[EM] Re: Empirical data on cycles

[Rob LeGrand]

Jobst wrote:
> From the following natural example, it seems to me DMC and wv are
> more truncation resistant than margins:
>
> Sincere:
>   46 A>C>>B
>   05 C
>   49 B>C>>A
> C is Condorcet Winner.
>
> Truncation:
>   46 A>C>>B
>   05 C
>   49 B
> C still wins in DMC and wv, but B wins in margins.
> What do you think?

I agree that winning-votes is more truncation-resistant with
information than margins; my understanding is that it was
explicitly designed to be so.  But truncation is nothing more than
voting tied ranks near the bottom of the ballot, which is usually
an ineffective offensive strategy under margins or winning-votes.
Under either, if I were the C-first faction, I'd vote

 4:C>A>B
 1:C>B>A

to protect C's position as much as possible, making margins just as
truncation-resistant as winning-votes with information.

Apparently we all agree that rational zero-info margins strategy is
sincerity while rational zero-info winning-votes strategy usually
prescribes equal ranks near the top of the ballot.  But even with
perfect information, it seems to me that the equilibria that are
found when each faction protects from others' strategy as much as
possible will be the same whether under margins or winning-votes.
I think it would be a mistake to insist on winning-votes if it
turns out that margins is an easier sell for public elections.

Kevin, could you post an example that gives a favorite-betrayal
incentive under margins but not winning-votes?

--
Rob LeGrand, psephologist
rob at approvalvoting.org
Citizens for Approval Voting
http://www.approvalvoting.org/


	
		
______________________________________________________
Click here to donate to the Hurricane Katrina relief effort.
http://store.yahoo.com/redcross-donate3/