[EM] Re: first wave Condorcet proposals

James Green-Armytage jarmyta at antioch-college.edu
Mon Mar 14 23:03:04 PST 2005


I wrote:
>	Assuming no pairwise ties, how does SD differ from SSD? Can you give an
>example where they differ?

Mike, you wrote:
>I can't find an example in time to include it in this reply, and I didn't 
>want to delay this reply while I look for an example. 

	Markus was good enough to dig up such an example, from several years ago.
The example has 7 candidates in the Smith set. It is a good demonstration
of how SD differs from beatpath, and how SD lacks monotonicity, but it is
convoluted enough to be absurdly unlikely to occur in a public election
(given a reasonably small number of candidates). Hence I do not find it to
be very damaging in practical terms, and I can easily imagine a scenario
where the simplicity of SD outweighs the benefit of satisfying
monotonicity, etc. 
	In the long run, of course, it would be better to settle on one of the
"premium" base methods (beatpath, ranked pairs, river...).
...
	In my opinion, SD is just as intuitive and easy to explain as PC, if not
more so ("why bother to drop a defeat that isn't causing a cycle?"), but
you might be right in thinking that the matter should be settled via
public polling.

my best,
James




More information about the Election-Methods mailing list