for each rank position, against lower-ranked
candidates. Say you’re ranking x at a certain rank position. UP’s majority rule
guarantee protects x as long as hir voters don’t give it away by ranking
candidates you like less.  But, if that
candidate is sincere CW (SCW), and you’re part of a majority ranking hir  over those you rank lower, and order-reversal
doesn’t happen on a scale sufficient to affect the outcome, then less liked
candidates can’t win. 

 

So, for x to not be protected would require two things: x’s
voters giving it away to candidates you rank lower; and, if x is SCW, offensive
order-reversal against x.

 

That’s how MDD, ABucklin gives its own kind of
fairly-reliable protection at each rank position, while still letting you
distinguish among your favorites instead of equal-top-ranking them all.