[Election-Methods] [EM] RV comments

[Juho]

The mail was so long that I copied some parts of it here to be answered.

>> As discussed many times the strategy problems of Range may be too bad
>> to be overcome. We should keep seeking for better methods all the
>> time of course.
>
> Please describe a bad problem with Range. Specifically, and show  
> examples. Hopefully, they will be realistic.

I did but I assume you didn't accept the example. I'll use just  
slightly different words this time. The sincere opinions of Democrats  
are D=100, R=80, and for the Republicans R=90, D=80. If Range in used  
in a way that assumes all opinion values to be used extensively, not  
only min and max, then the one that votes more strategically (min and  
max only) is likely to win (assuming that the support of the parties  
is roughly 50%-50%). Individual voters are thus tempted to  
exaggerate. The votes of those voters that will not exaggerate will  
have smaller weight than the votes of the strategic voters.

One can fix this by letting the voters understand that in order to be  
fully efficient, an approval style vote is in most cases the  
strongest strategy for them. Letting some voters cast sincere opinion  
based votes e.g. for 50 years before they learn that their votes have  
had less weight than the votes of some other voters doesn't sound  
nice to me.

>>> This argument has been presented many times against Range, and I
>>> have never seen an analysis of what "too rewarding" is. Obviously,
>>> something is missing, there are assumptions being made that, for
>>> example, there should be no reward for strategic voting.
>>>
>>> Yet every method, to some degree, rewards strategic voting, and the
>>> reward can be large. What is "too rewarding?"
>>
>> If e.g. strategic voting is often possible, easy to apply, influences
>> the results and maybe elects some clearly "no-good" candidates.
>
> No. Please show an example. (In Range, the strategy is only to vote  
> Approval style. If a candidate is going to win because some people  
> vote approval style, I'd argue -- and many argue -- that this is  
> quite a reasonable winner. Not a "no-good" one unless the  
> electorate wants to elect a no-good.

Approval winners may be considered "not no-good". (Although also  
Approval has its strategic voting problems, but let's forget that for  
now.)

Range is likely to become Approval like if used in a competitive  
environment. Range may however provide worse results than Approval if  
there is a mixture of Approval like and sincere opinion like votes  
(and those votes are not evenly spread among the candidates).

An (exaggerated) example on how Range could elect a "no-good"  
candidate: 50% D=100, R1=80, R2=70,  30% R1=100, R2=90, D=70,  20%  
R1=100, R2=0, D=0. The "bad" Republican wins. In real life this is  
however not likely to happen since probably the D and R1 supporters  
will understand what's going on and will exaggerate too. Many R1  
supporters might take one step back and give more points (maybe max)  
to R2 too.

> I find it odd that some will argue against the mild alleged  
> vulnerability of Range Voting to "strategy," which merely means  
> voting range as a pure ranked method with two ranks, and then  
> swallow the much more insincere votes coming from strategic  
> vulnerabilities of Condorcet methods, perhaps claiming that the  
> latter are "rare," perhaps because it's too complicated for people  
> to figure out, so they won't do it.
>
> But with Range, these same people will claim that the sincere  
> voters are "suckers." That they are going to be taken advantage of  
> by conniving Approval voters.
>
> It's a double standard.

I don't think so. The main difference is that in (sincere opinion  
based) Range the (Approval style) strategies typically work in all  
elections and for all voters while in Condorcet the strategies work  
only in some scenarios and are often hard to implement. The  
vulnerability of Condorcet methods thus depends very much on how one  
estimates the various factors that influence the probability of  
"failure" due to strategic voting. In some calculations Condorcet  
methods can be considered fully safe while others may consider the  
threats more probable. (Note also again that Approval is also not  
without strategy problems, and plurality etc.)

> I am *not* saying that voting sincerely in zero knowledge is being  
> a sucker. But voting so, ignoring the identify of the frontrunners,  
> is, quite simply, foolish. There is *strategy* for voting Range.
>
> If people don't use it but do normalize, they will not be harmed  
> seriously.

I think you should refer to such normalization where at least one  
frontrunner gets min and one gets max (or something close to that).

(Note btw that normalization also destroys the otherwise nice  
behaviour of Range in the pizza example. If the sincere opinions of  
three pizza lovers are [A=100,B=99], [A=100,B=99] and [B=100,A=0]  
(the third voter is allergic to pizza A) the normalization changes  
the selection to pizza A.)

Juho





On Jul 31, 2007, at 2:45 , Abd ul-Rahman Lomax wrote:

> At 05:35 PM 7/30/2007, Juho wrote:
>> One possible definition of non-competitiveness is that voters
>> strongly want such an alternative to win that the society considers
>> best, not the one that they personally consider best.
>
> Here is the problem. In a healthy society, people do want what is  
> best for the whole society, but they also want what is best for  
> themselves. With Range Voting, we suggest, they tell us what they  
> want and how much they want it.
>
> It's possible to set up Range Voting with bids, where you are  
> effectively bidding your rating for the candidate. It is sometimes  
> argued that this would be plutocracy, *but* in most societies, the  
> bulk of the wealth and discretionary spending power is *not* with  
> the wealthy, it is distributed widely. The problem is that the  
> spending power of the poor and middle income people is not  
> organized, whereas the spending power of the wealthy is available  
> for rapid allocation. This is why the wealthy are considered more  
> powerful. Now, in some places, the concentration of power and  
> wealth into the hands of the few may be different than this, but I  
> don't think this is the case in what are called the western  
> democracies, and in quite a few other places as well.
>
> The distributed wealth of even the poor is enormous in some places.  
> But organizing it, that is another story. This is why FA/DP is so  
> important. It's a method of organizing poor people! It's one that  
> won't run away with their power and use it for narrow purposes. It  
> does this by leaving with power with them, and merely advising, and  
> it does the advising in such a way as to make it trustworthy. The  
> advice does not descend like manna from the top. Rather it filters  
> down through the proxy network from trusted proxy to client, each  
> link dependent upon the *maintenance* of trust.
>
> Consider you are a middle-level proxy in such a system. Your high- 
> level proxy, someone you have known for a long time, you talk  
> frequently, suddenly starts giving you advise that doesn't make  
> sense to you. You ask about it, and the answers aren't clear. Now,  
> do you
> (1) change your proxy.
> (2) complain and see how the proxy responds
> (3) pass down different advice that was given to you
> (4) pass down no advice
> (5) pass down the advice that you don't understand, knowing that  
> your clients trust you. Of course, you may have the same problem as  
> them.
>
> The structure is a filter, an intelligent filter, with no outside  
> control.
>
>>  Practical
>> examples on elections that may be non-competitive are voting on which
>> cloud on the sky is the prettiest, or voting on which flavour of ice
>> cream is the best flavour (everyone will buy their preferred flavour
>> after the election, they are thus not forced to buy and eat the
>> winning flavour ice cream), or what movie is the best ever (I believe
>> www.imdb.com has used some Range like method).
>
> Those are examples of non-competitive "elections," really, as  
> described, they are not elections at all, they are polls.
>
> What if a community has various options for how to spend a fund  
> that became available. The various options are presented in a Range  
> election.
>
> Now, would I recommend this? No. I'd recommend a Range poll,  
> followed by standard deliberative process in a Town Meeting or the  
> like. And it would still be a recommendation, perhaps, not a  
> control, upon a Council or whatever governmental body the town has  
> with the authority to make spending decisions.
>
> I recommend Range for public elections, even though I would much  
> rather see better process even than Range, but the latter processes  
> involve more significant structural changes. The arguments about  
> Range apply to Approval as well. What if some people are sincere so  
> they approve more candidates, others are "narrow and selfish" so  
> they only approve their favorite. The same arguments can be made.
>
> But they are specious. Implementing, Approval, Range 1, will not  
> give special power to the selfish. Nor will higher resolution  
> Range. It's a theory often stated, based on a shallow analysis of  
> what happens.
>
> Usually what is done is to assume that Approval style votes are not  
> sincere. It's a silly assumption. It is up to people to decide how  
> much they want something. I've pointed out that there is a natural  
> incentive to vote sincerely, there is a value to it that is not  
> easily quantified. It is reflected, perhaps, in the fact that Range  
> Voting, with sincere votes, actually does maximize overall  
> satisfaction; how can a method maximize your satisfaction if you  
> don't provide the information?
>
> Voting approval style puts maximal force in the direction of your  
> favorite. Now, is that accurate or not? Who are we to say? Here is  
> the paradox: if you want to exaggerate, you must have some motive.  
> If you don't care all that much, why bother to exaggerate?
>
> So if you exaggerate, you must care, enough to exaggerate. Oops,  
> that's sincere, you *aren't* exaggerating. Perhaps.
>
> It is up to the voters to decide what they really want and to  
> express that. The method will take this into consideration. If you  
> put all your eggs in one candidate's basket, that's great for you  
> if the candidate wins, and you have created some edge toward that.  
> But if you fail, you could end up with a worse outcome. So, in  
> fact, you might need to Approve two candidates, if you want to be  
> safe. But, oops!, by voting for two, Approval style, you are  
> telling the system that you don't care which of these is elected.  
> If you care, you just might be better off expressing it more  
> precisely.
>
> Generally, in such situations, in fact, the sincere strategy is as  
> good as the Approval strategy -- *almost*. The difference really  
> isn't great, particularly in Range N where N is relatively high.  
> Range 2 works well if you can exactly express your preference, but  
> error in expressing preference, oops, causes error in maximizing  
> your outcome. But all this really means is that you have to be  
> careful about rounding off. You might easily be better off rounding  
> up, for example, or down, to an Approval style vote. Which could  
> mean less error.
>
> I've suggested that a preference marker be added to Range  
> elections. It's not used in determining the net scores. It is used  
> for preference analysis. If runoffs are triggered under some  
> conditions, as shown by preference analysis, such as the existence  
> of a candidate who pairwise beats the Range winner, it becomes more  
> possible to vote sincerely... or to vote Approval style with less  
> risk. Generally, it should raise everyone's expected return.
>
> This is what we are aiming for. It makes little sense to me to  
> lower everyone's expected return in order to avoid giving "selfish"  
> people some benefit.
>
> The fact is that we want people to tell us what they want and how  
> much they want it. We will use this to determine what to give them.  
> If they lie to us, what can they expect? They may benefit in this  
> particular election, but overall, they will lose; it has to be that  
> way for actions that lower social welfare, unless some vary special  
> conditions exist that guarantee that some group is always on the  
> plus side in elections.
>
>
>
>>> By limiting ourselves to "competitive elections," we are limiting
>>> ourselves, actually, to dysfunctional societies. We need to know  
>>> that.
>>
>> You are quite ambitious.
>
> In a way, yes. I'm trying to change the world. Now, Juho, I've  
> found the lever and I've found the fulcrum, and I'm pushing. What  
> happens when you push, in free space, a really massive object.
>
> It moves. Any force moves it, but perhaps only a little, the  
> acceleration is small. However, acceleration accumulates. I've been  
> pushing for some time now, and I see motion. Even though that  
> motion is slow, should I assume that pushing is useless? Fact is,  
> some other people are starting to push with me. And this is,  
> indeed, the plan.
>
> I keep explaining it and a few get it. Those few may be busy with  
> this or that, but if they get it, opportunities will arise for them  
> to push. It starts to snowball.
>
> If it is a good idea, if it does what theory indicates,  
> essentially, if I'm write.
>
> This is not like other reforms. I'm not trying to *control* the  
> world. Rather, I am trying to seed structures that will grow that  
> will enable people to control their own lives and societies, far  
> more efficiently and effectively than present structures and systems.
>
> And it's not complicated. A few people have taken the trouble to  
> investigate it, to ask the challenging questions and reflect on the  
> answers.
>
> And, following FA/DP principles, if you were a proxy for other  
> people, and a client came to you with this idea, what would you do?
>
> Would you consider it?
> Would you pass it up among your peers in the structure for wider  
> consideration.
> Would you pass it up for higher consideration?
> Would you tell your client why it was a bad idea?
> Would you listen if the client said that he had considered all  
> those points, and here were the answers he came up with?
> Or would you decide to stick with your prior opinion, just because  
> you are always right?
> Would you drop the client for pestering you for explanations of why  
> you aren't passing it on or at least discussing it with peers,  
> perhaps connecting your client with some committee?
>
> I invented FA/DP because I needed a filter for ideas like FA/DP. If  
> we had FA/DP, we would get FA/DP *or something better* in short  
> order. It really could take only a few years to have a major  
> impact. Right now, it is in a very early stage, where one or two  
> people count. Later on, it will have its own inertia, and nobody  
> would be able to stop it.
>
> Some even worry about that.... indeed, I do. The law of unintended  
> consequences can bite hard. However, FA/DP does not destroy  
> anything. FAs don't attack *anything*, at least not head one. They  
> could be said to be attacking isolation, ignorance, powerlessness,  
> etc. But not the "bad guys."
>
> If the people have knowledge and power and the ability to find  
> consensus, bad guys will not be a problem. If they try to beat it,  
> they will just break themselves. It really could be quite hard to  
> stop, unless you nip it in the bud.
>
> I've had paranoid dreams about it. But not the slightest sign of  
> any opposition, really. FA/DP has major implications for places  
> like China. But it will be up to the Chinese, not me. And what  
> would it look like initially? Groups of civic minded people  
> cooperating to help fulfill broadly accepted public policy that is  
> being blocked, perhaps, by institutional inertia and corruption. It  
> won't attack the corrupt, at least the FA won't.
>
> But communication exposes corruption and makes it very difficult to  
> continue.
>
> In China, just so you understand better what I'm saying,  
> environmental protection is official policy. Which is largely  
> ignored because local officials are corrupt. Environmental groups  
> in China have been reasonably successful because they are promoting  
> official Communist Party policy and they are careful not to  
> challenge the Party.
>
> They are merely helping out.
>
> There is a saying, if you want to shoot the King, don't miss. In  
> Tiananmen Square, the hotheads wanted to shoot the King. There were  
> other leaders who were really sincere and who had a more  
> cooperative approach. But the students did not have an  
> organizational structure that could handle the opportunity that  
> Tiananmen Square represented. The moderates could not control the  
> hotheads, who shouted louder and provoked more strongly. The  
> government had two choices, essentially: surrender, resign in  
> disgrace, and quite possibly face severe punishment, or bring in  
> troops from outside, that did not understand the local dialect, and  
> suppress the students. Which one would you have chosen, in their  
> shoes?
>
> Chaos is very dangerous. Removing a tyrant from power is a very  
> delicate thing, it is quite easy to have someone else step into the  
> power vacuum who is worse. If you want to shoot the King, don't  
> miss. And don't do it unless you are fully prepared for what will  
> ensue. It's usually a mess, millions can die. In other words, don't  
> miss *and* don't remove the existing power center.
>
> There will be times for the people to take matters into their own  
> hands and shift power, and it can be done nonviolently, for the  
> most part. This whole thing about competitive elections vs.  
> cooperative elections is very important.
>
> And election methods, I suggest, should work well in both  
> environments. This *requires* collecting preference strength  
> information. If you don't you are flying blind.
>
>>  Making the societies non-competitive is a
>> huge task. The market economy is for example currently strongly based
>> on competition (harnessing competition to provide good results for  
>> all).
>
> That's actually nonsense. Competition is an issue and it keeps some  
> honest that might otherwise not be, but most business runs far more  
> on cooperation than on competition. Competition functions, in our  
> economy, to bound the transactions, but salespeople who are  
> cutthroat, we want to kill the competition, are generally not as  
> good as those whose goal is to serve the customer as well as  
> possible, and they are not even thinking about the competition,  
> most of the time.
>
> Read the stuff on sales, the training materials for salesmen. Some  
> of the best writing in psychology and sociology is there.
>
> There is an old joke about politicians. The most valuable trait is  
> sincerity, and if you can fake that, you've got it made!
>
> It's true, actually. However, faking it, deeply and for a long  
> time, is ... difficult. What can happen is that, acting out  
> sincerity, you become sincere. Not always.
>
>> And we don't need to go any further than to this mailing list to see
>> strong competitive attitudes (on e.g. which method is best).
>
> You've got to think of the sample. Now, consider this, raised by  
> your comment. There are some here who are really interested in the  
> truth. They would rather shrivel up and blow away than lie or  
> deceive to promote the methods they favor. If they find that an  
> argument of their own is specious, they stop using it, and they may  
> even admit that it was wrong. If someone else comes up with an idea  
> that is better than their own, they will change their views. You  
> take a person like this who's been doing it for a while, it can  
> look like they have strong, even inflexible opinions. But that  
> could be an error, it could be depth instead. They really have  
> considered, long ago, the arguments you are coming up with, and  
> rejected them after much thought.
>
> And sometimes you will come up with something that they didn't  
> think of. A new argument. If the "expert" I'm describing hasn't  
> become attached, he will recognize this and will entertain this  
> argument anew.
>
> And there are others, you can rub their noses in their repeated  
> errors, and they will continue to repeat them, never budging an inch.
>
> I wrote in one of these lists that some people would rather be  
> right than be happy. In other words, some people won't allow  
> themselves to be "wrong," ever. Of course, this means that they get  
> stuck in what is actually wrong.... and this will, almost  
> certainly, make them unhappy.
>
>
>
>> I think Warren Schudy already pointed out that there is a risk that
>> changing the voting method with good intentions may in some cases
>> lead to worse results, not better. The equation is of course complex
>> (and changes in the spirit of the society are often (but not
>> necessarily) slow).
>
> Of course, what we are proposing is the tiniest of changes, the  
> tiniest change of any of the proposals here. Not some world-shaking  
> complicated reform, requiring large investment, and, associated  
> with it, large risk.
>
> Simply Count All the Votes. Simply do what we should have been  
> doing all along, but did not do because of defective analysis,  
> which wasn't realized because the analysis and its assumptions were  
> never made explicit.
>
> If you read Robert's Rules on voting procedure, it says that, in  
> the standard method (which is Plurality), the clerk should discard  
> ballots with more than one vote on them because it cannot be  
> discerned which one was intended.
>
> There is an assumption that multiple votes are an error.
>
> Why is it an error? Because it is against the rules. Why is it  
> against the rule? Because it confuses the clerk.
>
> In face-to-face, show of hands elections, it has never been  
> prohibited to raise your hands for more than one candidate, and I'm  
> sure people do it from time to time. I've never heard it  
> challenged. Translating this to secret ballot, an old error was  
> made, an assumption that people would only vote for their favorite.  
> It wasn't discussed, I think, or at least not deeply, and  
> alternatives were not considered.
>
> And the fact is that in close, cooperative societies, Plurality  
> works quite well! It is competition and scale that make Approval  
> and Range more important. They simulate deliberative process.
>
>> As discussed many times the strategy problems of Range may be too bad
>> to be overcome. We should keep seeking for better methods all the
>> time of course.
>
> Please describe a bad problem with Range. Specifically, and show  
> examples. Hopefully, they will be realistic.
>
> The simulations show that the utility maximization of Range does  
> not seriously damage the sincere voters. So what is "too bad." I  
> asked that question before. No answer. But maybe it is below, I  
> haven't read the rest.
>
>>> "Too" rewarding is a quantitative judgement. What is "too  
>>> rewarding"?
>>
>> The problem of being too rewarding may become obvious to people e.g.
>> if after ten elections where Democrats have voted strategically
>> (D=100, R=0) and Republicans have voted sincerely (R=90, D=80) (I
>> assume that voters were requested to mark their utility values in the
>> ballots) Democrats have won every election despite of receiving only
>> a clear minority of votes.
>
> The scenario is preposterous. Republicans or Democrats have no  
> monopoly on sincerity.
>
> And your assumption that voters will be requested to mark utility  
> values on the ballot is one contrary to my own recommendations.  
> Rather, the ballot instructions I favor focus on what the method  
> does, how it works, and do not presume to tell the voter how to vote.
>
> "For each candidate, mark a rating. That rating will be considered  
> a number of votes for that candidate, added together with ratings  
> from all other voters, and the candidate with the most votes wins."
>
> And I would avoid, at least at first, any "Quorum Rule"  
> complications. Sum of Votes, quite traditional, actually, and a  
> continuation of Plurality and a continuation of Approval. Nobody  
> has proposed using average approval in Approval. Well, that may be  
> wrong. There are varieties of Approval where there is a Yes/No vote.
>
> However, there is strong tradition on how to handle that, existing  
> practice. Conflicting referenda, if more than one pass, are  
> resolved by implementing the one with the most Yes votes. Not the  
> highest average vote.
>
> And I'm not sure even as much as I wrote would be on the ballot.  
> But I do think it is a good idea, at least at first. People should  
> know what to expect will be done with their vote.
>
> Then, the voter is an action, not a "sincere response to a  
> question," which votes presently are not, nor should they become.  
> People should decide for themselves what their votes mean, but they  
> should know how the system will interpret the vote.
>
>>> This argument has been presented many times against Range, and I
>>> have never seen an analysis of what "too rewarding" is. Obviously,
>>> something is missing, there are assumptions being made that, for
>>> example, there should be no reward for strategic voting.
>>>
>>> Yet every method, to some degree, rewards strategic voting, and the
>>> reward can be large. What is "too rewarding?"
>>
>> If e.g. strategic voting is often possible, easy to apply, influences
>> the results and maybe elects some clearly "no-good" candidates.
>
> No. Please show an example. (In Range, the strategy is only to vote  
> Approval style. If a candidate is going to win because some people  
> vote approval style, I'd argue -- and many argue -- that this is  
> quite a reasonable winner. Not a "no-good" one unless the  
> electorate wants to elect a no-good.
>
>
>> In Range there are many possible ways to use the method. I think it
>> should always be made clear if one asks the voters to mark their
>> sincere utilities, normalized utilities or "fully extremized"
>> opinions (with option to cast also weak votes as needed). The Range
>> problems typically emerge when the voters use the scale in different
>> ways (strategically or with "more sincere intentions"). For me the
>> most sincere Range is one where I very seldom use the min and max
>> values (since I think a typical politician is not the worst choice
>> for the job nor the best).
>
> No examples have been shown yet. Yes, it's true, if people vote  
> their sincere ratings, *not* normalizing, the best outcomes are  
> shown by Range. However, there is a problem. If there is no  
> normalization, of any kind, then there is no commensurability.  
> There is a way, I think, to normalize votes with what I call the  
> first normalization, but I won't get into it here.
>
> The short answer is that we expect people, where there is some  
> reasonable choice in the election, normalize their vote to max  
> their favorite and min their most dislike, *among the  
> frontrunners.* Others may not do this, consider frontrunners, but  
> they will generally still normalize. And these voters are helping  
> to make the system work better, those are closer to commensurable  
> sincere votes.
>
> It's a voting method, not a poll. It is a way for people to express  
> power with some delicacy, instead of all push this way for these  
> and all push this way for those, we can all push this way for some  
> set and all push that way for another, and everything in between.
>
> If you were a control systems engineer, you'd recognize this as  
> obviously safer.
>
>
>>> It is a very serious error to term Approval-style voting in Range
>>> as "insincere." There is no basis for it other than a convention,
>>> and such conventions are dangerous, where they create special
>>> terminology, accepted in a specialized field, with implications
>>> quite different from general usage. We get to use big words in
>>> special fields, and we can even coin terms.
>>
>> Depends on how the voters were expected to vote.
>
> Sure. I expect them, most of them, to exercise their vote to  
> maximize their own personal satisfaction.
>
> And that is what we want them to do. The question of intermediate  
> votes is complex; simulations and study, so far, show some small  
> advantage in some situations to Approval style voting, but that  
> presumes knowledge of the probabilities, and there are complexities  
> of the interation between probabilities and sincere Range ratings,  
> which are then used to determine the optimal Approval Vote.
>
> It's really less complicated to simply normalize; for the  
> frontrunners, put your favorite in max, the worst in min, and then  
> decide where to put the rest, and not sweat about it a lot. Really  
> don't like a third party candidate? No problem. Even if you like  
> this one better than the frontrunner, the bad one, it's perfectly  
> okay to rate him min.
>
> It is not a poll, it is an action with consequences. The meaning of  
> the action is in the consequences, for those who care. As they say,  
> the road to hell is paved with good intentions -- which means  
> intentions that are supposedly good but aren't, and which are  
> implemented without regard to consequences because they are "good  
> intentions," not paying attention to all the messages that say  
> "Don't do that!"
>
> If you've got an election with a good guy, a bad guy, and a *really  
> bad* guy, what you should do depends on your assessment of what is  
> realistic. If the good guy can't win, in your judgement, your vote  
> is clear. Max the bad guy. But if it is close, then min the bad guy.
>
> Intermediate? Depends on the ratings and probabilities. They both  
> count. Zero knowledge, a very good strategy is to vote sincerely,  
> it equals Approval, actually, if we have Range of sufficient  
> resolution.
>
> But when you have probabilities, it shifts. You then know what  
> pairwise elections are important. And in pairwise elections, you  
> want to cast full strength votes, generally. Unless you really  
> don't care.
>
> When there is a pairwise election, when the other candidates are  
> moot, it's just silly to vote intermediate ratings for them. That  
> there are other candidates on the ballot does not affect how your  
> vote is going to change the real world, if it has any effect at all.
>
> And we should always act as if our actions will affect the result.  
> That was why I did the study of Range 2, basing utilities on those  
> restricted votes, the ones that would affect the outcome.
>
>>> Range *is* "Approval with option to cast weakened votes," it does
>>> not become it. And it turns out that those who cast weakened votes,
>>> even if only a few (so few, indeed, that one voter can accomplish
>>> it), *help* not only the strategic voter, but also the sincere ones.
>>
>> What I mean by "becoming" is that majority of the voters come to the
>> conclusion that they will use min and max values and forget the mid
>> values (assuming that they were originally expected/requested to use
>> also other than min and max values).
>
> Get this. I expect most voters, if we start to Count All the Votes,  
> will just cast one. Does this mean that it's useless?
>
> Well, for starters, it is actually easier than excluding overvotes
>
> But then a few percent doing something else improves the outcomes  
> in many elections.
>
> And in Range, interesting fact. The value of a vote, under the  
> assumption that the vote counts, jumps if the election goes from  
> Approval to Range 2. It appears that even a single voter can make  
> this happen, so if only a few voters use intermediate ratings, it  
> improves the outcome for everyone.
>
> I'm not sure I fully understand it, so it may be speculation, but  
> controls systems sometimes work better when a little noise is  
> added. The single votes create some "buzz," and could make the  
> response of a system better.
>
> But those are the numbers, so far. Range 2, 3 candidates, sincere  
> candidate utilities of 210 for the voter, zero knowledge, relative  
> utility of voting an exactly sincere Range vote under the  
> assumption that the vote counts, 40% over not voting. Approval  
> Vote, 40% over not voting, doesn't matter which, in large  
> elections. (Does in small.) Change the election to Approval,  
> expected satisfaction, 33% over not voting.
>
> Isn't that interesting?
>
> Now, I could easily have made mistakes. However, Warren Smith has  
> now taken my study, refined it, and published it with both our  
> names on it. There may be some mistakes he introduced, and it  
> remains possible that I led him down the rosy path with some  
> assumptions that looked good but which were false (certain  
> assumptions about probability), but that would really only impact  
> the exact results, not the method. The method I've suggested has  
> some weight.
>
> It's a calculation, not a simulation.
>
>>> Please explain it to me, why we should consider strategic votes as
>>> something to prevent. They are a medicine that voters use when they
>>> are sick, when society is sick.
>>
>> Strategic votes themselves are maybe not a problem if they are just
>> noise in the election, but if the strategists are able to change the
>> outcome of the election to something else than the society wants it
>> to be, that'd be a problem.
>
> They are not noise, they are binary data. And how do you know "what  
> society wants the outcome to be"? What is the standard?
>
> Look there is no way of doing it that makes sense except some kind  
> of Range poll, if you could get people to vote sincerely.
>
> And if you can't, then you *still* get better information. The  
> serious distortions of strategic voting are when people reverse  
> preferences, which some other methods practically require in some  
> circumstances.
>
> I find it odd that some will argue against the mild alleged  
> vulnerability of Range Voting to "strategy," which merely means  
> voting range as a pure ranked method with two ranks, and then  
> swallow the much more insincere votes coming from strategic  
> vulnerabilities of Condorcet methods, perhaps claiming that the  
> latter are "rare," perhaps because it's too complicated for people  
> to figure out, so they won't do it.
>
> But with Range, these same people will claim that the sincere  
> voters are "suckers." That they are going to be taken advantage of  
> by conniving Approval voters.
>
> It's a double standard.
>
> Voting methods should be judged, at this point, by simulations, and  
> major work should be put into making the simulations as realistic  
> as possible.
>
> And then methods should also be studied by gathering much more  
> complete ballot data from real elections. We need to know what is  
> on each ballot. And there are good reasons for making all this  
> public. Among other things, done properly, it could make election  
> fraud almost impossible.
>
>
>>> Essentially, trying to maximize my personal gain in a Range
>>> election by voting Approval style is short-sighted. If everyone
>>> does it, everyone loses, on average.
>>
>> Note that the "on average" addition is critical. If the strategists
>> will benefit, that's a good enough reason for them to vote
>> strategically.
>
> Maybe. But in Range, the "strategy" is little more than deciding  
> where to place your scale and how to expand it. It's still honest,  
> actually.
>
> I expect voters to vote in the way that makes them happiest, with  
> Range. They can vote sincerely, they vote Plurality, they can vote  
> Approval. (In the small election case, with balanced utilities,  
> there was an edge to voting 200 over 220. I have a six-voter Range  
> 2 election, complete list of possible votes, cooking. It's being  
> done to test the formulas for our exact study of many-voter elections.
>
> So the optimal strategy in small Range 2 elections is actually to  
> vote Plurality style.... even if your utilities are balanced. It  
> follows, then, that your actual Approval cutoff is above your  
> sincere rating for the candidate..... in large elections, this  
> disappears in the noise, but it's still there.
>> If Range is presented to voters as "vote min or max, or a weak
>> intermediate vote if you like" then voting in Approval style is no
>> problem. If Range is presented as "mark your sincere preferences,
>> e.g. R=90, D=80" then Approval style voting should be considered
>> strategical and those voters that voted as requested may feel
>> themselves betrayed.
>
> Only an idiot would use the word "sincere" in ballot instructions.  
> Might as well say, "Vote like the suckers that most people are."
>
> I am *not* saying that voting sincerely in zero knowledge is being  
> a sucker. But voting so, ignoring the identify of the frontrunners,  
> is, quite simply, foolish. There is *strategy* for voting Range.
>
> If people don't use it but do normalize, they will not be harmed  
> seriously.
>
> I've never seen a ballot instruction that told people how they  
> should vote, other than the rules, such as "Vote for One," which is  
> what is common.
>
> For Approval, that changes to "Vote for each candidate as you  
> choose." Or if the Nos are made explicit, which guarantees a  
> majority winner or *no* winner, by the way then "For each  
> candidate, Vote Yes or No."
>
>> Range with 100% of voters using Approval strategy gives same results
>> as Approval.
>
> Duh. However, it ain't gonna happen. Some people will vote their  
> sincere ratings, of some kind of sincerity. Intermediate ratings  
> will be used.
>
>>  Range with some voters voting in Approval style gives
>> more power to the Approval style voters. Range with one opinion group
>> voting more in Approval style than other groups favours that
>> political group.
>
> But a "political group" voting this way isn't a "political group."  
> Political groups promote agendas.
>
> Range will not favor some political group over another, unless we  
> form a "We Gotta Vote Sincerely If It Kills Us" political group.  
> Would you join? I wouldn't. I'd vote sincerely if I thought it  
> wouldn't harm me more than a little -- might be *very* little --  
> and I'd vote Approval style if I cared.
>
> For intermediate candidates who aren't going to win no matter what  
> I do, in my opinion, I will vote intermediate ratings. Why not?
>
> And this is part of the benefit of Range. We start to see the true  
> strength -- or weakness -- of third parties. Warren calls it the  
> "incubator effect," and he's written a lot of hype about it, but  
> it's real, I think.
>
>> A general comment. The main problem in these discussions seems to be
>> the problem of mixing Approval style, normalized votes and fully
>> utility based votes in Range. Different groups have different power.
>> The dynamics of the system may drive it towards Approval style
>> voting, not towards more sincere utility oriented voting.
>
> I expect if we have Range 2, the likely first implementation in  
> public elections (CR-3, votes from 0 to 2 or Minus, Zero, Plus),  
> most voters won't vote the middle at all. They will vote as they  
> voted before. And with Approval, even before that, the large  
> majority of voters will vote as they did before.
>
> But we will get better outcomes. Who what is this nonsense about  
> power?
>
> People who vote sincerely are very unlikely to see outcomes where  
> they seriously regret their vote. So where is the pressure alleged  
> toward voting Approval Style. Politicians will still try to get  
> people to think that they are 100% and everyone else is a zero. So  
> new?
>
> Frankly, I'm getting tired of the B.S., analysis that is dreamed up  
> with no connection to reality, horror scenarios that have no  
> connection with what is likely to happen.
>
> There are *real* horrors that actually happen, Warren has written  
> about quite a few. There are real dangers from some of the election  
> reform proposals, though I don't want to exaggerate them, the sky  
> will not fall if we get IRV, though we might get some bad election  
> results (it happens a surprising percentage of the time,  
> apparently, if third parties start to get strong, as supposedly IRV  
> will help happen.)
>


		
___________________________________________________________ 
The all-new Yahoo! Mail goes wherever you go - free your email address from your Internet provider. http://uk.docs.yahoo.com/nowyoucan.html