[EM] Consdervation of information

Richard Lung voting at ukscientists.com
Mon Jan 23 12:49:05 PST 2023



Forest Simmons recently reaffirmed that policy confined them to forms of 
elimination count, in IRV. As he said before, they are all 
non-monotonic. The fact that colleagues are persisting in this course, 
however, implies that this is not a fatal objection, to them.

I believe that it is only an issue, if it all, on the basis of a flawed 
assumption that elections are like an axiomatic system, not allowing any 
untoward consequences, such as a Riker example, and many others.

>From the statistical viewpoint, non-monotonicity might be just “noise” 
in a system. In an attempt to disprove this, one team resorted to the 
case of NASA using the (traditional) single transferable vote (the best 
available system) to estimate the collective view of their engineers, on 
best trajectories for a satellite launch.This is hardly a convincing 
instance of supposedly “chaotic” consequences of non-monotonicity in STV 
elections of human candidates.

Well over a century of STV usage demonstrates, all things being equal, 
that the preference vote reliably corresponds to the proportional count. 
The more proportional the count, in larger constituencies, the greater 
the proportion of first preferences elected, together with a relatively 
few higher preferences.Negligible evidence of chaotic results with STV 
has been provided.

“The possibility of later harm,” alleged of parents, is an argument used 
by social services, in Britain, to excuse forced adoptions. It cannot be 
disproved. It is a superstition and persecution, incapable of scientific 
disproof, unworthy of a modern knowledge-based society. Likewise, a 
similar unproven and unprovable excuse against the use of STV, made, by 
a career party, in the Plant report, citing Riker, without 
demonstration, that stv is "chaotic."

Elections have been derided precisely because they are only “a 
statistic,” as if the public are an uninformed rabble, unworthy of 
consultation, except perhaps to humor them.But elections are not an 
axiomatic system, with a predetermined result to every election.

Thus, I submit that failure to conserve preferential vote information, 
or indeed proportional count information (failure to conserve individual 
and collective information) is a better criterion of representation, 
than non-monotonicity, in first past the post election counts and last 
past the post elimination counts.

I invented a binomial stv system conserving all the preferences in the 
count (over a 4 or 5 seat minimal proportionality).

  Regards,

Richard Lung.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20230123/ee30fe55/attachment-0001.htm>


More information about the Election-Methods mailing list