First Choice Criterion

[Mike Ositoff]

> 
> This is a follow up to a discussion with Mike about his new
> First Choice Criterion (strong form).  
> I address the remainder to Mike.
> 
> So far, I think it means something like this,
> 
> It should never be necessary to vote a less-liked alternative equal
> to or over one's favorite in order to ensure the sincere CW wins.
> 
> But what's missing is the "necessary for WHO" part.  That is, who
> has to be co-operating to ensure the CW win.  For example, "all
> voters" or "any majority of voters".

As I was saying in another reply, I meant that _no one_ should
have to do that. But I think I know what you're getting at:
With Margins, in a 3 candidate race, if the middle voters
are willing to reverse their lower choices, then no one will
have to vote anything less-liked over their favorite.

Maybe that criterion's wording is lacking, then. Mayve
the majority wording I discussed would be better.

But unless the middle voters somehow know from which direction
the order-reversal will be attempted (an offensive strategy
involving reversal of a preference ordering is what can
make Margins fail that criterion), then they can't do anything
to prevent it. By deterring one side, they make it really easy
for the other side. For that reason, I don't believe Middle's
defensive strategy in Margins, against order-reversal can
be said to protect the CW, since it royally sets the CW up
for defeat by order-reversal. Therefore, the only way to
really protect the CW is for at least 1 of the extreme sides
to vote the CW over their favorite.

So maybe that criterion is useful as written.

Mike



> 
> Also, could you provide an example of ballots where VA Condorcet
> passes, but Marginal Condorcet fails this criterion.
> 
> ---
> 
> Blake
> 
> 
> 
> -----== Sent via Deja News, The Discussion Network ==-----
> http://www.dejanews.com/  Easy access to 50,000+ discussion forums
> 
>