[EM] Naive random UWA, timeshare, etc.

[Markus Schulze]

Dear Joe,

random single-winner methods have already been promoted by
other participants of this mailing list. For example: Lucien
Saumur promoted Smith//RandomCandidate; Albert Langer promoted
RandomDictatorship; Smith//RandomDictatorship has been mentioned
from time to time as an example for a single-winner method that
meets Pareto, monotonicity, and independence from clones.

When we talk about Condorcet methods, then I have to say that
I haven't yet seen a random tiebreaker that can solve a problem
that cannot be solved by a deterministic tiebreaker. The argument
for UWA-Random is that it meets "very local IIAC". (That means:
An additional candidate who isn't in the new Nash Set cannot
change the winning probability of an already running candidate?)
However, in so far as there hasn't yet been posted an example
where SSD or Ranked Pairs violates this property, I guess that
also SSD and Ranked Pairs meet this property (without having
to violate monotonicity like UWA-Random).

Markus Schulze