[EM] "Implied ranked choice" method

Rob Brown rob at karmatics.com
Tue Aug 31 13:50:21 PDT 2004

One of the biggest problems with ranked choice voting (whether it be tabulated 
by a condorcet method or IRV or whatever) is that ballots can be rather 
complex and it would presumably be expensive to implement as well as to 
educate people on how to vote  (true, interfaces like this one I did might 
help! :) http://karmatics.com/voting/ ). Another problem is that voters would 
need to be familiar with a larger selection of candidates than they do today, 
since there would be less need to filter them out via primaries.

Here is an alternative I was thinking about:

Instead of the voters having to rank candidates, they just vote like they do 
today, for a single candidate.  But each candidate would provide the list of 
second choice candidates.  So if I vote for Nader, I am automatically voting 
for a ranked choice list, with Nader at the top, but which is Nader's idea of 
who else would be the best alternative choices.

Even if Nader's only strategy is to get himself in office, he would be wise to 
pick the list in a way that would seem reasonable to his potential voters.  
His list would be public, and presumably finalized at least a few weeks before 
the election (so the public and media have time to analyze things).

I am thinking this would have most, if not all, of the advantages of having 
the voters themselves rank the candidates.  Since it would be tabulated using 
one of the systems for tabulating ranked choice methods -- hopefully a 
condorcet method -- it would be mostly immune from vote splitting and 
strategic voting and such.  It would generally be in people's interest to vote 
for their favorite candidate, even if that candidate doesn't have a great 
chance of winning.

There are interesting side effects this would probably have (candidates 
cooperating in heretofore unseen ways comes to mind).  It might even reduce 
the chances of cyclical ambiguities vs. "voter ranked" methods, but I am not 
sure about that one.

Has anyone ever proposed such a thing?


More information about the Election-Methods mailing list