CR pairwise

[Forest Simmons]

On Wed, 29 Aug 2001, Roy One wrote:

> --- Rob LeGrand wrote:
> > Yes, I know it's tempting to use the information more precisely,
> > and if we could count on voters' being sincere, it would be
> > better... Any time you don't elect a voted Condorcet
> > winner, you create serious strategy problems
> 
> See my posting "Dyadic approval implented as CR" for my suggestion
> for addressing this. See also Richard Moore's comments about weak
> information obscuring strategy in that thread.
> 

This recent CR pairwise is meant to be simpler than Dyadic approval via
CR, otherwise probably inferior.

Compare it with Bucklin.

In Bucklin you collapse certain preferences until some candidate has a
majority. [That's equivalent to choosing the candidate with the greatest
median rank.]

Suppose you collapsed preferences only until some candidate attained beats
all status.  That would be an improvement.

Now suppose that you had Dyadic Ballots and you collapsed the weakest
preferences until you had a beats all winner.  That would be a further
improvement, and is equivalent to my CR pairwise suggestion.

Forest