[EM] Joe's ballot--I accidentally missed posting it.
LAYTON Craig
Craig.LAYTON at add.nsw.gov.au
Sun May 6 23:14:03 PDT 2001
I've had a more comprehensive look at the pairwise count. There is a
natural order, meaning that I didn't count this using any Condorcet method
in particular. But, as it happens, every numbered method on the following
list pairwise beats every method lower down on the list, is defeated by
everything higher on the list, and ties with every method on the same level
(with one exception, SD pairwise ties with PC). Testing if Condorcet
methods match the following order might be a good test of whether or not
they are adequate ordering methods;
1. Approval
2. Ranked Pairs (m)
= Beat Path (m)
3. Cardinal Ratings
4. Approval Completed Condorcet
5. Cloneproof SSD
6. SSD
= Ranked Pairs (wv)
7. SD
8. Smith//PC
9. PC
10. IRV
11. Plurality
= Borda
-----Original Message-----
From: MIKE OSSIPOFF [mailto:nkklrp at hotmail.com]
Sent: Monday, 7 May 2001 13:30
To: election-methods-list at eskimo.com
Subject: [EM] Joe's ballot--I accidentally missed posting it.
Sorry, I accidentally missed Joe's ballot when I gathered the ballots
for posting and printing-out. That means that I didn't count it either,
since I counted from my printout.
But Joe's ballot further increases the magnitude of Approval's win.
Approval's main rivals were some pairwise-count methods--Cloneproof SSD,
ordinary SSD, RP(wv), and RP(m). Joe's ballot increases the amount
by which Approval beats those methods.
Since Joe didn't designate a method, Manual is the default designation,
and his ballot directly increases the final Approval score of Approval,
but not of the pairwise-count methods. Likewise, that ballot
further helps Approval against the pairwise count methods in the
Approval count and in all the pairwise-counts.
So Approval won even more soundly than it had initially seemed.
It must be obvious that I'd have no motive to intentionally not count
& not post a ballot that counts in favor of Approval against Margins.
Besides, Rob LG had a copy of the ballot too, and so it wouldn't have
been possible to successfully leave it out, even if I'd wanted to.
Lastly, if I were going to intentionally omit a ballot, I wouldn't
state the correct number of ballots (nine).
More information about the Election-Methods
mailing list