[Election-Methods] IRV/Approval/Range comparisons on Wikipedia

[Abd ul-Rahman Lomax]

At 11:11 AM 9/27/2007, Andrew Myers wrote:
>Warren Smith wrote:
> > strong candidates that you actually approve of but are not your first
> > choice"
> >
> > --this "strategic" move actually is "strategicaly stupid" i.e.non-optimum,
> > generally speaking.  See   http://rangevoting.org/RVstrat3.html
> > and compare columns B and E for large # of voters.
> >
> > The writer here is advising somebody to be UNstrategic, then 
> complaining that what
> > happens if they obey his advice, is the horrible consequence of 
> strategic voting!
> >
> > Warren D Smith
> > http://rangevoting.org
> >
>The strategy proposed is essentially truncation. To be convinced that it
>doesn't work, I'd want to see a simulation with a more realistic model
>of voter behavior. The reported simulation has voters choosing their
>preferences uniformly at random.

No, actually. It generates preferences at random, that particular 
simulation, but the simulated voters then vote various strategies 
based on those preferences. It is just one approach. Warren's IEVS 
simulator includes provision for many different models of voter 
preferences, that was just a simple one, if I'm correct.

At this point, I'd suggest that the burden is on you to show that the 
strategy *works*. With one particular distribution of preferences in 
the electorate (and such a distribution is, in fact, a reasonable 
first approximation, though it could get much better), it does *not* 
work. Is there an example of a reasonable distribution where it *does* work?

Or is it just a hunch?