To Blake re: strategy

[MIKE OSSIPOFF]

Blake said:

I wonder if anyone can find a newspaper or magazine article or editorial
that hints at some awareness of strategy in IRV. I doubt it, although
I've been wrong before.

I reply:

Since neither you nor I reside in Australia, that's a pointless thing
for us to speculate about. All 3 of the Australians that I've gotten
information from (2 by e-mail, one in person), none of whom know
eachother, told me that it's common for preferrers of small parties
to insincerely vote one of the big-2 parties' candidates in 1st place
, to avoid "wasting their vote". One of those Australian voters with
whom I spoke had just voted in that way in the most recent election.
But they all said that such voting is common in Australian IRV elections.

I've made the effort to get some information about IRV strategy in
Australia. I realize that it's easier for you to question that than
to get information yourself.

Blake quotes me:

>You said that you hope that FBC won't continue to be
>used, presumabley because you prefer the unattainable Strong FBC.

Blake replies:

I don't understand why you're doing this presuming. I thought I was
fairly clear. Voters don't seem to like voting candidates above their
favourites, but nor do they like voting candidates equal to their
favourites.

I reply:

As I presumed, your Strong FBC is more important to you, which
makes you dissatisfied with FBC.

Blake, we can't always have everything that we'd like.

You said that voters don't like voting candidates equal to their
favorites, but I don't know of evidence from ordinary voters on that.
All we know is that progressives, and many Libertarians, are very
willing to vote a lesser-evil over their favorite. Presumably those
people would be considerably more willing to vote that lesser-evil
equal to their favorite. Every Nader-peferrer who voted Gore over
Nader in 2000 would have voted Gore equal to Nader in Approval--
unless previous Approval results, or better candidate coverage
resulting from Approval, had shown that Nader was winnable.

What you meant (I presume) was that some reform advocates express a
dislike for voting someone equal to their favorite. The IRVies, in
particular.

Yes, because the IRVies don't understand that IRV will often give
voters a strategic need to vote someone over their favorite.

Methods that pass FBC have none of the former, but a lot of
the latter. On average, in the proposed FBC methods, I should rate half
the candidates as equal to my favourite, unless the candidates are
skewed bad or good. So, FBC is at best a mixed blessing.

I reply:

It isn't entirely clear what that last sentence means. Which is
better, passing FBC or not passing FBC? Say there's a method that
meets your Strong FBC. Do you think it won't pass FBC? Passing FBC
doesn't mean not passing Strong FBC. Methods that fail Strong FBC
don't fail Strong FBC because they fail FBC. They fail Strong FBC
because Strong FBC is unmeetable by any nonprobabilistic method that
I've heard of. If you know of a multiballoting nonprobabilistsic
method that will never  give strategic incentive to vote someone
equal to or over their favorite, I hope that you'll be kind enough
to post it.

Typically, as you said, you'll be voting for about half of the
candidates in Approval, or giving them maximum points in strategically-
voted CR. But you'll typically still be voting about half of your
sincere pairwise preferences.

I understand your dislike for voting someone equal to your favorite,
which is why I asked you if you know of a proposable method, or any
nonprobabilistic method, that will never give strategic incentive
to do that.

You see, what you're asking for is _complete_ freedom from any
strategy need. I doubt that that's attainable in any nonprobabilistic
method. So give your quest for complete strategy-freedom, and
settle for what's attainable. What's attainable?

Some methods never give incentive to vote someone over your favorite
(FBC).

Some methods that don't meet FBC still meet WDSC, which speaks of
a majority defeating a disliked candidate without reversing any
preference.

Some methods, while not meeting FBC, meet SFC. SFC speaks of conditions
under which voters don't need any strategy, not even truncation,
to make a disliked "greater evil" lose. GSFC generalizes SFC to
situations where there's no sincere CW.

Some methods meet SDSC, which speaks of a majority defeating a
disliked candidate without reversing a sincere preference or
voting a less-liked candidate equal to a more-liked one. In particular,
that includes not having to vote a less-liked candidate equal to
their favorite.

So even if Strong FBC is important to you, I suggest that you
consider the methods that meet related criteria that are about
desiderata that _are_ attainable: FBC, SARC, WDSC, SDSC, SFC, GSFC.

Of course Ranked-Pairs(margins) fails every one of those criteria,
in addition to Participation, Consistency, IIAC, Regularity, &
Heritage.

I presume that you mean that you prefer rank methods like IRV &
Ranked-Pairs(margins) because, though they fail FBC, the voter
_sometimes_ won't regret not voting a lesser-evil equal to his
favorite.

I don't know about you, but guarantees that contain the word
"sometimes" or "maybe" don't sound at all reassuring. Of course
for one thing, the voter might not be able to determine whether
a particular election is his lucky day when your method won't
require him to vote someone equal to his favorite (or over his
favorite, in the case of Ranked-Pairs(margins) ).

Mike Ossipoff



_________________________________________________________________
Join the world’s largest e-mail service with MSN Hotmail. 
http://www.hotmail.com