On 12/17/05, <b class="gmail_sendername">Simmons, Forest</b> <<a href="mailto:simmonfo@up.edu" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">simmonfo@up.edu
</a>> 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;">
Rob, in my experience typically when there is a Condorcet Cycle there is no Approval Strategy A style equilibrium of the kind you posit in your second message on this topic.</blockquote><div><br>It had been under the understanding that there will always be at least one equilibrium, but that iterative limit-seeking programs don't always find it. Maybe I misinterpreted something someone else said. I haven't yet written a program to test it.
<br><br></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">However, the method you propose in your earlier message might indeed approach an approval strategy A configuration as a limit without this limit configuration itself being an equilibrium configuration.
</blockquote><div><br>If it does converge on a limit that is not an actual nash equilibrium, I assume it would be creating Range ballots in the end, not having converged all the ballots to be Approval ballots. That is not ideal, but not so bad either. The problem I have with Range ballots is they are not stratigically optimal in the last round of voting (from the point of view of each voter), and having the "agent" vote in a way that is not strategically optimal for the voter it represents makes this a bit less pure.
</div><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">This is analogous to the fact that a function with a jump discontinuity will approach a different value then the value of the function. *
</blockquote><div><br>Got it. Your example, of course, converges toward a "jump discontinuity" (what I was calling a "cusp"). Now certainly there will be conditions like this if there are true ties, which should be less and less common as the number of voters gets larger. If these cases happen that rarely, I don't see it as a problem. However, if they are relatively common (like Condorcet cycles), then this technique is nowhere near as interesting.
<br></div><br>I have no way of knowing which is the case, without writing something to test it.<br><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Anyway, I think your first approach is worth exploring further. I'm sure there is something valuable there to be learned.
</blockquote><div><br>Thanks....<br>-rob</div></div><br>