<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
<title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Warren Smith posted (24 Aug 2010) a link to page discussing a simple
IRV<br>
election:<br>
<br>
18: A>B>C<br>
24: B>C>A<br>
15: C>A>B<br>
<br>
Quoting from the page:<br>
<br>
<blockquote type="cite">FAILURE OF THE SNIFF TEST:
First of all, without any analysis at all, who do you think ought to
win this election?
<br>
It sure looks to me like B is the "most correct" and "most democratic"
winner. But IRV elects A.
<br>
</blockquote>
<br>
That verdict might be justified on positional grounds: B has both the
most first preferences and the most<br>
second preferences and so looks the prettier winner.<br>
<br>
But the objection is mostly based on ranking information which the IRV
voters were content to give because IRV meets<br>
Later-no-Harm. If the A and B supporters are mostly concerned to
elect their favourites ( perhaps encouraged<br>
by accurate pre-election first-preference polling) then with a method
that fails Later-no-Harm and meets Later-no-Help<br>
(such as Bucklin or Range / Score) the cast ballots would more likely
look like:<br>
<br>
18: A<br>
24: B<br>
15: C>A<br>
<br>
Does A now look like the wrong winner?<br>
<br>
<blockquote type="cite"><small><a
href="http://rangevoting.org/RangeVoting.html">Score voting</a>
considers this election an easy call. It would elect B if all voters
gave score X to their first choice, Y to their
second, <br>
and Z to their third, for <i>any</i> X≥Y≥Z, not all equal.
</small><br>
</blockquote>
<br>
Really?<br>
<br>
18: A9, B1, C0<br>
24: B9, C1, A0<br>
15: C9, A8, B0<br>
<br>
A wins. Doesn't this example qualify?<br>
<br>
<br>
Chris Benham<br>
<br>
<br>
<br>
<pre>
<a href="http://rangevoting.org/CompleteIdioticIRV.html">http://rangevoting.org/CompleteIdioticIRV.html</a>
</pre>
<br>
</body>
</html>