[EM] Depicting PR-STV in an issue space

[Richard Lung]

Users of STV usually say that for cases of a single seat, the next best 
thing is the Alternative Vote (AV; IRV) but that does not mean that the 
two systems are the same in principle. AV/IRV is an elimination count, 
over which the voters have no direct control, and the mechanisms of 
which may frustrate the voters order of choice. With conventional STV, 
this "last past the post" feature of STV becomes less and less important 
with more seats, giving a more proportional count, so that most voters 
elect their first preferences, and then mostly their highest preferences.

A random STV election should take the form of a normal distribution. It 
is analogous to giving local communities numbers of seats in proportion 
to their size. Thus an isolated rural area may only gain one seat, at 
one end of the distribution (next to the origin of the graph) whereas, 
at one time, in the UK, Leeds would have got perhaps 9 seats.

Arithmetically challenged Winston Churchill made the best electoral 
reform aphorism: I would rather be one-fifth of the Members for the 
whole of Leeds than one Member for a fifth of Leeds. (Mentioned by Joe 
Rogaly in Parliament for the People.)

Regards,

Richard Lung.


On 19/04/2024 00:30, Kevin Venzke wrote:
> Hello,
>
> Here's a new feature that tries to show graphically the outcomes of random STV
> elections. There are always 10 candidates, but you can control the number of voters,
> the plot height (vs. the width), how far down voters' sincere rankings they rank,
> and how many seats will be filled. (So, you can reduce the seats to one to see IRV
> if you wanted.) There are a few voter distribution options.
>
> The ranking threshold and number of seats to fill are dynamic controls, so you can
> change them without rerolling everything.
>
> votingmethods.net/stvplot
>
> One interesting phenomenon is how truncation can change the apparent Condorcet
> winner or Smith set. There are some notations for this.
>
> My original interest was the three-seat case with a plot close to 1D, and the
> likelihood that this would carve out a clear center. This could feed a second round
> of voting with three finalists. Some notations specific to three seats remain, but I
> guessed there could be more interest in STV in general.
>
> Kevin
> votingmethods.net
> ----
> Election-Methods mailing list - see https://electorama.com/em for list info