Tie-breakers

[Steve Eppley]

Bruce A wrote:

>>>> Bruce, would you care to suggest the definition(s) appropriate 
>>>> to the Electoral College reform issue, so that we could discuss 
>>>> "voting methods" of the form EC//M (where M is some single-
>>>> winner method).

>>> if there are no candidates that are capable of winning in 
>>> this manner, then EC should select all of the candidates
>>> running in the election as its winners.

>> But that complicates the definition of the function, and for 
>> what gain?  The electoral college is deadlocked and moot no 
>> matter how the reformed states award their delegates.  So 
>> why do you say "should"?

>In my "definitions document" a Single-winner Ranked-ballot Voting
>Method must always result in at least one winner.  This is a
>standard definitional criterion, and I do not choose to relax it.
-snip-

This doesn't appear to me to be a good enough reason to change 
the ec reform proposal.  Plenty of other methods also won't meet
this "narrow" definition, such as allowing voters to insert a 
"None of the Rest" choice in their rankings.  I think the onus 
is on academics to model real proposals, rather than on reformers 
to make things simpler for academics. 

This is a small enough addition to the proposal that maybe I 
shouldn't be too concerned with the added complexity.  More
significant is the case where exactly one candidate is capable
of winning nationally: the simple ec reform proposal will elect
this candidate, but the reformed states might be happier to
let the House pick the winner in some circumstances.

---Steve     (Steve Eppley    seppley at alumni.caltech.edu)