[EM] Truncation

[matt at tidalwave.net]

On 18 Sep 2002 at 18:30, Dave Ketchum wrote:

> I do offer three levels under Condorcet that could be worth some thought. 
>   I rank:
>       Those I like, just as has been.  Count each of these toward winning 
> in their pairs against all except those above - nothing new.
>       NOTB to represent those I choose not to rank, treating these as has 
> been done.
>       The lemons, with the rottenest one last.  Count each of these toward 
> losing in their pairs against all except those below - but do it as a 
> negative count as to how many voters approve them.
> 
> To clarify:
>       2 A ... NOTB ... Z
>       some who leave both A and Z unranked
>       1 Z ... NOTB ... A
>       Gives A>Z net of 1 (2-1)
>       Gives Z>A net of -1 (1-2)
> 
> This much ranking seems doable to me - I can identify those I like, those 
> I DISlike, and happily leave the rest in the middle.
> 
> As to 1-man-one-vote I claim no violation - each voter gets the same 
> opportunity.

It seems to me to be the same violation that occurs with truncation when ballots are 
not completed.  The more candidates left off uncompleted ballots the smaller the 
voter's voice in deciding the outcome.  Please keep in mind Bert Ingles 
demonstration that Adam Tarr was wrong about truncation never negatively 
effecting the outcome for the voter that truncates.

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em