[EM] A new criterion: The Co-operation/Defection Criterion

[Jameson Quinn]

First: please do not make minor changes in the subject line, such as
deleting the article "a". It breaks the thread in threaded email viewers,
making it harder to respond.

Here's my response. I'd written:

That looks right. In fact, let's make it more extreme:
>
> 39 C
> 10 C>A
> 21 A>B
> 30 B
>

To clarify:

These ballots are consistent with your criterion, if the sincere preferences
are
39 (C)
10 (C>A)
21 (A>B)
30 (B>A)
Thus, your criterion mandates that A wins with the ballots as first
described (with B rather than B>A). I am not convinced that A is always the
correct winner with these preferences. It would be very easy to assign
utilities to these preferences such that A is a distant third. I do not
think that it is a good idea to make hard criteria for borderline situations
like this.

These ballots could also plausibly arise from the other sincere preferences
I noted:
21 (A>C)
10 (C>A)
39 (C>B)
30 (B>C)
In which case it is very obvious that A is nowhere close to being the
correct winner. This is a separate argument for why those ballots should not
lead A to win.

Is this clear enough for you now?

Jameson
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20111026/6cc74abc/attachment-0003.htm>