[EM] as to Favorite vs Compromise vs Worse.

[Dave Ketchum]

Whatever the election method, voters are concerned with three groups  
of  candidates:
.     Favorite:  Desire electing one of these.
.     Compromise:  Not desired, but can help to avoid electing Worse.
.     Worse:  Want to avoid these getting elected.

Not covered here, seeing to better candidates getting nominated is  
also a worthy goal.

On Apr 26, 2012, at 11:21 PM, Michael Ossipoff wrote, as:
      Re: [EM] Article with some wording improvements, and ommissions  
fixed.

> Adrian:
>
> This is (what I consider) an improvement on my article. I've  
> replaced some awkward wordings, and added things that I'd left out,  
> and left out a few things too.
>
> But you should feel free to edit in any way necessary, &/or use an  
> earlier version.
>
> For instance, if I've now made it too long, then I have no objection  
> if you take out whatever necessary, to shorten it--or use a previous  
> shorter version.
>
> If it's of use to tell this, I wrote it in Word, then copied it, and  
> pasted it in this e-mail.
> Also, of course, use any title you prefer.
>
> My article:
>
> Some problems and undesirable social results of our current voting  
> system, Plurality Voting.
> Suggesting a minimal but powerful  improvement called “Approval   
> Voting”.
>
> Our current voting system, of course, is the “vote-for-1” method.   
> Also called "Plurality", or the "single mark method".
>
> In our Plurality elections, we often hear people saying that they're  
> going to vote for someone they don't really like, because he/she is  
> the "lesser-of-2-evils". Note that they're voting for someone they  
> don't like, and not voting for the people they really do like,  
> because the people they like are perceived as unwinnable.
>
Unable to vote for more than one, above voters give up on Favorite and  
vote for Compromise as the best they can do.

They fear that voting for Favorite is more apt to let Worse win.
>  What do we get when we vote for people we don't really like? We get  
> something that we don't like. Everyone complains about how all the  
> viable politicians are corrupt and bought. Does it really make sense  
> to believe corrupt and un-liked candidates to be more "viable"? How  
> viable would they be if everyone could feel free to support  
> candidates whom they actually like?
>
> We'd be voting from hope, instead of just from fear and dismal,  
> pessimistic resignation. And the results would reflect that. Voting,  
> and its results, would become something positive.
>
> So how does this strange situation come about? What causes it? When  
> you compromise in Plurality, for a "lesser-evil", you're saying,  
> with your vote-support, that s/he is better than your favorite.
>
> Plurality can be regarded as a point-rating system, but a funny one  
> in which you're only allowed to give a point to one candidate.  
> You're required to give 0 points to everyone else. Top rating to  
> one, and bottom to everyone else. Those bottom-ratings that you must  
> give to all but one are materially real, in the sense that you're  
> giving to those candidates zero points instead of 1, the lower of  
> the two ratings levels--and thereby voting for them to lose.
>
> Note that it is _not_ that Plurality only lets you rate one  
> candidate. You're rating them all. But you're required to rate all  
> but one of them _at bottom_, voted to lose. That's why I referred to  
> Plurality as a _funny_ point-rating system.
>
> Someone at the forum said that Plurality doesn't count enough  
> information. But that isn't true. Plurality counts plenty of  
> information, but it's mostly false information--all those compulsory  
> zero ratings.  When you say something because you have to, even if  
> you don't feel it, that's falsity. It's no exaggeration to say that  
> Plurality forces falsification.
>
>  It should hardly be surprising that this results in a lot of  
> dissatisfaction with the results of that falsification.
>
>
> Some defend Plurality by saying that, with it, we vote for our  
> favorite. But millions of voters—when they need to "hold their nose"  
> when insincerely helping someone they don't like, over someone they  
> do like--might not agree.
> If Plurality is supposed to have us voting for our favorites, then  
> it is failing miserably. If Plurality assumes that we're voting for  
> our favorites, then Plurality is assuming wrong.
>
> How to avoid this problem? Why not repeal the rule that makes  
> Plurality so funny?  Let people rate _every_ candidate with a  1 or  
> a 0. Rate every candidate as "Approved" or "Unapproved". The  
> candidate with the most "Approved" ratings wins. The result? Well,  
> we'd be electing the most approved candidate, wouldn't we.  Who can  
> criticize that?
>
Anyone who realizes that there is more to wish for.

Here you can vote for both Favorite and Compromise to help defeat  
Worse, but cannot vote for both without implying equal liking for each  
- and thus risking unwanted election of Compromise.

> When everyone can support the candidate(s) they really like, instead  
> of just a "lesser-evil", that can only mean that we elect someone  
> more liked.
>
>  That voting system, the minimal improvement on Plurality to fix its  
> ridiculous problem, is called "Approval voting", or just "Approval".
>
>  Occasionally we hear a claim that Approval violates “1-person-1- 
> vote”.
>
> But Approval is a points rating system. Every voter has the equal  
> power to rate each candidate as approved or unapproved.  1 point or  
> 0 points.
>
> If you approve more candidates, does that give you more power?  
> Hardly. Say you approve all of the candidates. You thereby have zero  
> influence on the election.
>
> And obviously, any ballot will be cancelled out by an oppositely- 
> voted ballot.
>
> Suppose you approve all of the candidates but one. I approve the  
> candidate you didn’t approve, and not ones that you approved . My  
> oppositely-voted ballot cancels yours out. You voted for nearly all  
> of the candidates. I voted for only one. But I cancelled you out.
>
> Some Approval advantages:
>
> Approval  is one of the few voting systems in which you never have  
> any reason to not top-rate your favorite(s).  For the first time,  
> everyone would be able to fully support their favorites.
>
> As said above, when people can fully support the candidates whom  
> they really like, we elect someone better-liked--someone to whom the  
> most people have given approval. That makes an Approval election  
> into something positive and hopeful.
>
> In a presidential straw-poll, using Condorcet, I’ve personally  
> observed someone ranking compromises over their favorite.  In  
> Plurality and Condorcet, that can be the only way to maximally help  
> the compromises against someone worse.  But never in Approval.
>
> That observed favorite-burial in Condorcet suggests to me that many  
> would feel a need to bury their favorites in Condorcet, as they do  
> in Plurality.
>
> Never underestimate voters' need to help a compromise all that they  
> can, even when that's at the expense of their favorite.
> I should add that, in Approval, not only does the voter never have  
> any reason to not top-rate their favorite(s), but it is  
> transparently obvious that that is so. If you have given 1 point to  
> Compromise, and 0 points to Worse, then it’s obvious that also  
> giving a point to Favorite won’t change the fact that you’ve fully  
> helped Compromise against Worse.
>
The above sentence emphasizes what happens to Compromise vs Worse,  
ignoring that it destroys Favorite's desired advantage over Compromise.

> Another thing, which really counts as a separate advantage:  In  
> Plurality, whether people compromise (as they seem so prone to do),  
> or whether, instead, they all vote for someone they like--either  
> way, their votes will be split between their various compromises or  
> favorites. Suppose the progressives add up to at least a majority.  
> That won’t do them any good in Plurality unless they can somehow  
> guess or organize exactly which candidate they’ll combine their  
> votes on That’s especially a reason why voters now let the media  
> lead them by the nose.
>
> That wouldn’t be a problem in Approval, where each person is  
> approving a _set_ of candidates, maybe various favorites and various  
> compromises. It would no longer be necessary to guess where everyone  
> else will combine their votes. In Plurality, that need, especially,  
> makes voters let the media lead them by the nose.
>
> Approval , as I said, is the minimal change that gets rid of  
> Plurality’s ridiculous problem.  There won’t be any question about  
> whether that’s an improvement.  When Plurality’s falsification  
> problem is discussed,  Plurality’s inexplicable problem-causing  
> rule,  then anyone trying to claim that that problem should be kept  
> will be arguing an indefensible position, and will be seen by all  
> for what he is.  I’m not saying that desperate arguments for keeping  
> Plurality’s problem won’t be made. I’m saying that they won’t work.
>
Agreed that Approval was an easy, but valuable, step up from Plurality.

But, Approval does not help us vote our preference for Favorite over  
Compromise.  I offer Condorcet as one easy step for this capability.

Easy for the voter - rank each approved candidate:
.     Each candidate ranked by a voter is preferred over each  
candidate unranked, with much the same power as in Approval.
.     Among those a voter ranks, each given a higher rank is preferred  
over each given a lower rank.

Picking the winner is based on the candidate pairs - best is for a  
candidate to win all its pairs.  Note that, like Approval but unlike  
such as IRV, batches of ballots can be counted into arrays and the  
arrays summed.

The negatives below suggest this is a difficult step.  Agreed, but its  
value says it is worth trying.

Dave Ketchum

> In contrast, when anything more complicated than Approval is  
> proposed , opponents, media pundits and commentators, magazine  
> writers, politicians, and some hired academic authorities will point  
> out that it could have unforeseen and undesired consequences.  
> They’ll take advantage of the fact that the public can’t predict all  
> of the method’s consequences.
>
> They’ll point out that the method could cause disaster, because we  
> don’t know what it would do. Sure, we voting system reform advocates  
> all agree that Condorcet is better than Plurality. But the public  
> won’t know that.
>
> Authorities and pundits will say “It needs a lot more study”, and...
>
> ...it will never happen.
>
> That objection won’t work against Approval,  because Approval is so  
> elegantly simple and transparent,
>
> Approval has a unique optimization. All of the Approval strategies  
> (which I’ll get to in a minute) amount to approving all of the  
> candidates who are better than what you expect from the election.  
> That means that the winner will be the candidate who is better-than- 
> expectation for the most voters. That’s the candidate whose win will  
> pleasantly surprise the most voters.
>
> Anyway, it’s obvious that electing the candidate to whom the most  
> people have given approval is, itself, a valuable optimization.
>
> Approval strategy:
>
> Experience with the several interesting and instructive presidential  
> mock-elections that we've conducted at the election-methods mailing  
> list suggests to me that, in an Approval election, people will  
> typically just know whom  they want to approve. People will have an  
> unmistakable intuitive feel for whom they want to approve.
> For instance, you likely will approve all the candidates whom you  
> like, or who deserve your support. You'll know who they are.
> The suggestions below are just for when you don't have a feel for  
> whom to approve:
> First, you can just approve the candidate you’d vote if it were a  
> Plurality election, and also for everyone whom you like better than  
> him/her (including your favorite).
> That would be good enough.
>
> But Approval has strategy instructions that aren’t available for  
> Plurality, because, for Plurality, they’d be too complicated to  
> fully describe, and much more difficult to implement. So don’t let  
> these suggestions make you think that Approval is more complicated.  
> Approval’s strategy is incomparably simpler than that of Plurality.
>
> If there are unacceptable candidates who could win, then approve all  
> of the acceptables, and none of the unacceptables.
>
> If there are no unacceptable candidates who could win, and if you  
> have no predictive information or feel about winnability, then  
> Approve all of the above-mean (above average) candidates.
>
> If neither of the above 2 paragraphs applies, then Approve all of  
> the candidates who are better than what you expect from the election.
>
> One way to judge that directly would be to ask yourself: “Would I  
> rather appoint him/her to office than hold the election?”  If so,  
> then approve him/her.
>
> But, because we don’t have the power to appoint officeholders, we  
> might not have a good feel for that judgment.  A better question  
> would be:
>
> Is s/he better than what I expect? Do I expect less? If so, then  
> approve him/her.
>
> In other words, vote optimistically.
>
> In fact, even if s/he is right _at_ the merit-level that you expect  
> from the election, then approve him/her if you like him/her.
>
> Why does that maximize your expectation? Because, when (by approving  
> him/her) you improve the win-probability of someone who is better  
> than your expectation, that will raise your statistical expectation.
>
> _All_ of the Approval strategy suggestions are special cases of the  
> rule just given.
>
> For example, maybe you have a feel for who the top-two vote-getters  
> will be. Then, of course, approve the better of those two, and  
> everyone who is better still.  But I hasten to emphasize that the  
> candidates who you might expect to be frontrunners in Plurality are  
> very unlikely to be the frontrunners in Approval.  Never let anyone  
> tell you who the frontrunners will be.
>
> Mike Ossipoff
>
>
>
>
>
> ----
> Election-Methods mailing list - see http://electorama.com/em for  
> list info

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20120427/d01a98b7/attachment-0004.htm>