Loading ...
Sorry, an error occurred while loading the content.
 

Re: [RangeVoting] The R vs D vs D (mythologically anti-Approval) scenario...

Expand Messages
  • Juho
    ... Democrats might split their votes between their two candidates even if they have majority. (These cases are not really in line with the example since in
    Message 1 of 26 , Aug 4, 2008
      On Aug 4, 2008, at 21:14 , warren_d_smith31 wrote:

      >
      >>> In a scenario with 2 democrats running one of whom is likely to
      >>> win,
      >>> republicans would be strategically foolish not to approve a
      >>> democrat.
      >>
      >> The assumption was that also the Republican candidate was a potential
      >> winner.
      >
      > --what'll presumably happen in the
      > R vs D vs D approval-voting scenario
      > (which is also what raphfrk had said, but I will repeat) is this:
      >
      > 1. If the republican looks like he is going to win, then the
      > republican
      > voters will bullet-vote and the dem voters will approve both dems.
      > Because that is the strategic thing for them individually to do.
      >
      > Result: whichever (R or D) side has the majority, will win.

      Democrats might split their votes between their two candidates even
      if they have majority.

      (These cases are not really in line with the example since in these
      cases we have less than three potential winners. There are also some
      corresponding cases further below.)

      >
      > 2. If a dem looks like going to win (but unclear which one),
      > then the republicans will approve both
      > the republican and one dem, while the dem voters will approve only
      > 1 dem.
      > Because that is the strategic thing for them individually to do.
      >
      > Result: whichever (R or D) side has the majority, will win.

      If it looks CERTAIN that one of the Democrats will win then the many
      of the Republicans could vote for their favourite Democrat (not all).
      Otherwise I'd expect large part of them to bullet vote for their own
      candidate and not "surrender".

      If all Democrats bullet vote then the Republicans need only some 34%+
      support to win if they all bullet vote.

      >
      > 3. If it is going to be either R or D1 who wins (D2 out of
      > contention) then
      > eveyrbody will cast a vote for R but not D1, or for D1 but not R
      > (and they also will approve D2 or not, but that will be irrelevant).
      > Because that is the strategic thing for them individually to do.
      >
      > Result: whichever (R or D) side has the majority, will win.

      Yes, in a two candidate race the obvious strategy is quite simple to
      follow also in Approval.

      >
      > 4. If it appears exactly tied and hence unclear which side (R or D)
      > will win,
      > then voters will do whatever they will do. These votes will then
      > cause one side
      > or the other to win. If they appear to be causing
      > one side to win, then please transition to one of the previously
      > analysed cases.
      > If it stays unclear all the way through all the voting until the
      > very last second, then
      > it stays unclear.

      My assumption was that the race is close enough to make all
      candidates and their supporters try to win (and believe in that) all
      the way until the election day.

      >
      > So UNDER THE ASSUMPTION that voters act sensibly and strategically
      > (i.e do not
      > cast votes which they are pretty confident will be wasted because they
      > do not distinguish between the two likely winners) in Juho's
      > scenario, the majority-side (R or D) will always win and if exact
      > tie then
      > the result will be unpredictable. Exactly as it should be.

      So, what was the strategy that they applied? (Sincerity?)

      >
      > But - Juho's entire scenario had been devised to show that would
      > not happen!?!?

      What was the "that" that is not supposed to happen?

      >
      > Why the difference? The reason is simple. Juho supposed asymmetic
      > non-strategic
      > behavior for his voters. He postulated R (or D) voters were
      > asymmetrically
      > more stupid. He then got an asymmetric result.

      Could you elaborate how the voters were different / more stupid.

      >
      > However, in a hypothetical future world with better communication,
      > voters would
      > effectively not be stupid. Hence it would go down the way I said.

      What is the "way" that you refer to?

      >
      > Now in the present world, stupid voters exist. And perhaps even
      > asymmetrically,
      > e.g. perhaps republicans really are stupider than democrats, or the
      > reverse.
      > As Juho appears to believe.

      The only difference between Democrats and Republicans that I see in
      the example is that one of the parties has two candidates while the
      other has only one. That leads to the risk of splitting votes.

      > In that case, Juho's problem could indeed arise.
      > The side containing stupider and less-informed voters would then
      > suffer.
      > The party leaders in the Stupid party would then have incentive to
      > try to improve
      > communications, education, and informedness so that they could
      > compete equally. The Smart party would have incentive not to
      > improve those things, but they could not prevent the Stupid party from
      > trying to improve it for their own people. If technology were
      > cheap and good enough,
      > such improvement efforts would succeed. Furthermore, even without
      > such efforts,
      > the Dem voters would simply SEE from historical experience that
      > they'd just lost an
      > election they should've won (say) and it was purely as a result of
      > their own
      > stupidity. This would incentivize them not to be stupid in the
      > same way next time.

      I lost you here since I don't know who the stupid voters are and how
      you would like them to behave. Please clarify.

      >
      > This kind of historical lesson works to fix the problem. For
      > example, the USA is now
      > extremely 2-party dominated. That is the result of historical
      > lessons where some party
      > lost due to votes for a 3rd-party "spoiler." Consequently next
      > time voters "got wise" and
      > refused to vote third party even if that was their favorite.
      > Consequently 3rd parties in the
      > USA have died out. This is obviously and unquestionably the reason
      > - cumulative
      > historical lessons causing strategy readjustments to the point
      > where votign 3d party in the
      > USA is now extremely rare, below 1%.
      >
      > I see no reason to postulate the same thing will not happen with
      > approval voting, leading
      > to voters voting "intelligently" in approval voting 99% - causing
      > this whole "problem" to be
      > very rare. Indeed, since the "intelligent" approval vote is not
      > even dishonest - unlike the
      > plurality case where thanks to these lessons voters now are
      > massively dishonest - we
      > expect it to work even better in the approval case.

      Approval surely is better than Plurality here. But as claimed,
      Approval may still have problems if one of the major parties
      nominates more than one candidate.

      >
      > So Burr-dilemma can happen, but NOT as a result of intelligent
      > voter strategy. It happens as a result of stupid and/or uninformed
      > voters and gets
      > repaired quite well and quite permanently by historical lessons.
      >
      > But all that was theory. What about practice?
      > I've been examining the 44 or so Pope Elections that were done via
      > approval voting.
      > Did Burr-dilemma-type scenarios ever occur in these elections?
      > Well... that's a bit hard to answer and they had a supermajority
      > approval threshold
      > which kind of changed things.
      > Apparently whenever 1 faction had a large enough majority, they
      > always won, no
      > problem, no issue with vote "splitting".
      > But if no faction had a majority (or 1 faction did, but not
      > large enough) then
      > there was often a deadlock because rival factions had the ability
      > to block the top faction.
      > In that case, they'd sometimes find a compromise or try to bribe or
      > intimidate
      > the blockers.
      >
      > There was at least one Burr-dilemma-type instance which worked
      > thusly. There was a
      > big faction F. Surprisingly, a big rival faction R voted for a
      > particular member f of F.
      > Now if the members of F had known that was going to happen, they
      > also could
      > have all approved f, and f instantly would have won, making them
      > happy.
      > However, they did not know it would happen, and instead only
      > approved certain
      > "frontrunners" in F. Hence f was not elected.
      >
      > Now why was this? Well, it was a game. The game by R was to
      > surprise F with the big
      > unexpected support for f, but which (R felt confident due to the
      > Burr thing) would
      > not elect f. They then would use this surprise as a psychological
      > tool to try to split
      > up F and make them worry there were games afoot they did not know
      > about, make
      > them doubt one another's veracity, and perhaps later gain the
      > support of a subfaction of F.
      >
      > So in that sense a Burr-dilemma did occur, but it was part of a
      > larger game.
      >
      > --Warren D. Smith

      And back to practice and to the old question once more. The claim was
      that Democrats will have hard time voting any sensible way. How
      should the Republican voters vote in the R vs D vs D scenario? How
      should the Democrats vote in the R vs D vs D scenario? You say that
      in my analysis the voters voted in some stupid/foolish way. What is
      the sensible way for them to vote?

      Juho






      ___________________________________________________________
      The all-new Yahoo! Mail goes wherever you go - free your email address from your Internet provider. http://uk.docs.yahoo.com/nowyoucan.html
    • warren_d_smith31
      - ... But I explained how to vote in all cases. You then ignored how I said people should vote, and asked how I thought people should vote. To return the the
      Message 2 of 26 , Aug 4, 2008
        -
        > And back to practice and to the old question once more. The claim was
        > that Democrats will have hard time voting any sensible way. How
        > should the Republican voters vote in the R vs D vs D scenario? How
        > should the Democrats vote in the R vs D vs D scenario? You say that
        > in my analysis the voters voted in some stupid/foolish way. What is
        > the sensible way for them to vote?
        >
        > Juho

        But I explained how to vote in all cases.

        You then ignored how I said people should vote, and asked how I thought
        people should vote.

        To return the the case which causes you the most difficulty:
        the republican voters, if they think a Dem will win, should approve 1 dem
        and the Repub.

        If they do not, then they are stupid and their vote, since it does not distinguish
        between the 2 likely winners, will be wasted.

        You postulated the R-voters were stupid, but the D-voters were not stupid.
        This is asymmetric.

        I actually gave 2 cases. The above was just one case.
        In both cases, if voters are smart, the majority side wins.

        If the voters are unsure which of the 2 cases they have, then they guess case 1 or 2.
        If the R- and D-voters guess in the same proportions, then the same thing
        happens that I'd said and the majority side wins.
        If they guess in different proportions, then
        they were behaving asymmetrically - each side believing in a different
        version of the facts. There is however only one fact.

        Indeed as Juho himself stated, he postulated both sides believed they were winning.
        If so, the Rs would have an artificial advantage.

        Why did Juho not make the equally ridiculous postulate that both sides believed they were
        losing? Then the Ds would have an artificial advantage!

        ---

        If however, the R-leadership blared propaganda that they were winning, trying to gain
        advantage. then the Ds would probably hear that propaganda. It would not
        be a secret. The R's advantage would then vanish.

        Similarly, if the D-leadership blared propaganda that they were losing, that would only
        be accentuated by the Republican propaganda that the Rs were winning...
        which only would help the Ds. (However, the D propaganda wuld help the Rs when Rs
        heard it.) The Rs could not counter it by declaring that THEY were losing,
        since that'd hurt their side...

        Moral is that even if one side has a big propaganda organ and the other side is
        muzzled, that still does not help, provided they, when propagandizing, cannot
        keep the propaganda a secret.

        This all is assuming logical voters. Real voters might not be logical and their
        illogic might be assymetric. But again, I repeat, the problem is NOT a consequence
        of voter logic. It is a consequence, when it occurs, of voter stupidity or ignorance..

        -wds
      • Juho
        ... The closest you came to giving an answer to this basic question was in point 4 (that was about the original scenario where all three candidates are
        Message 3 of 26 , Aug 4, 2008
          On Aug 5, 2008, at 3:52 , warren_d_smith31 wrote:

          > -
          >> And back to practice and to the old question once more. The claim was
          >> that Democrats will have hard time voting any sensible way. How
          >> should the Republican voters vote in the R vs D vs D scenario? How
          >> should the Democrats vote in the R vs D vs D scenario? You say that
          >> in my analysis the voters voted in some stupid/foolish way. What is
          >> the sensible way for them to vote?
          >>
          >> Juho
          >
          > But I explained how to vote in all cases.

          The closest you came to giving an answer to this basic question was
          in point 4 (that was about the original scenario where all three
          candidates are potential winners).

          You wrote:
          > 4. If it appears exactly tied and hence unclear which side (R or D)
          > will win,
          > then voters will do whatever they will do.


          This is also why I had some doubts that you might recommend sincerity
          to the voters. Was this a correct assumption?

          (Sincerity is hard to define in Approval but here it should be read
          to mean voting according to the basic Approval strategy => approve
          those candidates that you want to promote when compared to the
          expected outcome.)

          > You then ignored how I said people should vote, and asked how I
          > thought
          > people should vote.
          >
          > To return the the case which causes you the most difficulty:
          > the republican voters, if they think a Dem will win, should approve
          > 1 dem
          > and the Repub.

          I don't see any difficulties in this specific case. If they are sure
          they will lose (this was not part of the original scenario) they
          should approve one of the potential winners (this is the well known
          basic Approval strategy). If they think they still have chances to
          win and their preferences are R>>>D1>D2 (the default case in the two
          party example) then they probably should stick to approving the
          Republican candidate only. If they are somewhere in between then they
          will experience some of the Approval related pain of not knowing for
          sure which strategy would be the best.

          > If they do not, then they are stupid and their vote, since it does
          > not distinguish
          > between the 2 likely winners, will be wasted.

          We can drop the two winner cases from the discussion since we agree
          what the best strategy is in those cases. The problems appear only
          when there are three potential winners.

          (The case proposed by raphfrk where Republicans were expected to like
          one of the Democrat candidates more than the other is also
          interesting but it is maybe better to handle that case (with three
          potential winners) only after knowing how you recommend the basic
          problem case with three potential winners but no generally known bias
          in the Republican opinions) to be handled.)

          > You postulated the R-voters were stupid, but the D-voters were not
          > stupid.
          > This is asymmetric.

          You said they would be stupid. I thought they did in the example
          (=had a tendency not to approve Democrat candidates) what was quite
          natural for them to do.

          > I actually gave 2 cases. The above was just one case.
          > In both cases, if voters are smart, the majority side wins.
          >
          > If the voters are unsure which of the 2 cases they have, then they
          > guess case 1 or 2.

          This need to guess is one of the problems of Approval (as mentioned
          already above). There is no sincere easy way to vote. In many cases
          the outcome of the election depends on how the voters will decide.
          And it is possible to influence the voter decisions by propaganda
          (I'll skip further analysis for the time being since this mail is
          getting lengthy; the discussed example is maybe one of the most
          problematic such cases).

          > If the R- and D-voters guess in the same proportions, then the same
          > thing
          > happens that I'd said and the majority side wins.

          I assume that you mean that n% of the Democrats (maybe n% of the
          supporters of both Democrat candidates) will bullet vote and n% of
          the Republicans will approve also one of the Democrat candidates
          (maybe giving equal number of approvals to both Democratic candidates).

          > If they guess in different proportions, then
          > they were behaving asymmetrically - each side believing in a different
          > version of the facts.

          The situation is not symmetric. It is probably much easier
          psychologically for the Democrats to bullet vote for their favourite
          candidate than it is for the Republicans to approve one of the
          candidates of the competing party. (This might be the key difference
          in our thinking.)

          > There is however only one fact.
          >
          > Indeed as Juho himself stated, he postulated both sides believed
          > they were winning.
          > If so, the Rs would have an artificial advantage.

          I assumed that all candidates are potential winners and that also
          their supporters believe so (not that their supporters would be
          convinced that their favourite would actually win).

          > Why did Juho not make the equally ridiculous postulate that both
          > sides believed they were
          > losing? Then the Ds would have an artificial advantage!

          See the previous answer. I can also add here that generally
          candidates and voters tend to be optimistic or a least act as if they
          were, which means that they will try to present their side as a
          potential winner for as long as they can.

          > ---
          >
          > If however, the R-leadership blared propaganda that they were
          > winning, trying to gain
          > advantage. then the Ds would probably hear that propaganda. It
          > would not
          > be a secret. The R's advantage would then vanish.

          I believe all candidates would be spreading the "I'm the likely
          winner" propaganda. I don't know how that would change "R's advantage".

          > Similarly, if the D-leadership blared propaganda that they were
          > losing, that would only
          > be accentuated by the Republican propaganda that the Rs were
          > winning...
          > which only would help the Ds. (However, the D propaganda wuld
          > help the Rs when Rs
          > heard it.) The Rs could not counter it by declaring that THEY were
          > losing,
          > since that'd hurt their side...

          Of course no candidate will give up and declare oneself as a loser if
          the polls show that they are winning or few votes away from winning.

          > Moral is that even if one side has a big propaganda organ and the
          > other side is
          > muzzled, that still does not help, provided they, when
          > propagandizing, cannot
          > keep the propaganda a secret.

          I assume all propaganda and all polls and all discussion in the media
          and strategic recommendations will be available to all. That is the
          nature of large public elections (my default election type if nothing
          else is assumed).

          > This all is assuming logical voters. Real voters might not be
          > logical and their
          > illogic might be assymetric.

          I assume about the same number of strategy aware, strategic, sincere,
          uneducated etc. voters on both sides.

          > But again, I repeat, the problem is NOT a consequence
          > of voter logic. It is a consequence, when it occurs, of voter
          > stupidity or ignorance..
          >
          > -wds

          My best guess of the strategy that you recommend to the voters is now
          that they should vote according to the regular Approval strategy in
          the case where there are three potential winners (R, D, D) (in
          approximate balance in all directions).

          That would mean something like Democrats (with feelings D1>D2>>>R)
          approving the two Democrats, and Republicans (with feelings
          R>>>D1>D2) approving only R.

          That would then mean that in the case that Democrats actually will
          get majority of the votes on the election day the Democrat candidates
          will be tied. Or in real life the decision will be made
          (undemocratically, not taking into account the preferences of the
          whole voter community) by those few (or numerous) voters that did not
          follow the strategy, just like proposed in the original example.

          Of course in real life all would not follow the optimal strategy. My
          assumption was that it would be more typical among the Democrats to
          bullet vote than among the Republicans to approve one of the Democrat
          candidates (for psychological reasons). That would mean advantage to
          the Republicans.

          Do we agree that the voters have a dilemma of not being able to
          support both Democrats in general and one of them in particular at
          the same time? Do we agree on that Republicans would find it
          psychologically harder to approve one of the Democrats than Democrats
          would find to bullet vote for their favourite candidate?

          Juho









          ___________________________________________________________
          All new Yahoo! Mail "The new Interface is stunning in its simplicity and ease of use." - PC Magazine
          http://uk.docs.yahoo.com/nowyoucan.html
        • warren_d_smith31
          ... --yes. They cannot do both in approval voting. ... --no. I see zero actual evidence for this, I just see you proclaiming it. It might be true. Might not.
          Message 4 of 26 , Aug 5, 2008
            > Do we agree that the voters have a dilemma of not being able to
            > support both Democrats in general and one of them in particular at
            > the same time?

            --yes. They cannot do both in approval voting.

            > Do we agree on that Republicans would find it
            > psychologically harder to approve one of the Democrats than Democrats
            > would find to bullet vote for their favourite candidate?

            --no. I see zero actual evidence for this, I just see you proclaiming it.
            It might be true. Might not.

            Try this thought experiment. I offer you $10, $1 or $0,
            but with the caveat that it is very likely that my offer of $10 will be rescinded before
            I hand you the money. You can approve or disapprove all choices.
            What do you do?

            --You also proclaimed that "of course" no candidate would declare that they think they
            are losing.

            But why not??

            I can see why not with the current plurality-voting system (if you are seen as losing,
            then it is not worthwhile to vote for you since vote will be wasted in a >=3-candidate
            election).

            But with approval voting, if you are seen to be losing then that in no way prevents
            or makes it wise for anybody to cease to vote for you.
            Therefore, nobody will have incentive to pretend to VOTERS that they are winning
            when they are not.

            There will still be incentive to pretend one is winning in order to attract bribe money.
            There also will still be incentive to pretend one has winning chances to attract
            donor money (pointless to donate to a sure loser). However, since it also is
            pointless to donate to a sure-winner, there will be incentive
            to pretend you have losing chances also.

            So in summary, there is no reason to pretend to VOTERS you are winning.

            There is no reason to pretend to HONEST DONORS you are winning.

            There however is reason to pretend to BRIBERS you are winning.

            However, if candidates want to argue for bribe-purposes they are winning, then
            they will have to argue the others are losing. All these arguments conflict.
            Only one can be true. With good communication and technology, the true argument will
            be seen to be true and the false arguments will be seen to be false, because there
            will be actual evidence.
            And those bribers will demand actual evidence. They are not about to give away big
            money on the basis of no evidence.
          • Juho
            ... Ok. My conclusion from this was that if the Democrats want to make sure that a Democrat will be elected in the 50%-50% situation the choice between the
            Message 5 of 26 , Aug 5, 2008
              On Aug 5, 2008, at 20:26 , warren_d_smith31 wrote:

              >> Do we agree that the voters have a dilemma of not being able to
              >> support both Democrats in general and one of them in particular at
              >> the same time?
              >
              > --yes. They cannot do both in approval voting.

              Ok. My conclusion from this was that if the Democrats want to make
              sure that a Democrat will be elected in the 50%-50% situation the
              choice between the Democrat candidates becomes quite random.

              >> Do we agree on that Republicans would find it
              >> psychologically harder to approve one of the Democrats than Democrats
              >> would find to bullet vote for their favourite candidate?
              >
              > --no. I see zero actual evidence for this, I just see you
              > proclaiming it.
              > It might be true. Might not.

              Ok. To me this looks quite natural. Actually many Democrats might
              even find it more natural to bullet vote than to approve also the
              competing Democrat candidate.

              (This was the reason why Republicans were in a better position (more
              likely to win) than the Democrats even if the Democrats are made
              aware that they should approve both Democrat candidates.)

              > Try this thought experiment. I offer you $10, $1 or $0,
              > but with the caveat that it is very likely that my offer of $10
              > will be rescinded before
              > I hand you the money. You can approve or disapprove all choices.
              > What do you do?

              Maybe you would pick the offer randomly from all the choices but
              adding some additional weight on those choices that I will approve.

              With these small sums of money, and assuming that I don't like nor
              dislike gambling, I could simply count the expected utilities and
              approve the first two. Higher sums would change my behaviour (also in
              elections high risks might e.g. not be something to gamble about, but
              high wins would).

              I'm not quite sure how you link this to the election example.

              The Republican voters will take a risk if they approve one of the
              Democrats. They might get the (slightly) better one of the Democrats,
              but they know they are sabotaging their possible 51% victory. With
              the assumed utilities that would not make much sense.

              The Democrats are gambling with the Democrat candidates. Widespread
              vote splitting may mean that Republicans will win with high
              probability. With the assumed utilities the risk is not worth taking.
              But since bullet voting is easy and very natural many might do that
              even though it is not strategically optimal.

              > --You also proclaimed that "of course" no candidate would declare
              > that they think they
              > are losing.
              >
              > But why not??

              My feelings are based on my experiences of the politicians. I have
              seen some presidential candidates with 5% support and still refusing
              to accept claims that they are not going to win, and still claiming
              that it is possible that they will win or are expecting a landslide
              of votes since they are the best candidates anyway and voters
              understand and support that. And I agree that this is sensible
              behaviour. There is no point in taking part in an election but not
              taking the election seriously, with the intention to win. Also the
              sincere supporters of those candidates could feel betrayed if the
              candidate herself stops believing in herself and her chances earlier
              than the loyal supporters do.

              In some countries there is a tradition that candidates may withdraw
              during the race. That is ok too if the situation is already clear.

              In the given example all three candidates however had good chances to
              win, so in that case it would be quite exceptional to give up.

              > I can see why not with the current plurality-voting system (if you
              > are seen as losing,
              > then it is not worthwhile to vote for you since vote will be wasted
              > in a >=3-candidate
              > election).

              Yes, in Plurality most third party candidates are spoilers and should
              therefore (in principle) withdraw or not even be nominated. Except
              that there may be some secondary reasons for participation like
              preparing for the following elections, being a channel for protests etc.

              (One could even nominate competing spoilers for strategic reasons.
              Republicans could set up a campaign for a Democrat looking candidate
              (or finance some already existing campaign) just to steal the crucial
              votes from the Democrat candidate.)

              > But with approval voting, if you are seen to be losing then that in
              > no way prevents
              > or makes it wise for anybody to cease to vote for you.
              > Therefore, nobody will have incentive to pretend to VOTERS that
              > they are winning
              > when they are not.

              I think all candidates tend to lean in this direction anyway (i.e.
              behaving as if they were more popular than they are), and many
              candidates will be presented as potential winners. For some it may be
              enough to say that they are not going to win but trying to collect
              e.g. 30% approval (when their expected result is 15%) (this could
              still be interpreted among the voters as a positive attitude, maybe
              preparing for the next elections).

              I think also the "VOTERS" like candidates that look like credible
              winners. Candidates that exaggerate too much might be punished
              (credibility lost) but some exaggeration is probably strategically wise.

              > There will still be incentive to pretend one is winning in order to
              > attract bribe money.
              > There also will still be incentive to pretend one has winning
              > chances to attract
              > donor money (pointless to donate to a sure loser). However, since
              > it also is
              > pointless to donate to a sure-winner, there will be incentive
              > to pretend you have losing chances also.

              Yes, also these reasons are valid.

              > So in summary, there is no reason to pretend to VOTERS you are
              > winning.

              I think there is. One psychological explanation (there are many) is
              that many people tend to vote potential winners (that way their vote
              "is not lost"). For this reason it might sometimes even pay off to
              publish false polls (to gain a place among the identified potential
              winners).

              > There is no reason to pretend to HONEST DONORS you are winning.

              Depends on the definition of "HONEST DONOR". If they just want to
              support the ideology then they might be immune to whether the
              candidate will win in these elections or not.

              > There however is reason to pretend to BRIBERS you are winning.

              Yes. I assume that a "BRIBER" is an instance that expects some
              benefits (that they have paid for) after the candidate has won. (A
              donor that just expects benefits but not as a return gift for the
              paid money is maybe not called a "BRIBER" but something else.)

              > However, if candidates want to argue for bribe-purposes they are
              > winning, then
              > they will have to argue the others are losing. All these
              > arguments conflict.
              > Only one can be true. With good communication and technology, the
              > true argument will
              > be seen to be true and the false arguments will be seen to be
              > false, because there
              > will be actual evidence.

              I think people are quite good at believing in what they want to
              believe in. It is very easy to believe that the other candidates are
              wrong (e.g. with respect to the anticipated result), not our candidate.

              > And those bribers will demand actual evidence. They are not about
              > to give away big
              > money on the basis of no evidence.

              Companies giving out big money are probably more rational than
              individuals and ideological groups.

              Juho








              ___________________________________________________________
              All new Yahoo! Mail "The new Interface is stunning in its simplicity and ease of use." - PC Magazine
              http://uk.docs.yahoo.com/nowyoucan.html
            • warren_d_smith31
              Juho, you are trapped in plurality-based thinking! ... --that s an example. You think plurality is natural. Therefore this makes sense to you. In an
              Message 6 of 26 , Aug 5, 2008
                Juho,
                you are trapped in plurality-based thinking!

                > Actually many Democrats might
                > even find it more natural to bullet vote than to approve also the
                > competing Democrat candidate.

                --that's an example. You think plurality is "natural." Therefore this makes sense to you.
                In an approval-voting world, that thinking is irrelevant.

                Basically, your criticism is "they'll plurality vote." If you are going to make that criticism,
                why not criticize, say, Condorcet methods too? The voters will ust plurality
                vote since it is "more natural" therefore Condorcet will just
                elect the plurality winner, and we'll still have spoilers, etc etc,
                and so Condorcet is no improvement, "QED."

                > But since bullet voting is easy and very natural many might do that
                > even though it is not strategically optimal.

                --it is not "easier" because voters have to approve or disapprove every candidate.
                If they approve 1 and disapprove N-1, that is not less work
                than approving 4 and disapproving N-4.

                And it is only "natural" for those of us trapped in plurality-based thinking --
                which would be irrelevant in the future approval-based world.

                .
                > > --You also proclaimed that "of course" no candidate would declare
                > > that they think they
                > > are losing.
                > >
                > > But why not??
                >
                > My feelings are based on my experiences of the politicians. I have
                > seen some presidential candidates with 5% support and still refusing
                > to accept claims that they are not going to win, and still claiming
                > that it is possible that they will win or are expecting a landslide

                --you again are trapped in plurality-based thinking and so it all your
                "experience." They are doing that becasue it makes sense to do that with plurality voting.
                (And those politicians know full well they are lying. I have seen
                them claim they'd win 1 day and withdraw the very next day. Look right now
                at Hillary Clinton and the VP slot.)

                > In the given example all three candidates however had good chances to
                > win, so in that case it would be quite exceptional to give up.

                --saying "it looks like I will lose" is not the same as "giving up."
                You again are trapped in plurality-based thinking! Get over it!

                > I think all candidates tend to lean in this direction anyway (i.e.
                > behaving as if they were more popular than they are), and many
                > candidates will be presented as potential winners.

                --WHY? It is actually as though you PREFER your
                candidates to be unrealistic and regard them being realistic as some kind of sin!!
                That is how incredibly trapped in plurality-based thinking you are - you
                actually regard plurality's horrible flaws as a bonus feature!

                > I think also the "VOTERS" like candidates that look like credible
                > winners. Candidates that exaggerate too much might be punished
                > (credibility lost) but some exaggeration is probably strategically wise.

                --the "wisdom" as far as I can see is because they want to attract bribe money.

                > > So in summary, there is no reason to pretend to VOTERS you are
                > > winning.
                >
                > I think there is. One psychological explanation (there are many) is
                > that many people tend to vote potential winners (that way their vote
                > "is not lost").

                --again you are trapped in plurality-based thinking. This reasoning is not
                valid with approval voting.

                > I think people are quite good at believing in what they want to
                > believe in. It is very easy to believe that the other candidates are
                > wrong (e.g. with respect to the anticipated result), not our candidate.

                --yeah right. Dream on. How many Nader voters thought Nader was going to win?
                Neither Nader nor his voters thought he would win. Admit the obvious.

                > > And those bribers will demand actual evidence. They are not about
                > > to give away big
                > > money on the basis of no evidence.
                >
                > Companies giving out big money are probably more rational than
                > individuals and ideological groups.

                --I disagree... people analyse harder the more money is at stake...
                including (I suspect) you.
                And it is well known that many companies+PACs making big political contribs did so in
                a highly bribe-strategic way like donating to both sides, "switching horses"
                when it looked like X would no longer win, donating to people with no opponent, etc.
                Regular honest people don't do that.
              • Juho
                ... Not me. I think the voters are to some extent bound to that, both for historical reasons and for general psychological reasons. ... I said that bullet
                Message 7 of 26 , Aug 5, 2008
                  On Aug 6, 2008, at 1:36 , warren_d_smith31 wrote:

                  > Juho,
                  > you are trapped in plurality-based thinking!

                  Not me. I think the voters are to some extent bound to that, both for
                  historical reasons and for general psychological reasons.

                  >> Actually many Democrats might
                  >> even find it more natural to bullet vote than to approve also the
                  >> competing Democrat candidate.
                  >
                  > --that's an example. You think plurality is "natural." Therefore
                  > this makes sense to you.
                  > In an approval-voting world, that thinking is irrelevant.
                  >
                  > Basically, your criticism is "they'll plurality vote."

                  I said that bullet voting is natural. That correlates with the
                  Plurality case but is more general. I also said that voting for the
                  worst competing party is unnatural. Together these will make the
                  position of the party with several candidates worse than the position
                  of the party that has nominated only one candidate.

                  I btw think it is also natural for human beings to think to some
                  extent like A>>>>B>>>C>>D>E. That means that they idolize their
                  favourites, and next idolize their favourite grouping, think that all
                  others are generally wrong. (=> bullet vote, "party vote") (have you
                  noted this in the election method discussions :-) ) There is also
                  some tendency to demonize the worst competitors and worst enemies.
                  (=> no votes to the evil ones)

                  > If you are going to make that criticism,
                  > why not criticize, say, Condorcet methods too? The voters will ust
                  > plurality
                  > vote since it is "more natural" therefore Condorcet will just
                  > elect the plurality winner, and we'll still have spoilers, etc etc,
                  > and so Condorcet is no improvement, "QED."

                  I'm sure there is some tendency also in ranked methods to cast short
                  votes. One reason is pure laziness (also in Approval). Ranked methods
                  however lack the dilemma of being forced to lift all the others to
                  the same level with one's favourite. It is thus still very easy to
                  vote D1>D2. This is also a short vote that doesn't rank any of the
                  reminding candidates (there could be several Republicans an others).

                  In Approval one of the key problems (behind this problem case) is
                  that by approving both D1 and D2 the voter must give up giving full
                  support to her favourite candidate and force her to "rank" her
                  favourite at the same level with the next best ones (that may
                  actually be seen as the worst competitors of her favourite if the
                  voter believes that her party is likely to win).

                  I have also often criticized winning votes based Condorcet methods on
                  that they to some extent favour truncation (=> also bullet voting).
                  But this is luckily a very marginal feature.

                  Condorcet is maybe the best group of methods from this discussion
                  point of view in the sense that vote D1>D2>R is counted as D1>D2,
                  D1>R and D2>R, and all these three preferences in full strength. Most
                  other methods do not allow the voter to express all these preferences
                  as clearly.

                  >> But since bullet voting is easy and very natural many might do that
                  >> even though it is not strategically optimal.
                  >
                  > --it is not "easier" because voters have to approve or disapprove
                  > every candidate.
                  > If they approve 1 and disapprove N-1, that is not less work
                  > than approving 4 and disapproving N-4.
                  >
                  > And it is only "natural" for those of us trapped in plurality-based
                  > thinking --
                  > which would be irrelevant in the future approval-based world.

                  I claim that these tendencies are more general than just Plurality
                  history (or why not also Top Two Runoff history) based.

                  The easiness means also psychological easiness, not only the required
                  labour when voting. And in the ballots I'd expect the typical default
                  value to be to count unmarked candidates as "not approved" rather
                  than as "approved". I wonder if you thought that it would be
                  mandatory to mark all candidates explicitly either "approved" or "not
                  approved" (or otherwise the ballot would be rejected).

                  > .
                  >>> --You also proclaimed that "of course" no candidate would declare
                  >>> that they think they
                  >>> are losing.
                  >>>
                  >>> But why not??
                  >>
                  >> My feelings are based on my experiences of the politicians. I have
                  >> seen some presidential candidates with 5% support and still refusing
                  >> to accept claims that they are not going to win, and still claiming
                  >> that it is possible that they will win or are expecting a landslide
                  >
                  > --you again are trapped in plurality-based thinking and so it all your
                  > "experience." They are doing that becasue it makes sense to do that
                  > with plurality voting.

                  The examples are from Top Two Runoffs. The probability that I would
                  be trapped in Plurality-based thinking for historical reasons is
                  about zero :-). Same with my "experience".

                  > (And those politicians know full well they are lying. I have seen
                  > them claim they'd win 1 day and withdraw the very next day. Look
                  > right now
                  > at Hillary Clinton and the VP slot.)

                  That is also natural. I wouldn't call that lying (although that might
                  be true from a "purely mechanical fact based and time and change
                  ignoring thinking" point of view) since the candidates are generally
                  expected to present themselves as potential winners of the election
                  (until they withdraw, win or lose).

                  >> In the given example all three candidates however had good chances to
                  >> win, so in that case it would be quite exceptional to give up.
                  >
                  > --saying "it looks like I will lose" is not the same as "giving up."
                  > You again are trapped in plurality-based thinking! Get over it!

                  I think saying "it looks like I will lose" will in most cases
                  contribute heavily towards giving the voters an impression that this
                  candidate is not one of the main contenders. And although some could
                  vote this candidate for sympathy reasons or for ideological reasons
                  it is more common that her ranking will drop (rather than rise) in
                  the minds of the voters.

                  Approval allows the voters to approve also irrelevant candidates
                  (even those that have "given up"). they don't pose similar threat to
                  the voter as the candidates that are potential winners (the root
                  cause of the failure example we are discussion about).

                  Coming back to my referred text above. I think saying "it looks like
                  I will lose" would also make no sense in that situation.

                  >> I think all candidates tend to lean in this direction anyway (i.e.
                  >> behaving as if they were more popular than they are), and many
                  >> candidates will be presented as potential winners.
                  >
                  > --WHY? It is actually as though you PREFER your
                  > candidates to be unrealistic and regard them being realistic as
                  > some kind of sin!!
                  > That is how incredibly trapped in plurality-based thinking you are
                  > - you
                  > actually regard plurality's horrible flaws as a bonus feature!

                  Why did Obama go to Berlin and give a speech next to the (now
                  demolished) Berlin wall? Did he catch some of the glory of Kennedy
                  and Reagan that way? (Well, I don't know if this wen unnoticed in the
                  USA - probably not.) Did he make himself look like a potential
                  president instead of just one of the politicians? I think he could
                  have done that even if the election method would be some other than
                  Plurality.

                  >> I think also the "VOTERS" like candidates that look like credible
                  >> winners. Candidates that exaggerate too much might be punished
                  >> (credibility lost) but some exaggeration is probably strategically
                  >> wise.
                  >
                  > --the "wisdom" as far as I can see is because they want to attract
                  > bribe money.
                  >
                  >>> So in summary, there is no reason to pretend to VOTERS you are
                  >>> winning.
                  >>
                  >> I think there is. One psychological explanation (there are many) is
                  >> that many people tend to vote potential winners (that way their vote
                  >> "is not lost").
                  >
                  > --again you are trapped in plurality-based thinking. This
                  > reasoning is not
                  > valid with approval voting.

                  That argument is valid in the sense that in Approval one can continue
                  voting also those that are not going to win. But it still makes sense
                  to advertise oneself as a potential winner in order to get on the "to
                  be approved" list. And also to avoid falling from that list.

                  >> I think people are quite good at believing in what they want to
                  >> believe in. It is very easy to believe that the other candidates are
                  >> wrong (e.g. with respect to the anticipated result), not our
                  >> candidate.
                  >
                  > --yeah right. Dream on. How many Nader voters thought Nader was
                  > going to win?
                  > Neither Nader nor his voters thought he would win. Admit the obvious.

                  It is not necessary to believe that Nader is going to win. It is also
                  possible to attached to him and be one of his loyal supporters,
                  thinking that the candidates of the main parties are wrong, even if
                  Nader wouldn't have a chance. (It is possible that other candidates
                  had more "believing" supporters that Nader had but certainly that
                  phenomenon is present also in the minds of the Nader supporters.)

                  >>> And those bribers will demand actual evidence. They are not about
                  >>> to give away big
                  >>> money on the basis of no evidence.
                  >>
                  >> Companies giving out big money are probably more rational than
                  >> individuals and ideological groups.
                  >
                  > --I disagree... people analyse harder the more money is at stake...
                  > including (I suspect) you.

                  Ok, poor people have more reasons to count their money and that may
                  increase their level of rational thinking (of course the amount of
                  donated dollars is also smaller).

                  > And it is well known that many companies+PACs making big political
                  > contribs did so in
                  > a highly bribe-strategic way like donating to both sides,
                  > "switching horses"
                  > when it looked like X would no longer win, donating to people with
                  > no opponent, etc.
                  > Regular honest people don't do that.

                  That could be also rational.

                  In general I do believe that companies are also often irrational due
                  to the difficulties of making decisions in a group. They could e.g.
                  be bound by the legacy of budgets and habits of previous years, by
                  some high boss having some weir political ideas. Companies may also
                  fear that the competing companies have already donated money or are
                  about to donate, and candidates might take advantage of this
                  competitive situation etc. etc.

                  Juho






                  ___________________________________________________________
                  Try the all-new Yahoo! Mail. "The New Version is radically easier to use" � The Wall Street Journal
                  http://uk.docs.yahoo.com/nowyoucan.html
                • warren_d_smith31
                  ... --with top-two-runoff voting, politicians also have incentive to pretend they are winning or have winning chances. Because It is a wasted vote in TTR to
                  Message 8 of 26 , Aug 6, 2008
                    > > Juho,
                    > > you are trapped in plurality-based thinking!
                    >
                    > Not me.
                    > The examples are from Top Two Runoffs.

                    --with top-two-runoff voting, politicians also have incentive to pretend
                    they are winning or have winning chances.

                    Because It is a wasted vote in TTR to vote for somebody not in the top 3.

                    With approval voting, it is not a wasted vote to approve somebody even if
                    they have extremely tiny winning chances.

                    Therefore, if your experience+examples are based on TTR, I still deny its applicability
                    to approval voting.
                  • Juho
                    ... It is typically quite ok to vote for some of the lesser candidates at the first round in TTR. In some cases like if your second favourite is the third
                    Message 9 of 26 , Aug 7, 2008
                      On Aug 6, 2008, at 19:46 , warren_d_smith31 wrote:

                      >> The examples are from Top Two Runoffs.
                      >
                      > --with top-two-runoff voting, politicians also have incentive to
                      > pretend
                      > they are winning or have winning chances.
                      >
                      > Because It is a wasted vote in TTR to vote for somebody not in the
                      > top 3.

                      It is typically quite ok to vote for some of the lesser candidates at
                      the first round in TTR. In some cases like if your second favourite
                      is the third strongest candidate then you maybe should vote for that
                      second favourite at the first round instead of your first favourite.
                      Just like Approval requires strategy analysis before voting also TTR
                      requires that - at least in theory. In practice TTR voters probably
                      mostly do not bother.

                      (My experiences are typically from situations where there are at
                      least three strong candidates. I believe voting for one's top
                      favourite has been dominant. The second round seems to offer
                      sufficient opportunity (in the minds of regular voters and
                      politicians) to fix the possible randomness resulting from allowing
                      only the top two candidates (with most first place support) to go to
                      the second round.)

                      > With approval voting, it is not a wasted vote to approve somebody
                      > even if
                      > they have extremely tiny winning chances.

                      Yes, Approval allows multiple approvals that makes this possible. In
                      TTR the first round has somewhat similar function.

                      > Therefore, if your experience+examples are based on TTR, I still
                      > deny its applicability
                      > to approval voting.

                      The TTR experiences (I didn't give any such examples) were only to
                      provide example cases of human psychology. They should apply to most
                      election methods. Similar examples are not available on Approval
                      since I don't have experiences not other good data on Approval
                      elections available. The claims are thus not TTR specific in any way.

                      Juho









                      ___________________________________________________________
                      All new Yahoo! Mail "The new Interface is stunning in its simplicity and ease of use." - PC Magazine
                      http://uk.docs.yahoo.com/nowyoucan.html
                    Your message has been successfully submitted and would be delivered to recipients shortly.