Circular Tie Percentages???
dfb at bbs.cruzio.com
Sat Nov 16 22:40:43 PST 1996
Regarding the question about how frequent circular ties would be:
If voters & candidates can be considered to be positioned on
a 1-dimensional political spectrum, then there would always
be a Condorcet winner. But truncation could still create
a circular tie. Though Condorcet's method doesn't reward
truncation, truncation would still surely take place on a
large scale in a public elections, just as it has in all
the rank-balloting elections I've participated in. But I
don't know if people would refuse to rank a likely Condorcet
winner just because they felt their candidates could beat the
ones they like less than the CW, since there's no reward in
it, and there could be a penalty if their candidates turned
out to _not_ beat the ones they like less than the CW.
But then it's sometimes hard to judge who's the CW, and,
combining that with the tendency to vote short rankings, it
might well be not so unusual for truncation to cause circular
ties when there's a CW.
But when there are several issue-dimenstions, instead of just
1 political spectrum, there's less & less likely to be a Condorcet
winner, which means that it no longer would take truncation to
make a circular tie: There's often (maybe usually) be natural
circular ties. I emphasize, however, that the most important
situations are the ones when there's a CW. And it seems to me
that, even then, there could be non-naturall circular ties,
due to truncation and uncertainty about who's the CW.
More information about the Election-Methods