On 5/27/05, <b class="gmail_sendername">Bishop, Daniel J</b> wrote:<div><span class="gmail_quote"></span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Simmons, Forest wrote:<br>>1. Exactly how do you define correlation?<br><br>My suggestion is this:<br><br>The "absolute Borda difference" (ABD) between two candidates on one<br>ballot is the absolute value of the difference of their Borda scores on
<br>that Ballot.<br><br>The "total absolute Borda difference" (TABD) between two candidates is<br>the sum of their ABDs on all ballots. "Correlation" is the inverse of<br>the TABD.<br><br>(I had intended to send this on Wednesday night, only to find out that
<br><a href="http://neo.tamu.edu">neo.tamu.edu</a> no longer lets me log in as "dbishop". In the meantime, Ken<br>posted his suggestion. It looks like mine is identical, but simpler.)</blockquote><div><br>
<br>
Agreed. Thank you for the simplified statement. <br>
<br>
<span class="gmail_quote"><br>
<b class="gmail_sendername"><span style="font-weight: bold;"></span>Daniel continues:</b></span><br>
</div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"><br>In this particular example, the Condorcet Winner is elected. However, this is<br>
not always the case.<br><br></blockquote></div><br>
Again, absolutely true. However, it should come as no surprise that I consider <br>
"failing" CC to be a positive. Failing CC stems from obeying symmetry, which is<br>
a Borda property I tried to maintain. <br>
<br>
While your CC failure example is helpful, my favorite is Condorcet's original <br>
critique of Borda: <br>
<br>
30:A>B>C<br>
10:B>C>A<br>
10:C>A>B<br>
1:C>B>A<br>
29:B>A>C<br>
1:A>C>B<br>
<br>
Condorcet picks A & Borda & CIBR pick B. Here's the explanation (summarized<br>
from Saari): If symmetrical ballots, (which represent ties & should cancel), <br>
are factored out, the election outcome should be unchanged. <br>
<br>
The symmetrical ballots in Condorcet's critique are: <br>
10:A>B>C<br>
10:B>C>A<br>
10:C>A>B<br>
and <br>
1:C>B>A<br>
1:B>A>C<br>
1:A>C>B<br>
<br>
The reduced profile is then: <br>
20:A>B>C<br>
28:B>A>C<br>
<br>
All reasonable methods pick B in this election. Since Condorcet picked A from the<br>
original profile, it is not "symmetry proof." <br>
<br>
CIBR is both symmetry & clone proof, which is what is exciting about it. <br>
<br>
-Ken<br>
<br>