[EM] CR with AERLO

MIKE OSSIPOFF nkklrp at hotmail.com
Thu Apr 22 19:04:01 PDT 2004


AERLO isn't just for rank methods. AERLO is useful for CR too, and maybe for 
the various other rating-ballotng methods that have been proposed.

With AERLO, the voter in CR could rate sincerely, while still fully 
protecting his above-line candidates, as if he'd given them all maximum 
rating. In fact, if the voter also applied the automatic trunation line 
option (ATLO), at the same point, his ballot would optimally help the 
above-line candidates in the event that none of them initially won.

ATLO is triggered in the same way as AERLO, but, when triggered, it drops 
the below-line candidates.

In general, ATLO isn't as useful as AERLO. AERLO is the one that I suggest 
for all ranking & rating balloting methods. But ATLO would often be useful 
too. For instance it could be a useful addition to AERLO in CR & Bucklin. In 
Bucklin it could be useful if the voter wasn't confident that the other 
voters would protect his preferred set (with manual equal-ranking or AERLO), 
and so he could use ATLO to avoid giving any votes below that preferred set 
if no one in his preferred set initially wins, in case the votes that they 
need will be delayed in their arrival.

I've already posted about the possible value and the drawbacks of adding 
ATLO to Condorcet, when AERLO is also used. With ATLO, the supporters of a 
CW could more easily protect him by deterring offensive order-reversal, if 
they aren't confident that others would protect him by manual equal-ranking 
or AERLO. But the trouble is that if they can do that, then others might 
expect them to, and might not protect the CW. It could become a strategy 
game about who will protect the CW from offensive order-reverssal. Certain 
other voters could have their candidate elected if the CW supporters 
truncate and offensive order-reversal is attempted against him.

There could be additional options that would combine AERLO & ATLO, maybe in 
a system that uses three counts instead of just two, and in which ATLO is 
only applied, if needed, after the 2nd count.

Mike Ossipoff

_________________________________________________________________


More information about the Election-Methods mailing list