<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html;charset=ISO-8859-1">
<title></title>
</head>
<body>
<div class="moz-text-html" lang="x-western"> Marcus,<br>
You wrote (Wed.Jun.16):<br>
<pre>Dear Chris Benham,
you wrote (15 June 2004):
><i> According to Mike, it meets his "Weak Defensive Strategy
</i>><i> Criterion" (WDSC): If a majority prefers one particular
</i>><i> candidate to another, then they should have a way of
</i>><i> voting that will ensure that the other cannot win,
</i>><i> without any member of that majority reversing a
</i>><i> preference for one candidate over another.
</i>
You wrote (16 June 2004):
><i> Thanks for your interest and good example. Maybe somewhat
</i>><i> confusingly, Steve Eppley gives two versions of this
</i>><i> criterion (same name and author, but slightly different
</i>><i> definitions).
</i>><i>
</i>><i> Here is the other one:
</i>><i> > non-drastic defense: If more than half of the voters
</i>><i> > prefer alternative y over alternative x, then that
</i>><i> > majority must have some way of voting that ensures x
</i>><i> > will not be elected and does not require any of them
</i>><i> > to rank y over any more-preferred alternatives. (This
</i>><i> > is promoted by Mike Ossipoff under the name Weak
</i>><i> > Defensive Strategy Criterion. Non-satisfaction means
</i>><i> > some members of the majority may need to misrepresent
</i>><i> > their preferences by voting a compromise alternative
</i>><i> > over favored alternatives if they want to ensure the
</i>><i> > defeat of less-preferred alternatives.)
</i>><i>
</i>><i> In your example, y is A and x is E. The top three groups
</i>><i> of voters, who all ranked A equal first with two other
</i>><i> candidates, can ensure that E is not elected by ranking A
</i>><i> alone in first place. There is no candidate that they
</i>><i> prefer to A, so ER-IRV(fractional) seems to meet this
</i>><i> version (as Mike Ossipoff in effect claimed).
</i>What do you think about this example?:
10 B>C>A>E>...
10 B>D>A>E>...
10 C>B>A>E>...
10 C>D>A>E>...
10 D>B>A>E>...
10 D>C>A>E>...
7 B>E>...
7 C>E>...
7 D>E>...
38 E>...
A majority of the voters strictly prefers candidate A
to candidate E. In my opinion, this example demonstrates
that ER-IRV(fractional) violates Mike Ossipoff's WDSC and
Steve Eppley's "non-drastic defense" criterion. What do
you think?
Markus Schulze
CB:I agree with you. With the votes as they are above, A has no first preferences and so is the
first to be eliminated. The best that the majority who prefer A to E can do for A without ranking
A above any candidate they prefer (to A)is for sixty of them (the top six groups of ten in your
example) to rank A equal-first with two other candidates.
This will result in these first-preference tallies: A20, BCD each 20.3333, E38.
A still has the lowest tally, and so is eliminated.
Chris Benham
</pre>
<br>
</div>
</body>
</html>