[EM] margins axiom

[Steve Eppley]

Hi,

There is a possible compromise between margins and 
majorities (but I prefer straight majorities, not 
this compromise):  

   If the voter leaves two candidates unranked, 
   don't count that as a half vote for each, count it 
   as zero for each.  But if the voter explicitly ranks 
   two candidates as equal, count that as a half vote 
   for each.  

Thus, if more than half the voters rank x but not y, 
then y cannot have a count over any z that is as large as 
the majority for x over y.  This means it's straightforward 
to satisfy the minimal defense criterion; for instance, 
the corresponding variations of MAM and BeatpathWinner 
satisfy minimal defense.

Truncation resistance, as I've defined it in my webpages,
is also satisfied.

--Steve
-------------
James G-A wrote:
> Sorry, I just said that Blake's argument treated the fairness of margins
> as an axiom in itself, but now I realize that he does provide a specific
> justification for it.
> 
> I quote from
> http://lists.topica.com/lists/RankedPairs/read/message.html?mid=1600781094&sort=d&start=15
> 
> Blake wrote:
> >The argument for using margins is fairly straight-forward. If the
> >voters on the winning side are evidence that a proposition is true, than
> >the voters on the losing side are evidence against it. Therefore, it
> >makes sense to consider both when deciding our confidence in the
> >decision. That was Condorcet's approach. View the election in terms of
> >evidence and probability. That is my approach as well. However not
> >everyone agrees.
> 
> 	This is a good point. However, I still think that the strategic issue
> gets more priority. When Condorcet made this probabilistic argument, I
> doubt that he was taking strategic manipulation into account.
> ----