[EM] Rankings made by another method

[James Gilmour]

> James Gilmour wrote:
> >Your definition relies on the answer to a question that is 
> asked only 
> >very rarely, and is never asked in a real election.  (It was 
> asked in 
> >the Scottish Social Attitudes Survey, shortly after the
> >two elections to the Scottish Parliament, but that is 
> exceptional.)  In a
> >real election the voters,
> >or at least those who are awake, will always take into account any
> >effects inherent in the
> >arithmetic of the voting system being used for that 
> election.  It was for
> >this reason that I
> >suggested it would not give very useful results to take 
> ballots recorded
> >for one voting system and
> >count them by a different voting system.  It would still not 
> be useful if
> >the ballots were marked
> >sincerely (ie sincerely as determined by the unasked 
> question) because
> >such ballots might well not
> >produce the result the voters wanted.  

> On Behalf Of James Green-Armytage
> Sent: Friday, March 26, 2004 3:32 PM> 
> 	I don't follow your train of reasoning in the last 
> sentence, to be honest.

NB I had hoped my first two sentences above had made it clear that I had accepted your definition of
"sincere voting" or at least, that I had acknowledged the difference between a "sincere vote" and
what a voter would almost certainly do in a real election, according to the voting system in use.

But the sincere vote is something of an abstract concept.  To turn it into an election outcome you
need to apply particular voting system arithmetic.  And different voting systems may (will) give
different outcomes.  How then do you decide which of these different outcomes is the best
representation of the totality of the "sincere votes", given that every outcome is an artefact of
the arithmetic used to produce it?  And what value system do you then use to decide "best"?

Some of the example ballots used in this list are artificially created and are declared to be
"sincere".  But some of the datasets are taken from real elections.  For these, no assumption should
be made about the votes being "sincere".  That surely introduces an additional distorting factor
into the comparison of outcomes when these datasets are counted by voting systems other than the
ones used in the real elections from which the datasets were taken.


> 
> >So you have still discovered nothing useful.  Worse, you 
> have probably 
> >misled yourself.
> 
> 	I haven't the faintest idea what you're talking about 
> when you say that I have misled myself.

MY apologies  -  I should have been much more careful about my use of "you", given what I wrote in
another post recently!  I was not intending to make a comment on anything you yourself had written.
What I should have said was that analysts who take this approach are at risk of misleading
themselves.

<<Big CUT>>

> I now say:
> 	Your equation of sincerity with strategic praxis in a 
> given method is contrary to usage.

As first written above, I had hoped I had made it clear in the previous message that I understood
and accepted this distinction.  Mea culpa!

James Gilmour