[EM] Instant Runoff Voting 3-candidate elections - pathologies considerably more common than you may have thought

[robert bristow-johnson]

On Aug 27, 2010, at 2:29 PM, Jameson Quinn wrote:

> Bucklin, too.

Bucklin uses the same ranked-choice ballot that Condorcet or IRV or  
Borda uses (ignoring that some allow for equal ratings), so it can be  
compared directly to any.  and if we assume that the highest rank vote  
is the same as the "single affirmative vote" of FPTP, it can be  
compared to that.

my position is that sometimes these methods elect the Condorcet winner  
and sometimes they do not.  i would say that at least one pathology  
exists when they elect someone different than the Condorcet winner.


> >> It is reasonable, in the face of such massive and frequently- 
> arising
> >> evidence that IRV has (obvious) problems, to promote it, as  
> opposed to
> >> some simpler method largely free of such problems?
> >
> >
> > is the some simpler method either Approval or Score Voting?
> >
> > what simpler method are you suggesting, Warren?



> 2010/8/27 Warren Smith <warren.wds at gmail.com>
> --score & approval are simpler and are comparatively free of crazy  
> pathologies.

i started engaging Clay Shentrup about this, but i really felt "piled  
upon" at his ESF group (i quickly unsubscribed) and that they were not  
listening at all.  but this main thesis you're selling, Warren, needs  
discussion and defending.  and it should be here rather than at ESF,  
which really should be renamed the Score Voting Advocacy Forum.

i will say this again; both Score (or Range) Voting and Approval  
Voting have *inherent* strategic (or "tactical", i really don't know  
which word is best used, i *do* know the difference in a military  
context) burdens for the voter.  because, as i examine my own feelings  
about voting, and as i talk with other voters about election systems  
(because of IRV and a credible third party, there is a lot of  
discussion about it here in Vermont), both Score and Approval will  
degenerate to Plurality for voters that just do not wish to harm their  
favorite candidate.  and, in talking with people and examining my own  
motivations, i really believe that this is most voters.  if the large  
majority of voters score their fav with 99 and everyone else gets 0  
(and the equivalent with approval), how is that any different than a  
scaled outcome of FPTP?  these methods will differ from FPTP (which i  
*know* sucks) only to the degree that people will forsake their  
favorite, at least a little.

and, in the Chittenden State Senate district in Vermont, we  
*virtually* have Approval (vote for up to 6; top 6 vote-getters are  
seated) and i can tell you first hand that nearly *everyone* that is  
politically savvy at all bullet vote for one or maybe two candidates.   
and the vote totals (that are about 1/4 of the maximum) bear that  
out.  we actually *worry* about if voting for someone we would rather  
see elected over the jerks in the other party will actually harm our  
favorite candidate.  i haven't met a single person who told me they  
used all 6 of their votes.

with Score and Approval, it's easy to mark your favorite candidate  
(Score=99 or Approval=1).  and we know that Satan gets Score=0 or  
Approval=0.  then what do you do with other candidates that you might  
think are better than Satan?  that question has never been answered by  
Clay.  and any answer must be of a strategic nature.

the ranked-order ballot does not have such a strategic burden.  if the  
tabulation method is fair and equally-weighted (one-person-one-vote),  
then no one can argue that anyone other than the Condorcet winner is  
more preferred than the CW.  there is no other candidate that a  
majority of voters would prefer to see in office.  now, i am not yet  
suggesting that Condorcet (or something like it) be used for multi- 
winner elections (maybe STV would work), but i *am* saying that  
marking the ranked-order ballot tabulated "fairly" seldom requires  
strategic thinking from the voter.  if the voter prefers Candidate A  
more than Candidate B (sufficiently that if a traditional "simple  
majority" election were between only those two, this voter would vote  
for A), then all the voter needs do is rank A higher than B.  as far  
as Candidates A and B are concerned, no other information is assumed  
from this voter nor need be extracted from this voter.

the "simple majority" ballot accepts too little information from the  
voter and the Range ballot requires too much.  and although the  
information in an Approval ballot is less than the ranked-choice  
ballot, it is too little and causes the voter to think strategically  
in how (or whether) he/she will express his/her electoral position in  
a particular race.


--

r b-j                  rbj at audioimagination.com

"Imagination is more important than knowledge."