[EM] Re: Optional Later-no-harm AER

[Forest Simmons]

On Mon, 5 Jan 2004, Chris Benham wrote:

> Forest,
> I   am interested in any comment you have on this method (that I posted
> Sun.Dec.21, 2003).
> I  didn't steal one of  your ideas, did I ?

No, but I wish I could claim it ... especially your idea of minimal
adjustment of approval cutoffs to keep all possible ballots relevant.

My idea for keeping all ballots relevant was to use dyadic approval
ballots or (equivalently) adjusting (at every stage) the approval cutoff
to the exact halfway mark between the two most extreme CR values still in
play on the CR ballot.

Your idea is simpler, in that it doesn't require CR ballots or dyadic
approval ballots, and better in that it makes the minimal adjustment
necessary to keep all of the ballots relevant, thus more nearly reflecting
the original approvals of the voters.

Forest