The "Turkey" problem (Re: [EM] 2-rank and N-rank Condorcet)

[Adam Tarr]

Rob Lanphier wrote:

>I think that with this system, the proper strategy is to vote for the most 
>favored of the two frontrunners, and all candidates preferred to them 
>(same as a general election with Approval).

As has been pointed out many times, this is not the optimal strategy for 
Approval.  The best strategy is to vote for everyone you like more than the 
frontrunner, and the frontrunner as well if you like him/her more than the 
second place candidate.  So if the higher-polling of the frontrunners is 
your least favorite of the two, you should not just vote for the second 
place candidate any any candidates you like more, but ALSO for candidates 
you like less than the second frontrunner but more than the frontrunner.

If the election is a dead heat between two frontrunners, the best strategy 
is to approve candidates that you like more than the average of the two 
frontrunners.

>  My only concern with not having a cap is that it might lead to "attack 
> of the clones", where a party would run as many candidates as possible, 
> in order to pack the field with party candidates.

Definitely a valid concern.  Probably enough of a concern to go with simply 
a percentage cutoff.  If nobody can make a (say) 25% cutoff in an Approval 
election, then you don't need a Condorcet election; you need new 
candidates!  It's very rare in approval polls taken before elections to see 
zero candidates with that much support.

-Adam