[EM] Re: Utilities?

[Adam Tarr]

Jobst Heitzig wrote:

>EXACTLY! Don't worry about cyclic preferences! I certaintly don't since
>I know they are no problem at all for most serios methods we discuss
>here! That's exactly my point! Don't worry what properties the
>individual preferences might have that you might dislike or consider
>silly, since those properties just don't matter at all! Exactly what I'm
>saying all the time :-)

While they don't create a problem for most methods, allowing cyclic 
preferences creates significant problems for most _ballots_.  This is the 
key point.  Is the ability to express cyclic preferences requires a massive 
ballot with every pair being voted on; contrasts that to the "A, B, C, D, 
F" graded ballot that is trivial to vote properly on.

For every voter who really wants to express cyclic preferences, who is 
helped by the pairwise comparison ballot, there will be dozens who submit 
spoiled ballots because they don't understand the more complicated ballot, 
who mark cyclic or mistaken (for them) preferences simply because they lose 
track of things, or who fail to mark preferences they truly have because 
they grow confused by the ballot.

For the record, I agree with James's criticism of the logic of cyclic 
preferences.  I have had this argument on the list a few times - I have not 
yet been convinced that personal cyclic preferences are anything but 
noise.  With enough thought, you can resolve the inconsistency and produce 
a transitive order.

Note that the previous paragraph is NOT the main reason I oppose pairwise 
comparison ballots.  It's just the reason I don't feel bad about it.  The 
main reason I oppose pairwise comparison ballots, and support graded 
ballots, is simplicity for the voter.

-Adam