[EM] Approval, FBC, voter influence

[Richard Moore]

MIKE OSSIPOFF wrote:

> 
> 
> That never occurred to me about Approval & FBC. The A, B, & C factions 
> in your
> example could be in different time-zones, and it could happen as you 
> describe.
> 
> Then the matter of whether voting is simultaneous & the matter of 
> whether voting
> is secret must be included in the definition of a voting system. 
> Approval fails
> FBC, then, if its voting is neither simultaneous or secret.


Given Forest's example, it seems unlikely that any method would hold up
to a criterion like FBC under conditions such as a roll call vote, where
the middle voters can change their ballots to lure the later voters to
a compromise they otherwise wouldn't make. I think it would be best if
the criterion itself specifies that the ballots are simultaneous and
secret, at least in the English definition.

This specification may not be necessary in the mathematical definition,
or rather, it may already be present (though not explicit). For example,
if a mathematical definition is based on what happens when we combine
ballot B with a fixed set S of other voters ballots, since S is a
constant, the voter who casts ballot B cannot influence the ballots
in S. Changing B to B' may change the election's outcome, but it won't
change what the other voters did, if the definition is expressed in
this way.


  -- Richard