[EM] Brian--spatial simulations

[MIKE OSSIPOFF]

  Brian--


  First, thanks for the simulation, and for the offer to include in the 
simulation methods sent in by others. It's a good thing that that work is 
being done.

  Samuel Merrill described his spatiatial silmulations in _Making 
Multicandidate Elections More Democratic_, and in at least one subsequent 
book. But ongoing simulations are important to compare new methods.

  Let me disagree with your statement that an ounce of data is worth a pound 
of theorizing. The criterion compliances and other 
theoretically-demonstrable facts and proporties have solid validity. The 
simulations add more, but they aren't more valid than the theory. No need to 
compare the importance of the different kinds of facts, when we can have 
both.

  What kind of distance are you using? Pythagorean (also called Euclidean) 
distance, or city-block distance. I told, on this list, why I claim that 
city-block distance is more meaningful for spatial simulations.

Correct me if I'm mistaken, but it seems to me that we use Pythagorean 
distance when the dimensions represent quantities that don't have a 
meaningful sum, such as mutually perpendicular spatial directions. Or 
different people's shoe-size, etc.

But, as I claimed in my posting some years ago, the dimensionis of issue 
space are, in principle, it seems to me, numerically comparable, and 
therefore meaningfully summable.

Say the dimensions are economics, individual rights, and militarism. Well, 
courts try to put a cost on a person's death or injury, which is part of the 
disadvantage of militarism. Mililtary actions cost money too, making it 
comparable to economics. Like deaths, violations of individual rights are 
quantified by courts.

Or we could approach it from the opposite direction, and maybe that would be 
better. Say the goal is a long and happy life. Money, military service, and 
loss of individual rights affect that goal in ways that could, in principle, 
be quantified.

That's why I claim that city-block distance is the most meaningful kind of 
distance for spatial simulations in issue-space.

  As for methods to simulate, there are different kinds of Condorcet. Most 
prefer the winning-votes kind of Condorcet. As for methods, BeatpathWinner 
is probably the most popular, and probably the best.

  Have you simulated the FBS-complying methods that Kevin proposed toward 
the end of my time on the list before I quit last time, a few years ago?

Mike Ossipoff

_________________________________________________________________
Share your latest news with your friends with the Windows Live Spaces 
friends module. 
http://clk.atdmt.com/MSN/go/msnnkwsp0070000001msn/direct/01/?href=http://spaces.live.com/spacesapi.aspx?wx_action=create&wx_url=/friends.aspx&mk