[Election-Methods] gaming dsv (rob brown still confused)

CLAY SHENTRUP clay at electopia.org
Wed Jan 2 17:26:05 PST 2008


>> 2. your rather confused belief that dsv shouldn't leave voters with
>> any incentive to strategize.
>If it does, than it is not actually dsv. By definition, a dsv agent is
>working toward your interest and therefore it makes no sense to lie to it.

wow. you just don't listen. here we go again.

if there's a condorcet winner, then dsv will indeed act in your best
interest. but the problem is when there's a cycle, in which case dsv
will cycle infinitely, trying to improve each faction's result.
eventually, there has to be a cycle resolution system. so say x/y/z
are going in a cycle, and x is declared the winner of that resolution
process.

now what does dsv do for the majority who prefers y to x? nothing!
it says, "sorry, i have to stop cycling and pick a winner."

that's when they wish they had lied and said that z was their
favorite, preventing the cycle, and getting their second choice
instead of their third.

bottom line: dsv = condorcet, and can be gamed. your next step, after
finally accepting this reality, is to understand why your subsequent
objection (that it is infeasible to game condorcet) is also wrong.

but let's go one step at a time.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20080102/69451e6d/attachment-0002.htm>


More information about the Election-Methods mailing list