[EM] Burlington Vermont repeals IRV 52% to 48%

Dave Ketchum davek at clarityconnect.com
Wed Mar 3 18:13:16 PST 2010

Actually, a chance to move ahead, if we grab it.

IRV lost, and good riddance.

Let us look at an extreme example - not something to expect to be  
common, but to clear the air:
      100C - 2/3 like C>A; 2/3 like C>B

IRV has A win for this exact count; B or C as likely for near counts.

Plurality would have expense of a runoff with A or B winning for this  
exact count; C could expect to be included in 2/3 of runoffs for near  
counts, and win most of those.

Approval is better than Plurality, and properly awards C for being  
liked by all.

Condorcet is better for various reasons:
      Looks at all that the voters say, unlike IRV.
      Lets them vote for more than one, unlike Plurality.
      Ranking lets them express relative liking, unlike Approval.  The  
ranking is familiar to those having heard of it with IRV - same votes  
are permitted, though equal ranking is also permitted here.
      Unlike IRV, all that any collection of ballots (such as for a  
precinct) says can be summed in an N*N matrix, and collections of such  
matrices (such as for the whole district) can be summed.  Matrix  
content can also be released, showing all concerned how liking of  
various candidates varies.

Dave Ketchum

On Mar 3, 2010, at 1:58 AM, Kristofer Munsterhjelm wrote:
> robert bristow-johnson wrote:
>> Well, that's sad.  Even with a sorta narrow victory the anti-IRVers  
>> will swagger down Church Street like they own the place. We will  
>> now all accept that God instituted the "traditional ballot" for use  
>> forever and that a 40% Plurality is a "winner".
>> It would have been optimum if IRV survived this vote by a narrow  
>> margin.
>> It's sad that when FairVote introduced and promoted the ranked  
>> ballot that, from square 1, they always coupled it to the IRV  
>> tabulation of votes.  When enough disasters (at least anomalies)  
>> happen like in Burlington or Aspen, some backlash, both ignorant  
>> and enlightened, is bound to happen.
> I think that shows that IRV is just not good enough. Of course, I  
> could be wrong: perhaps it is, as you said, an outcome on par with  
> "Bush wins the presidency -- in the supreme court", but if it was  
> IRV itself that gave the opposition enough proverbial ammunition,  
> then that does count against the method.
> Since I prefer Condorcet, I would say that a good method should have  
> elected Montroll. IRV didn't.
> What's really bad is that now people will probably think that the  
> ranked ballot and IRV are one and the same - that the only way to  
> conduct a ranked ballot election is by using IRV. That would, absent  
> favorable development elsewhere (Condorcet efforts that succeed,  
> showing another way is possible).
> What method will be used in Burlington now -- Plurality or runoff?  
> Since you said 40% earlier, I guess it's a runoff, but 40% sounds  
> odd as a runoff threshold. Shouldn't it be majority? Anything less  
> and the voters might have preferred someone else.

More information about the Election-Methods mailing list