[EM] Schulze definition (was: information content, game theory, cooperation)

[Juho Laatu]

--- On Mon, 8/6/09, Raph Frank <raphfrk at gmail.com> wrote:

> On Sun, Jun 7, 2009 at 11:52 PM, Juho
> Laatu<juho4880 at yahoo.co.uk>
> wrote:
> > My thinking was that if the question on the
> > referendum excludes IRV, then the final outcome
> > is anyway likely to be Schulze (and the
> > unlikely event of choosing some other one of
> > the good Condorcet methods would not be a big
> > problem).
> 
> But they could pick the bottom 2 runoff version of IRV, if
> all you
> want is Condorcet compliance.
> 
> Some possibilities
> 
> "elect the condorcet winner if 1 exists, or the candidate
> with the
> most first choices otherwise."
> 
> "elect the condorcet winner if 1 exists or the candidate
> chosen by the
> outgoing PM otherwise".
> 
> It depends on how "evil" the legislators are.
> 

Yes. Bottom 2 version of IRV is not one
of the best Condorcet methods because of
the rather random nature of the sequential
elimination, but it is Condorcet compatible
at least.

Since one can not describe the full method
in the referendum question one has to take
some of these risks in any case. One could
try to list all the key characteristics of
the Schulze method but still the legislators
could decide to take into use ballots that
have only two slots in them (if you forgot
to include the requirement of having more
slots in the referendum question).

For these reasons and to make the voters
understand the question and to avoid giving
too much space for general complexity
arguments it may be wise to write the
referendum question without all the details
that would tie it exactly to the Schulze
method. Using e.g. River or Ranked Pairs
would probably also not be a catastrophe.

So, I tend to think that the best approach
would be to use some common language and
make the question such that it gives some
rough understanding to the voters and at
the same time eliminates the worst pitfalls.
The risks include e.g. 1) picking some bad
method due to not understanding what is good,
2) use of complexity arguments against the
Schulze method, 3) incumbents intentionally
picking a method that favours them (could be
e.g. IRV).

I agree with Árpád Magosányi in that one
should pay lots of attention on how to
formulate the question. I'd however keep
most of the complex criteria and requirements
out since that gives too much space for
speculation and complexity arguments. And as
we know one can spend lots of time in arguing
about the benefits and problems of most of
the criteria (there are arguments for and
against all of them, and all methods have
some problems that some others do not have,
later-no-harm can be used against the
Condorcet methods, do we want winning votes
or margins for Schulze etc.).

I.e. keep it simple and close to what people
really understand. If one wants a definite
binding to the Schulze method, then one can
mention its name in the question (without
explaining the details).

Juho