[EM] RE : Re: Range voting, zero-info strategy simulation (Dave)

[Kevin Venzke]

Dave,

--- Dave Ketchum <davek at clarityconnect.com> a écrit :
> >> > D. Zero-info Approval strategy. The voter gives a 10 to every
> >>candidate
> >> > at least as good as the average value of all candidates, and gives a
> >> > 0 to the others.
> >> >
> >>Interpreting, I see, as a common case, a couple tolerable candidates,
> >>more 
> >>lemons, at least a few dozen voters - and thus every voter who is
> playing
> >>
> >>this strategy approving both tolerable candidates without indicating
> >>which 
> >>of them is preferred as best.
> >>
> >>I see such strategy as a lemon, and this research as a lemon if it 
> >>promotes such strategy as it seems to.
> > 
> > I don't understand why you say this, Dave. If for you there are a
> couple
> > of tolerable candidates, and a bunch of bad ones, but you have no idea 
> > which candidates are considered tolerable to the other voters, then it
> > is indeed your best strategy to give the top score to both of the
> > candidates you find tolerable. Although you may prefer to indicate a
> > clear preference of one candidate over another, you must expect that
> > Range will not reward you for doing so.
> 
> I do not see "for you" in the text I was responding to - and will ignore 
> it in this response.

The text you were responding to is considering the strategy of a single
voter, with the other voters' strategies and preferences being unknown.
The single voter could be anyone.

There was no simulation or assumption that *all* voters are using any
particular kind of strategy.

> If all the voters give top score to all the tolerable candidates, then 
> range will have no information to sort out which is best liked - BECAUSE 
> the voters are hiding that information.

You can't show a strategy to be a "lemon" by arguing that you don't 
like the result when people use it. You have to argue that the people 
using it would not like the result.

If it is known that the voters like the *same* candidates, then sure,
there is a better strategy than zero-info Approval. It doesn't involve
using all the rating positions, though.

> A normal election is usually not close enough to a tie for what ONE voter
> does to make a difference.  If, generally, the collection of voters that 
> consider A and B tolerable vote your strategy, A and B can tie; if each 
> who has a clear preference votes it, this should properly affect the
> result.

If these voters are strategic, then under Range the contest between A
and B will be decided by the voters who like A but not B, or who like
B but not A.

As you note, a single voter has hardly any ability to make a 
difference. If he wants the best result, he doesn't want to hand out a
9/10 and a 10/10. He wants to rate them both in the thousands, if only
it were allowed.

Kevin Venzke



	

	
		
___________________________________________________________________________ 
Découvrez une nouvelle façon d'obtenir des réponses à toutes vos questions ! 
Profitez des connaissances, des opinions et des expériences des internautes sur Yahoo! Questions/Réponses 
http://fr.answers.yahoo.com