<div> <font face="Arial, Helvetica, sans-serif">From: Jobst Heitzig <heitzig-j@web.de><br>

> I think we have! The reasoning is this: 55% like A best, 45% like B best.<br>

> Therefore the "democratic benchmark" solution with which we should compare<br>

> prospective solutions is the lottery that elects A with 55% probability and B<br>

> with 45% probability. Now, all voters prefer C to this benchmark, but only 55%<br>

> prefer A to this benchmark and only 45% prefer B to the benchmark. From this<br>

> point of view C is a better solution than A is.<br>

                                                                                                                                                            
<br>

That is interesting.  There has to be some baseline/fallback before negotiation<br>

can begin.  This is normally the status quo or some baseline cultural viewpoint.<br>

                                                                                                                                                            
<br>

The baseline would be random ballot.  However, 2 ballots can be combined into<br>

a single new ballot.  The actual ratings don't matter in the context of picking<br>

the winner, but they do matter for the combination algorithm.<br>

                                                                                                                                                            
<br>

Each vote pair can be replaced by a combined vote.<br>

                                                                                                                                                            
<br>

If I vote 100 - 80 - 0 and so does another, then that can be replaced by<br>

                                                                                                                                                            
<br>

2) 100 - 80 - 0<br>

                                                                                                                                                            
<br>

This represents no increase in utility.  This vote if picked will pick A as the<br>

winner.  This is the same as 2 separate votes with half probability.<br>

                                                                                                                                                            
<br>

It gets a little more complex when combining votes which are different.<br>

                                                                                                                                                            
<br>

Combine (100 - 80 - 0) with (0 - 80 - 100)<br>

                                                                                                                                                            
<br>

Is 0 - 100 - 0 an improvement from the perspective of both votes ?<br>

                                                                                                                                                            
<br>

Assuming no more combinations take place, then this represents an improvement.<br>

                                                                                                                                                            
<br>

(50 - 80 - 50) would also be an improvement in that context as if there is no<br>

more combinations, that that is a C win.<br>

                                                                                                                                                            
<br>

The question comes down to defining the value of a vote while also incorporating<br>

the fact that futher negotiation is possible.<br>

                                                                                                                                                            
<br>

If that can be done, then there could be an auction.  Each vote is combined with<br>

the other vote that maximises the combined value.  Votes remain uncombined if<br>

the combination results in a lowering of value from the perspective of one or<br>

other vote.<br>

                                                                                                                                                            
<br>

Bullet votes would never combine.  The bullet vote would never gain anything by<br>

giving up some chance of its main choice winning as it considers all the others<br>

equally bad.  I wonder if that can be used to help with the combination.<br>

                                                                                                                                                            
<br>

> If Candidates A and B agree to "trade" their power by transferring their complete<br>

> share of the probability to C, both factions will gain.<br>

                                                                                                                                                            
<br>

However, in practice, they won't.  Most candidates would prefer a 45% chance that<br>

they will win over a 100% chance that they won't.  Their own personal utlities<br>

will swamp any difference in factional utilities<br>

</font><br>
</div>

<div> <br>
</div>

<div style="clear: both;">Raphfrk<br>
--------------------<br>
Interesting site<br>
"what if anyone could modify the laws"<br>
<br>
www.wikocracy.com</div>

<div> <br>
</div>
<!-- end of AOLMsgPart_0_33f4826e-e740-4dab-980a-0042475dd18c -->

<div class="AOLPromoFooter">
<hr style="margin-top:10px;" />
<a href="http://o.aolcdn.com/cdn.webmail.aol.com/mailtour/aim/en-us/index.htm" target="_blank"><b>Check Out the new free AIM(R) Mail</b></a> -- Unlimited storage and industry-leading spam and email virus protection.<br />
</div>