[EM] Simulations

[Forest Simmons]

On Fri, 1 Jun 2001, Richard Moore wrote:

<snip>

> 
> Or (very easy method) do it by random ballot, a poll
> with a sample size of 1. The sim shows that this
> method has a pretty small amount of top bias, which
> is probably just about the effect that's needed.
> 
> 

It would depend on the type of election. In a professional society the
random factor should be less. In public election where big money and
personalities prevail above the issues, the sample size of one would
probably do.

Another suggestion for someone that wants to do the thankless task of a
simulation similar to Richard's:

Assume that voters want their candidates to be clear on all of the issues
that are important to them, i.e. their ideal candidates are at the corners
of the square.

In the simulation have random numbers of voters whose ideal candidates are
at the respective corners of the policy space square. 

Position the real candidates randomly throughout the square, but with some
bias towards the corners supported by the greatest number of voters.

Etc.

Forest