Positive Involvement; No-Show

[Mike Ositoff]

Markus posted examples to show that SC violates No Show and
Positive Participation, but how does Schulze's method do with
regard to those 2 criteria?

As I said, I don't consider those 2 criteria important. They
merely provide an opportunity for an impossibly well-informed
voter to offensively strategize. Not likely to ever be a real
problem.

But I still acknoledge that Schulze's method appears to be at
least somewhat better than SC. At least by its property of protecting
members of a clone set from shooting eachother down by subcycle
fratricide. It may be better in other ways too. Maybe Schulze
meets No Show & Positive Participation. Maybe not--I don't know.

Maybe there's a Schulze-tie solution that can provide more general
subcycle fratricide protection (not just clone sets). Again, I
don't know.

By "Schulze-tie", I refer to a situation where there's no alternative
that has a Schulze win against each of the other alternlatives,
meaning that, for every alternative X, it has a stronger beat path
to X than X has to it.

I've suggested that when there's a Schulze tie, the winner be
the alternative whose strongest beat path to it is the weakest,
which gives the same result as Condorcet(EM). But maybe there's
another solution that provides more general subcycle protection
than the clone set protection that Schulze ordinarily provides.

So yes, Schulze's method seems at least somewhat better than SC.
If a small-sample poll indicates that people accept both methods
equally well (in regards to complicatedness and whatrever other
considerations), then Schulze would probably be the better proposal.
But if they aren't equally preferred, then I claim that their merit
is so close, for all pracatical purposes, that polling, and not merit
should be the basis for choosing between those 2 methods.

If the other SC advocates agree, we can amend out recommendation
to ER to indicate that the SC advocates agree that Schulze's method
is better than SC, at least in the sense of clone protection.

Since EM has new members, as well as new methods proposals, I
wouldn't object to holding a new ER recommendation election here.
Maybe use Approval, and recommend everything that gets approval byk
more than a majority. But, considering Approval's strategy problems,
maybe it would be better to use SC, since it's our current recommendation,
and then just report the winner; and then the alternative that wins
after the winner is removed...etc.

Or maybe use Voter's Choice: Do the count by all the methods propposed.
Give each voter a Plurality vote, and assume that he/she wants to give
it to the alternative that wins by the methods that he/she has ranked
1st. 

(Of course, using Voters' Choice in a political election, voters
would have to explicitly state which method's winner they want to
give their Plurality vote to).

Mike