[EM] cyclic preferences

[James Green-Armytage]

"Paul Kislanko" <kislanko at airmail.net> writes:
>Your example is not relevant to the question. In an election only between
>Z
>and X, there are no other pairs to worry about.

	Yes, it is relevant. This is one of the more generally accepted
definitions of a preference ranking: in an election between X and Z, if
you would vote for X, then your preference is X>Z; if you would vote for
Z, then your preference is Z>X; if you would vote for both or neither,
then your preference is X=Z.
	This definition isn't perfect, and I don't want to get caught saying that
it is *the* definition of a preference relation, but it's generally useful
for most purposes.
	So, if you want to vote X>Y>Z>X, that seems to suggest that
1. in an election between X and Y, you would vote for X
2. in an election between X and Z, you would vote for Z
3. in an election between Y and Z, you would vote for Y
	In my e-mail, I was trying to argue that Jobst hadn't set forth a
convincing argument for why his voter would want to do this.
	I thought that my line of argument was fairly clear, but apparently it
was not clear to you. Hopefully I have clarified it now.

my best,
James