[EM] Remember Toby

[Kevin Venzke]

Hi Robert,

--- En date de : Sam 28.5.11, robert bristow-johnson <rbj at audioimagination.com> a écrit :
> will minimax of margins decide differently than ranked
> pairs?  if the cycle has only three candidates, it
> seems to me that it must be equivalent to ranked pairs.

It is the same with three.
 
> is there any good reason to use minimax of winning votes
> (clipped at zero) over minimax using margins?  it seems
> to me that a candidate pairing where Candidate A just
> squeaks by Candidate B, but where a lotta people vote should
> have less weight than a pairing where one candidate creams
> the other, but fewer voters weighed in on it.

Margins is basically what Peter originally suggested and what I was
trying to advise him away from.

Margins on average is closer to IRV in results, WV closer to Bucklin.
Though both are closer to each other, of course.

You say you find it more obvious to drop a close contest, but it's only
the winning side of that contest that's going to feel the outcome was
spoiled if they get overruled. The margins idea of "what looks right"
doesn't directly serve any purpose, yet by definition vetoes more voters'
opinions than WV does, making more people wish they had just voted FPP
style, or making candidates wish they hadn't entered the race.

Margins elects A here:
35 A>B
25 B
40 C

Is this going to be defensible when this method is proposed? Can you
argue a case for A without mindreading off of the blank areas of the 
ballots?

I don't think the tightest race is the one to drop. That could be the
only race people thought mattered.

Can you imagine if there were a very tight election between candidates
"B" and "G" let's call them, but because there was a third candidate in
the race we may pick the *loser* of the B-G contest?

I.e. the voters give you a single majority decision (more than half the
voters) and that's the one you don't respect?

Kevin