<html><head></head><body>Markus Schulze wrote:<br>
<blockquote type="cite" cite="mid:3B025326.A81537B1@sol.physik.tu-berlin.de"><pre wrap="">you wrote (15 May 2001):<br></pre>
<blockquote type="cite"><pre wrap="">It may be sufficient for some, but I would think at least a few voters<br>would demand more complete information about an election that will<br>determine many aspects of their lives for the next several years.<br>Exactly where do the numbers come from in each step? If there were<br>100000 A votes in the first round, and A got eliminated, how many were<br>ABC votes and how many were ACB votes? This can be determined by the<br>difference between first and second round votes for B and C. But how<br>many of the B votes were BAC votes and how many of the C votes were<br>CAB votes?<br></pre></blockquote>
<pre wrap=""><!----><br>However, also the pairwise matrix doesn't say how many BAC votes there are.<br>You seem to believe that the voters want to know how many BAC votes there<br>are only when IRV is used but not when a pairwise method is used.<br><br>Markus Schulze<br></pre>
</blockquote>
What the pairwise matrix doesn't tell you is anything about the order of<br>
elimination. You could provide both the elimination results at each stage,<br>
plus the pairwise matrix, and this would tell you a lot. But it doesn't tell<br>
you enough to play out alternate scenarios, such as "If X had received an<br>
extra 0.1% of the vote, and had not been eliminated in the first round, how<br>
would the outcome have changed?"<br>
<br>
As a practical example, any candidate thinking of asking for a recount<br>
to challenge his elimination in a close IRV race might wish to first<br>
determine whether doing so could produce a worse outcome.<br>
<br>
Richard<br>
<blockquote type="cite" cite="mid:3B025326.A81537B1@sol.physik.tu-berlin.de"></blockquote>
</body></html>