[EM] Practical Condorcet

Forest Simmons forest.simmons21 at gmail.com
Tue Mar 7 01:08:52 PST 2023


Remember the "strongest pair" winner  method?

The thing that side-tracked me was trying to guarantee Smith efficiency in
a simple way that would not scare away the innumerate.

I think I may have a solution to that challenge.

Here's my presentation idea:

In the context of single winner deterministic elections, when we have a
candidate that is preferred by a majority of the participating voters over
any competing candidate ... basic majority rule principles suggest electing
that candidate as a matter of policy.

But what about the one-in-a-hundred possibility of non-existence of such a
candidate?

In that case, almost certainly there will be at least one pair of
candidates such that no other candidate defeats both of them.

Such a pair is called a covering pair.

It seems reasonable to elect the pairwise winner of the strongest defeat
covering pair.

One possible measure of defeat strength is winner's pairwise support (wv):

x ABC (sincere is ACB)
y BCA
Z CAB

Assume z<min(x,y)

B beats C with strength x+y=n-z, where n is the total number of ballots
x+y+z.

This is the strongest defeat.

B wins thereby making the A faction burial of C backfire.

There may be a better choice of defeat strength ... we'll have to
experiment.

Note the simplicity of getting ISDA without mentioning Smith!

Suggestions?

-Forest
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20230307/3b14a5bc/attachment.htm>


More information about the Election-Methods mailing list