[EM] Stephane's poll--let's go with it.

Adam Tarr atarr at purdue.edu
Sat Apr 13 00:02:41 PDT 2002


Mike wrote:

>If there are fewer than 5 candidates, or fewer than 5 voters, or
>if someone offers to count the rankings by computer, then I vote
>for all of the methods that I've nominated except Bucklin.
>
>Otherwise, I vote for Approval and Bucklin.

I'm not sure what we are voting on here.  Are we voting on what we consider 
the best method, in theory, for public elections?  Or the best method for 
_any_ elections?  Or the best method to advocate?  These are very different 
questions in my mind.

If we're talking about best for public elections in theory, then I'd 
vote  Beatpath(wv) and Cloneproof SSD in tied first place, followed by 
RP(wv), followed by Approval-Competed Condorcet, followed by PC(wv), 
followed by Approval, followed by Beatpath(margins), RP(margins), Bucklin, 
IRV, lone-mark plurality, and Borda.  My approval cutoff would be at approval.

But if we're talking about what I'd actually advocate as a change, I'd 
suggest Approval, followed by Approval-Competed Condorcet, followed by 
RP(wv), followed by Cloneproof SSD, followed by Beatpath(wv), followed by 
PC(wv), and I would vote everything else in tied last place (i.e. not worth 
campaigning for).

Really, this just amounts to a reshuffling of my approved voting schemes, 
but it shows how dependent my answer is on the question being posed.


-Adam



More information about the Election-Methods mailing list