Before this reply, let me say that my demonstration that Approval soon arrives at the voter median, when there´s 1 issue-dimension, and probably Forest´s demonstration that Approval soon arrives at the CW, use the assumption that the voters and the candidates are the same as in the previous election. I haven´t seen Myerson´s & Weber´s demonstration of their similar claim, but it may very well make those assumptions too.
No, those assumptions won´t strictly be correct. Often they´ll be close enough. No, "close enough" isn´t a mathematical term. As I said, Approval strategy is based on best estimates, and so must be our statements about what Approval will do. No one is saying that those estimates or assumptions are reliably accurate, but they´re what one has, with Approval.
That´s something that has long been common knowledge here. Thank you for announcing your discovery of it. Now, are you finished repeating it?
Russ said:
I will read your postings whenever I please, and I will come down on you like a ton of bricks whenever I feel like it.
I reply:
You mean you´ll come down on EM like a ton of garbage.
Russ continued:
This is the Wild West, Mike. No rules
I reply:
Rob L, is that true here? You own and manage EM. Some of Russ´s postings, the ones devoted entirely to another list member, and filled with frothing, raging, raving namecalling, lead me to ask you, Rob, if there´s such a thing as something that can´t be posted on EM, and if there´s anything that a person can habitually post to be kicked off EM. Is there? And, if so, is it possible that Russ hasn´t qualified for eviction?
Russ continued:
One of these days you will realize the you f***ed with the wrong person
I reply:
But is there really any such thing as a _right_ f***-up to anger?
But that doesn´t mean that we have to go out of our way to not anger a f***-up.
F***-ups like you, pretentiously ignorant, angry cranks, aren´t unheard of on EM, and they´re understood to be a nuisance. One would hope that at some point they´d get kicked off EM. But if for some reason that won´t happen, then people like you are just a fact of life on EM.
Russ continued:
and you are in a no-win situation.
I reply:
I wasn´t aware that I was in some game that could be won or lost. I thought that I was just commenting on the postings of some crank f***-up.
Russ said:
I have little to lose here, Mike
I reply:
No doubt. But your presence is a loss for EM, when you waste time and archive-space.
Russ continued:
, because this mailing list is not my "career."
I reply:
You´ve said that EM is my "career". Not quite sure what you mean by that. But it´s a fact that I take EM more seriously than you do, since I don´t make use of it to vent grudges. And the pretentious ignorance that you post suggests that you don´t take the list´s topic as seriously as I do.
And what am I going to lose because of your postings? The list´s respect because I reply to you? Well maybe :-) But I think that I can be excused for replying to postings that are about me. And of course it´s acceptable to reply to postings about the list´s topic, even if they´re by a pretentous crank f***-up.
Russ continued:
I do respect and value the opinion of at least several persons on this list
I reply:
...but not enough to refrain from spamming them with your grudge-anger.
I´d said:
I had just finished saying: Many strategies can be related to Weber's strategy of voting for
candidates with positive strategic value, according to Weber's strategic
value formula.
Russ replied:
The Best-frontrunner strategy is not just "related" to Weber's Approval formula. It follows directly and trivially from it.
I reply:
Word it as you want. I and others have commented on EM that BF is a case Weber´s strategy of voting for candidates with positive strategic value. That´s common knowledge. But congratulations for discovering it. I also made that comment in an article that was at your website, till I withdrew permission for you to have my articles at your website. But that doesn´t mean that entirely different ways of choosing whom to vote for, based on entirely different information or estimates, must be called the same strategy, just because there´s only one optimal way of voting. Actually, in fact, different ways of votng can be optimal, with respect to their respective different kinds of estimates, when those different kinds of estimates, used with their different expectation-maximizing methods, lead to different ways of voting.
Russ continued:
If it was inconsistent with the formula for all reasonable input sets, it wouldn't be optimal.
I reply:
When the voter doesn´t have an estimate of Weber´s Pij, then consistency with Weber´s positive strategic value voting, using Weber´s formula for strategic value, can´t be directly tested.
But what you´re saying sounds as if you got it from one of my Approval strategy articles that was at your website. Í´d said that if we assume that there´s just one best way to vote, and that Weber´s method describes it, and if we make a few reasonable approximations by which Forest´s Better-Than-Expectation strategy conforms to Weber´s method, then, by the assumption that there´s only one optimal way to vote, and by those reasonable approximations, anyone who, by any means, votes to maximize their expectation must also be voting for his/her better-than-expectation candidates. Yes that´s based on assumptions. The conclusion that I then drew was that, with those assumptions and approximations, Approval then maximizes the number of voters for whom the winner is better than the expectation that they had for the election, better than their perceived value of the election. And I concluded that it´s reasonable to say that, with those assumptions and approximations, Approval maximizes the number of voters who are pleasantly suprised by the outcome.
You´ve latched onto someting that you copied from me, and are now presenting it as if you´d found something new on your own.
I´d said
But the resulting ways of choosing which candidates to vote for are different, even though all or most of them can be explained or justified in terms of Weber´s strategy method.
Russ replied:
I agree competely that one need not use Weber's formula explicitly to determine a reasonable or even "optimal" vote, depending on the definition of the word "optimal." However, I claim that a vote cannot be quantitatively optimal unless it satisfies Weber's formula for some reasonble set of inputs.
I reply:
That´s what I was saying at your website. Now it´s become your claim :-) But of course in practice, without an estimate of Weber´s Pij, there´s no way to test for whether or not a strategy conforms to Weber´s strategy.
Russ continued:
Any other "strategy" is essentially an attempt to make the formula more intuitive for special cases.
I reply:
No. Other stratgies use different estimates, estimates of information different from the Pij. They´re useful because people don´t usually want to try to estimate the Pij.
Russ continued:
That's fine, but don't fool yourself into thinking they are separate strategies unless theý´re non-optimal.
I reply:
They´re separate strategies because they´re completely different ways of choosing whom to vote for, based on entirely different estimates--estimates of information other than Weber´s Pij.
If you want to call all expectation-maximizing strategies the same because ideally they´re the same as Weber´s way of voting, if we disregard the fact that that can´t be tested without estimating the Pij, then fine, you can make up your own definitions and names. But you´re a pretentious crank when you try to then pass off a different definition as a new discovery.
Russ said:
Let me just say that you don't need to consider tie probabilities to derive Weber's formula.
I reply:
Weber´s formula for strategic value uses probabilities. There´s no way for anyone to know what you mean, or what you´re trying to say, when you speak of deriving, without considering probabiliies, a formula that´s based on probabilities. But we can´t expect to know what a confused, pretentioius crank is trying to say.
Mike Ossipoff
_________________________________________________________________
Don’t just search. Find. Check out the new MSN Search! http://search.msn.click-url.com/go/onm00200636ave/direct/01/
---- Election-methods mailing list - see http://electorama.com/em for list info
