[EM] Methods for various conditions, electorates, and goals

[MIKE OSSIPOFF]

I said yesterday that my preference for SSD over Approval and RV is purely 
an individual subective preference. But, actually, for particular 
conditions, electorates and goals, one can choose objectively.

I've mentioned some appeals of Approval. Its counting of emphatic 
preferences, and its social optimizations, and all the ways of voting. These 
things have irresistable appeal, even under conditions where Approval isn't 
optimal.

But I want to talk about what's optimal.

Because I consider the public elections to be an acceptable/unacceptable 
(a/ua) situation, then, as an individual voter, Approval is the most 
convenient method for me. But one shouldn't choose a voting system based on 
what is convenient for oneself. It's necessary to go by what works for one's 
overall faction, at least.

The conditions we have in our public political elections are: We have 
greater-evils, candidates who are so undesirable that many people would do 
whatever voting it takes to try to keep them from winning.

Our electorate contains many people who perceive an acceptable/unacceptable
(a/ua) situation, but most of whom are very mistaken about what is 
acceptable. Or, if that's too subjective, then let me just say that there's 
serious disagreement about what is acceptable.

Based on those things, it would be difficult to devise a more unfavorable 
voting situation, or one that is more demanding on the voting system.

The goal that governs our choice of voting system? That probably depends on 
whom you ask.

I suggest that the important problem to be solved is favorite-burial, and, 
in general, the failure to express one's preferences. I've told why. When 
people aren't expressing their preferences, democracy means little. When 
people are burying their favorite, democracy becomes a sad joke.

So, given these conditions, electorate, and goal, what is the best voting 
system?

Consider Approval. It meets FBC, and so it gets rid of the worst problem, 
the favorite-burial problem. Approval requires someone to vote a needed 
compromse equal to their favorite, and a lot of people don't like that. I 
too would rather not have that, because it's suboptimal, by the goal that I 
expressed.

But if progressives vote for Kerry and Nader, and Nader outpolls the 
Republican, then they know they don't need to vote for Kerry next time...or 
do they? I hope so. If so, then, as we discussed before, Approval quickly 
homes in on the CW, the voter-median candidate. So, if the giveaway 
progressives can cease giving-away when Nader outpolls the Republican, then 
Approval soon, but not right away, works fine.

But if Approval only overcomes equal-voting giveaway after a few elections 
(or if progressives won't stop voting for the Democrat no matter what), then 
Approval certainly is not optimal with our conditions and electorate.

I advocate Approval and RV as the best proposal, because of their 
proposability, and because they meet the most basic requirement, FBC, and 
should eventually converge to the CW. But they aren't optmail under these 
conditions, with this electorate, and with the above-stated goal.

Yesterday I told why I claim that MDDA is optimal for these conditions, 
electorate, and goal.

What if progressives became better judges of what is acceptable,  or, to put 
it more objectively, what if there were agreement, within factions, about 
what is acceptable?  Or what if the lesser-evils were being increasingly 
voted out of the political system?

Then at least Approval's equal-voting strategy would be used correctly. 
Approval would be much better than than it is now. So maybe the first 
difference would be that Approval would become a better choice. Approval or 
RV. Approval's worst problem now is the combination of the need for equal 
ranking strategy, and voters who would use that strategy contrary to  their 
own best interest. So, with better voters &/or better candidates, Approval 
would be fine.

FBC is the really important guarantee, but as these improvements in 
conditions &/or electorate continued, maybe FBC, too, wouldn't be necessary. 
Then SSD becomes a good choice, because of its deluxe array of 
majorilty-enforcement criteria compliances.

How would SSD be better than Approval then? Each has, potentially, a 
strategy need that it doesn't get rid of. And the one that SSD doesn't get 
rid of is worse than the one that Approval doesn't get rid of. But Approval 
doesn't even come close to getting rid of need for equal-voting. It merely 
doesn't any longer have a problem with it because (we're assuming), voters 
at least know who does and doesn't deserve the equal-voting.

So, if the lesser-evil favorite-burial problem has really stopped being  a 
problem, then we can direct our efforts to really getting rid of all 
strategy need. With SSD. As I was saying before, MDDA, with its SFC 
compliance, enforces majorities very well. But why not have even better 
majority protection, with SFC, GSFC, MMC, Condorcet's Criterion, and the 
Smith Criterion. So I claim that, when voters &I/or electorate have improved 
as much as I'm talking about now, SSD becomes the ideal best method.

Is that the end of improvement in conditions, voters, and goals? No. Voters 
are still trying to optimize the outcome for themselves or their faction, 
and the goal is still to minimize their strategy need, to ensure sincere 
expression of their preferences, when they want to optimize their outcome.

But what if voters weren't trying to optimize their outcome? What if, 
instead, they only wanted to sincerely express their preferences, to do 
their part to maximize social utility (SU)?

What I mean by "sincere", in this context is different from the sincere 
voting definition that I use with my criteria. Here, I just mean that people 
are expressing their preferences truly, without strategic intent.

Approval makes another comeback. Now RV is the ideal best voting system, and 
Approval too, because Approval is a rough RV.

What do I mean by sincere voting in Approval? What does it mean for people 
to truly express their preferences without strategic intent, in Approval? 
Well obviously if you perceive some candidates desirable and some 
undesirable, then your sincere Approval voting choice has already been made.

But says there's more of a gradation. Then vote for (only) all of the 
above-mean candidates.

Yes, that's the zero-info strategy for Approval. But it's also the sincere 
vote when there's a gradation and no obviously desirable and undesirable 
sets.

When you vote above-mean, you're maximizing the sum of the 
preference-strengths that you're voting.

Approval of course has no way of knowing exactly how you rate the 
candidates, and all it has to go on is which candidates have a positive sum 
of preference-strengths with the other candidates. They're the above-mean 
candidates.

If your preferences are Nader>Kerry>Bush, and your Nader-Kerry utility 
difference is greater than your Kerry-Bush utilty difference, then your 
negative preference for Kerry over Nader is greater than your positive 
prefrence for Kerry over Bush, and so, for you, Kerry's sum of 
preference-strengths is negative. Don't vote for him.

I don't know if society will ever reach the favorable state that I'm talking 
about now. But it's still of interest, because not all voting is public 
political voting. Many organizations and committees consist of people who 
are interested in voting sincerely, maximizing SU, rather than optimizing 
the results for themselves. So, in those organizations, SU maximization by 
Range Voting or Approval is the best voting system.

I used to say that BeatpathWinner is the best for organizations and 
committees. And it is, if individual outcome-optimization is what is 
important to people. But often that isn't the case, and often Range-Voting 
is the best for an organization or committee. With Approval, the rough range 
voting, a very good solution if convenience requires it.

If an organization has any significant percentage of strategists, then it's 
better off with BeatpathWinner. What if nearly everyone is sincere, and 
there's just one or two strategists? Then of course it could be a toss-up. 
But if re-voting is allowed, then the group could vote, and then, if it is 
obvious that the strategist has strategized, affecting the outcome, then the 
others could re-vote, strategizing so as to get the result that sincere 
voting probably would have given.

I recently voted in an organizational vote, by RV. For some alternatives, 
you don't have a strong preference between them. With Approval or 
BeatpathWinner, you'd have to vote zero or full preference between them. 
With RV you can vote a slight preference between them. Sometimes you want to 
vote one over the other, be counted as preferring one to the other, but it 
isn't a strong preference. RV lets you express that slight preference. On 
that RV ballot, I didn't use the ballot's entire available range, because my 
preferences weren't particularly strong, and so if others had stronger 
preferences, they would rightly have more influence.

Mike Ossipoff

_________________________________________________________________
Express yourself instantly with MSN Messenger! Download today - it's FREE! 
http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/