[EM] Another method idea

[Forest Simmons]

Daniel Bishop wrote:


>Forest Simmons wrote:
>
>> Ballots are ordinal rankings or cardinal ratings.
>>
>> Any candidate with more than average first place rankings or ratings 
>> gets a point.  Any candidate with fewer than average last place (or 
>> truncated) rankings or ratings gets a point or an additional point.
>>
>> The candidate with the most points wins.
>>
>> In case of ties, all but the tied candidates are stricken from the 
>> ballots, and the method is repeated.
>>
>> If the method gets stuck on a set, then the tie is broken by random 
>> ballot restricted to members of the tied set.
>>
>> Call this method "The first and last place point system."
>
>Consider the set of candidates {A, B, C} and ballots
>
>3: A>B>C
>2: B>C>A
>
>If I understand your method correctly, the relevant math is:
>
>First-place votes:
>
>A: 3 (above average)
>B: 2 (above average)
>C: 0 (below average)
>Average: 5/3=1.666...
>
>Last-place votes:
>
>A: 2 (above average)
>B: 0 (below average)
>C: 3 (above average)
>Average: 5/3=1.666...
>
>Points:
>
>A: 1
>B: 2
>C: 0
>
>So B is elected.  But A has an absolute majority of first-choice votes!
>

That's why I suggested the spruced up version!

Since A is the only member of the Dutta minimal covering set, A is the 
winner of the spruced up version.

Forest