[EM] another CR/Approval method

[Forest Simmons]

This method can also be called "raising the bar" analogous to the way high
jump and pole vaulting contests are carried out in track meets.

The "ante" is the bar level.  As the bar gets higher and higher, the
weaker candidates start to drop out.  Once a candidate misses, the
possibility for recovery still exists, but it is not likely.


On Thu, 2 Oct 2003, Forest Simmons wrote:

> Here's yet another approach that uses the same basic idea as Gradual Info
> Approval.  I call it "up the ante approval" or "Auntie" for short:
>
> Initialize the "approval ante" as 0.
> Mark all candidates as "viable" initially.
> Initialize all candidates' approval at 1 (or any other value greater than
> zero).
>
> While more than one candidate is marked "viable" ...
>
>    On each ballot calculate the new approval cutoff as a weighted average
> of the viable candidates' CR values, where the weights are the most recent
> respective candidate approvals decreased by the most recent "approval
> ante".
>
>    Calculate all candidates' new approvals based on the new approval
> cutoffs.
>
>    Up the ante, i.e. increment the approval ante.
>
>    Erase the old viability marks, and mark as viable all candidates with
> approval greater than the new approval ante.
>
> End While
>
> The winner is the candidate with the highest approval value in the last
> approval calculation inside the while loop.
>
> Note that in this method as the approval ante begins to overtake a
> candidate's approval, that candidate weighs less and less in the cutoff
> calculation (unless the other candidates are being over taken at a faster
> rate).
>
> To be continued ...
>
> Forest
>
>
> ----
> Election-methods mailing list - see http://electorama.com/em for list info
>