[EM] percentage support

[Juho Laatu]

Dear Curt, Daniel and All,

On May 3, 2005, at 02:06, Curt Siffert wrote:

> You cannot derive, from a Condorcet ballot collection, how much 
> percentage support each candidate got.  You can't give each candidate 
> a share of 100% in a way that all candidates would agree on.  If you 
> can, I'd love to know how.
>
> Is this an already identified criteria?  The ability to determine 
> percentage support?  The Siffert Criteria?  :-)  If so, Condorcet 
> fails it; at least, I haven't seen a technique that would allow it to 
> pass it.  What voting methods can convincingly a) identify the total 
> available support (in terms of that vote method) for all candidates, 
> and b) determine what percentage of that support each candidate 
> received ?

I think ability to give clear information that indicates who how 
clearly someone won and how well the other candidates performed is a 
useful requirement. I however think it is not necessary to have 
measures that can be summed up to 100% but any easily understandable 
measurements would do fine. I'd thus propose to widen the definition a 
bit and require only the ability to present some (numeric) values that 
clearly describe the strength of wins, losses and/or final strength of 
each candidate.

This kind of criterion is useful in real life where one may want e.g. 
to show graphics on TV illustrating how well each candidate performed - 
percentages, number of votes or whatever as long as people get a good 
understanding on what happened in the elections. Ability to see what 
really happened in the election may also increase the trust in the used 
election method.

Daniel Bishop's mail gave some interesting measures.

On May 3, 2005, at 05:56, Daniel Bishop wrote:

> Here's a method I tried a few months ago:
>
> Start by determining the number of additional "bullet votes" each 
> candidate would need in order to win or tie.
>
> For example, suppose you polled 7 people for an {A, B, C, D} election, 
> and:
> * D is the winner
> * A would win with the additional ballots 4:A
> * B would tie D with the additional ballots 3:B
> * C would win with the additional ballots 6:C

Counting additional ballots that are needed to become a Condorcet 
winner means the MinMax (margins) method (aka "Least Additional 
Votes"). Daniel seems to refer to additional ballots needed to win in 
general, not ones needed to become a Condorcet winner (but allow me a 
reference to MinMax since that is one of my favourite methods :-) ).

Question to David. Why weren't you happy with the numbers in the 
beginning part of your mail that I referred above? Those measures look 
very logical to me. A's defeat is worth 4 votes, which is a clear 
measure. B loses by 3 votes and C by 6. D's win is maybe worth 3 votes 
since that is the distance to the closest competitor (B).

Best Regards,
Juho