IRV mailing party adventures
mcnh2 at cam.ac.uk
Fri Apr 6 08:28:52 PDT 2001
Bart Ingles wrote:
> That's pretty close to Merrill's published results. Why not try running
> the same trial, but with the distribution of candidates having a
> standard deviation of about 1/2 that of the voters' S.D.
Haven't seen those... got a URL or the name of the paper?
st. dev of 1 gave <1%
st. dev. of 0.5 gave 3%
st. dev. of 0.25 gave 6%
st. dev of 0.1 gave 43%
For what it's worth... :-)
More information about the Election-Methods