# [EM] (no subject)

Forest Simmons fsimmons at pcc.edu
Mon Jul 21 12:55:02 PDT 2003

```The more I think about it the more I like the class of methods that
convert sets of CR ballots into sets of approval ballots.

If the conversion is according to some reasonable procedure, then there is
little incentive for the voters to second guess the method.  In any case
the over-wise voter can rate all candidates at the extremes of the CR,
thereby determining his own final approval cutoff.

Here's an example of a method of this kind:

(1) Use the CR ballots to calculate the pairwise Smith set.

(2) On each ballot use the average of the Smith set ratings (on that
ballot) as the approval cutoff.

If one or more candidates on the ballot fall precisely on this cutoff
value (for example, when the Smith set has only one member), then nudge
the cutoff slightly towards the center of the CR range.

[Be sure to use a CR ballot with an even number of levels so that the
center will not be a possible CR value for a candidate.]

This method would work well with a four level CR ballot, for example.

Forest

```