[EM] SSD vs MMPO. What's important in voting system results.

[MIKE OSSIPOFF]

First, I reaffirm that I consider MMPO to be the best rank-balloting public 
proposal. Maybe the best public voting system proposal, bar none, because 
its unique ultimate simplicity avoids the problem that there are innumerable 
rank-counts. Maybe, in a public proposal, MMPO would do better than CR or 
Approval. I'd offer MMPO first.

But, though I've been saying that, and meaning it, I didn't say that _for 
me_, aesthetically, and maybe even as a voter, I like MMPO better than SSD. 
It seems to me that SSD has a lot of beauty, and its own kind of compelling 
motivation and justification. Disregarding proposability, disregarding 
MMPO's essential guarantee for the timid voter, all else being equal or 
disregarded, I aesthetically prefer SSD.

How could I say that, though MMPO is the public proposal that I advocate, I 
might prefer SSD, for me, as a voter, when MMPO, but not SSD, meets FBC? 
Wouldn't we all benefit from FBC, even those of us who aren't 
lesser-of-2-evils voters? Depends on how the method gives FBC compliance.
It depends on whether it meets FBC by giving us a no-favorite-burial 
solution that SSD doesn't have, or whether it does it by removing a 
favorite-burial strategy opportunity that SSD has, without substituting 
another opportunity.

I haven't thoroughly looked into this, so this is just my first impression, 
but it seems to me that the latter is correct. I'm not criticizing MMPO. 
But, for someone who is not a lesser-of-2-evils voter, someone who won't 
mis-use or over-use favorite-burial strategy, I'm suggesting that SSD's lack 
of FBC compliance doesn't amount to a loss or a meaningful lack. SSD's 
favorite-burial strategy opportunity is just a disastrous mis-use-temptation 
for the lesser-of-2-evils voter, and that's why MMPO is better for the 
voting public. As I said, with MMPO, you can emphatically absolutely assure 
voters that if they rank a lesser-evil in 1st place, there's absolutely no 
reason to not rank their favorite in 1st place too. That could be what can 
reassure lesser-of-2-evils voters not to bury their favorite.

But, about wv's FBC failure, say you rank the expected CW equal to your 
favorite, but the other people sharing your favorite don't do so, or at 
least most of them don't. And the CW's voters don't truncate for deterence 
as maybe they should. So say the offensive order-reversers will win unless 
some small subset of those sharing your preferences vote the CW over your 
favorite.

What's different with MMPO? The offensive order-reversal will still succeed, 
and the only difference is that you, and the rest of your defensive-strategy 
subset, won't have that opportunity to save the CW, by favorite-burial 
strategy. So that could be called a lack rather than an advantage, for 
people who won't mis-use or over-use that favorite-burial CW-saving 
strategy.

So, for me, what I was saying in the previous paragraph means that SSD's 
favorite-burial strategy is an added option for me, which is ok if I'm not 
someone who will mis-use it.

I often refer to a situation in which, for a particular voter, the 
candidates are in an acceptable set and an unacceptable set, and the merit 
differences within those sets are negligible compared to the merit 
difference between those sets.  I should have an abbreviation for that, to 
avoid having to say the whole thing. I'll call it the 
acceptable/unacceptable situation. As I've said, I myself judge the 
candidates as either acceptable or completely unacceptable, completely 
undeserving of a vote.

With the acceptable/unacceptable situation, of course Approval or or CR is 
perfectly good, though not really deluxe, for the voter who perceives the 
candidates in that way.

In SSD, with the acceptable/unacceptable situation, and if the voter feels 
that a certain one of the acceptable candidates is likely the CW, that voter 
probably would best rank that candidate in 1st place, over the others, over 
his/her favorite. I hate to say that, but it seems obvious, since, strictly 
speaking, the wv methods and margin methods don't meet FBC.

In MMPO, there's no reason to do that. Provided that you're not a 
lesser-of-2-evils sucker, are you better off with that strategy 
availability, as it is in SSD? Maybe, but what about the fact that the 
opposition won't have it either? Does that tend to reduce the loss to you, 
maybe outright cancel it out, when you don't have that favorite-burial 
CW-saving strategy, and the opposition doesn't have it either? Well, maybe 
it could be a (very) little worse for SU.

By the way, maybe that's true of lots of lacks in a voting system, and that 
suggests that there are only two things that matter about voting system 
results: Strategy need, and social utility (SU). If failing a criterion 
doesn't cause one of those problems, then I suggest that it's a value-less 
criterion. (This paragraph is the reason for the 2nd part of this message's 
subject line). Of course you could say that CC relates to strategy, in the 
sense that you don't need it if everyone is sincere and if there's a CW; and 
that CC also relates to SU, because the CW usually maximizes SU. But, as I 
pointed out to James, if the voters are timid lesser-of-2-evils voters (they 
are), and if the method doesn't meet FBC, then you won't need to worry about 
what happens if all or nearly all voters vote sincerely.

Back to the topic of this message: But I guess that was pretty much all I 
was going to say about it. Maybe in some distant future, when better voting 
systems have been in use for a long time, and when the political system has 
become more honest, and the voters have better judgement, and no one is a 
lesser-of-2-evils sucker (maybe the worst "greater-evil"s have been voted 
out of the political system by then)--Maybe then SSD will be chosen because 
of its special aesthetic beauty, and the extra luxuries of SDSC (which MMPO 
has, with AERLO), GSFC, CC, Smith, ICC and MMC, which could be worth having 
when there isn' t a favorite-burial problem that makes FBC necessary.

In the meantime, BeatpathWinner/CSSD is a good proposal for organizations 
that won't have a favorite-burial problem. Either BeatpathWinner/CSSD or 
MMPO would be fine for such an organization. Of course the advantage of 
proposing MMPO to organizations is that it gives precedent for something 
that is really easy to propose to the public and has a timidity-proofness 
that makes it a better public proposal.

About criteria again, I suggest, as I said, that strategy-need and SU are 
the only important considerations, and that criteria whose faiure doesn't 
affect those things aren't important enough to mention. And I suggest that 
people reconsider their assumptions about criteria, and what is essential or 
important for a voting system. Maybe some criterion that you found in a 
journal isn't really so important.

Mike Ossipoff

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