[EM] multi-method combo

[Abd ul-Rahman Lomax]

At 01:57 AM 6/8/2006, Dave Ketchum wrote:
>I choke on your collection of methods:
>
>Imposing both range and ranked choice demands that the voter consider both
>methods.  It also complicates the ballot and the counting.

Ranked information can be extracted from a Range ballot. If someone 
doesn't want to equate two candidates, they don't give them the same 
range rating.... Simple.

(If a voter considers two candidates equal, within the resolution of 
the Range ballot, it is questionable to base a public election on 
preference between them; indeed, is this not the very reason to go to 
Range, to avoid forcing preference (but ranked ballots that allow 
equal ranking do avoid that).

>No point to having both IRV and Condorcet:
>       Usually they will select the same winner.

The difference is not in the ballot, but only in the analysis. I can 
see a benefit to publishing both analyses. The IRV winner is perhaps 
simpler to explain and, as you've pointed out, usually IRV will 
select the same winner. But, especially in a close election, it would 
not be too hard to explain why a Condorcet failure in an IRV election 
should result in the victory going to the Condorcet winner. Condorcet 
should be the rule, IRV merely a way of explaining the results.

(But actually, Condorcet is *easy* to explain: The winner is the 
candidate who would beat all other candidates in a two-person race, 
based on the rankings on the ballot.)