[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Conclusion of the discussion of voting and super majority concepts?

Hash: SHA1

Hi folks,

        This is a preliminary recommendation from the folks examining
 voting methods. I am not sure we can use this for the coming election
 as the primary method, given the wording of the constitution. I am
 likely to use the vote machinery already in place, and compute the
 winner by the new, recommended method as well, but barring a
 constitutional amendment, the previous voting method is still going
 to be the final word.

        Call this election the test period for the new voting methods.

ps: I am in the process of researching the algorithm, since to an
uninitiated layperson like me, terms like "weakest defeat" need
definition; anyone forwarding algorithms/psuedocode shall earn my

 Mike> "Mike" == MIKE OSSIPOFF <nkklrp@hotmail.com> writes:

 Mike> The constitution revision committee arrived at a concrete recommendation
 Mike> for the count rule, for counting the rankings and choosing a winner:

 Mike> We voted to recommend Cloneproof SSD. It's equivalent to another
 Mike> procedure that is known sometimes as BeatpathWinner and sometimes
 Mike> as Schulze's method. But we choose Cloneproof SSD as the procedure
 Mike> to recommend, the rank-count rule to recommend.

 Mike> Let me briefly state the definition of Cloneproof SSD:

 Mike> First, a definition of the Schwartz set:

 Mike> 1. An unbeaten set is a set of alternatives none of which are
 Mike> beaten by anything outside that set.
 Mike> 2. An innermost unbeaten set is an unbeaten set that doesn't contain
 Mike> a smaller unbeaten set.
 Mike> 3. The Schwartz set is the set of alternatives that are in
 Mike> innermost unbeaten sets.

 Mike> [end of definition]

 Mike> Of course, if there are no pairwise ties, then there will only
 Mike> be one innermost unbeaten set, but small committee voting can
 Mike> sometimes result in pairwise ties, and so the definition allows
 Mike> for the existence of more than one innermost unbeaten set.

 Mike> Cloneproof SSD:

 Mike> 1. Calculate the Schwartz set based only on undropped defeats.
 Mike> 2. If there are no defeats among the members of that set, then
 Mike> the count is over, and the members of that set are the winners.
 Mike> 3. Otherwise, drop the weakest defeat among the candidates of that
 Mike> set, and go to 1.

 Mike> [end of definition]

 Mike> Cloneproof SSD (CSSD) meets all the criteria that our committee
 Mike> consider important, including the Condorcet Criterion, the Smith
 Mike> Criterion, Monotonicity, Independence from Clones, and a number
 Mike> of strategy criteria that measure for the important goal of avoiding
 Mike> a lesser-of-2-evils problem.

 Mike> Most of these criteria can be found at http://www.electionmethods.org

 Mike> At the "technical evaluation" page, there's a criterion compliance
 Mike> table that compares SSD to some other commonly proposed voting
 Mike> systems.

 Mike> Independence from Clones requires that a faction not gain or
 Mike> lose advantage by running a number of identical candidates. Of
 Mike> course our committee used a more precisely worded definition of
 Mike> the Independence from Clones Criterion (ICC).

 Mike> Another criterion met by CSSD, and which we consider important is
 Mike> the Schwartz Criterion, which requires that the winner come from
 Mike> the initial Schwartz set.

 Mike> Unfortunately, we didn't make any progress on the other issues, and
 Mike> discussion slowed down & stopped when we moved on to the next
 Mike> constitution revision issue. But we definitely made an official
 Mike> decision, based on a vote, for the issue of what method should be
 Mike> used to count the ranked ballots in elections: We chose Cloneproof
 Mike> SSD.

 Mike> I should add that, though the algorithm that I've just sent tonight
 Mike> speaks of electing every member of the Schwartz set when there are
 Mike> no defeats among the members of the Schwartz set, nearly always the
 Mike> reason why there are no defeats among the members of the Schwartz
 Mike> set is because the Schwartz set has only one member.

 Mike> Only if there are pairwise-ties can there be more than one member
 Mike> of a Schwartz set that has no defeats among its members.

 Mike> Usually it's suggested that ties be solved by "Random Ballot",
 Mike> but I understand that Debian solves them by Chairman Ballot--the
 Mike> Debian Project Leader, or his ballot, solves the tie. That's
 Mike> of course another perfectly good solution, and is more convenient
 Mike> when e-mail is used, since it isn't possible for everyone to watch
 Mike> a coin-flip, or a number being drawn from a bag, etc.

 Mike> (In an e-mail poll, I once specified an elaborate pseudo-random
 Mike> procedure for choosing a "random ballot", but there wasn't a tie).

 Mike> Anyway, though we made a definite recommendation for the rank-count
 Mike> method, for voting, we didn't recommend on the
 Mike> supermajority issue. Discussion & progress just slowed & stopped when
 Mike> we got to the supermajority issue.

 Mike> But the rank-count issue isn't dependent on
 Mike> the supermajority issue, and so our rank-count recommendation could
 Mike> be implemented even though we haven't voted on the other issues.
 Mike> Let me know if there are any questions or comments.

 Mike> Mike Ossipoff

- -- 
 "Remember kids, if there's a loaded gun in the room, be sure that
 you're the one holding it" Captain Combat
Manoj Srivastava   <secretary@debian.org>  <http://www.debian.org/%7Esrivasta/>
1024R/C7261095 print CB D9 F4 12 68 07 E4 05  CC 2D 27 12 1D F5 E8 6E
1024D/BF24424C print 4966 F272 D093 B493 410B  924B 21BA DABB BF24 424C
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: Processed by: Debian GNU/Linux -> Emacs -> Gnus -> Mailcrypt


Reply to: