[EM] N/P Scales

[David Catchpole]

That is such a faulty version of utility I don't know why I'm
responding! Sorry, no offence really intended, but where's the point and
reason in "zero means neutral (do not care)"? Care about what? The common
utility formalism attaches a comparative value (utility) to each outcome,
according to which an actor acts rationally to optimise her mean
utility- mean utility, because when an outcome of an actor's behaviour is
not certain, the utility of that behaviour is considered to be the
probabilistically weighted mean of outcomes from that behaviour. However,
the funky probabilistic thing isn't considered much because it tends to
occur where the process has an explicit random element (like a
lottery/stochastic election system) or in certain types of "prisoner's
dilemma" games, where actors possess mutual information about their
utilities.

Utility is scale-insensitive in that any positive linear transformation on
a set of utilities will be considered exactly the same way as that set of
utilities. Hence Von Neumann / Morgenstern normalisation of utilities
(0-1).

Disinterest in action is considered to be driven not by the scale
of a utility (otherwise the paragraph above would be untrue) but by
disincentives- erosion of utility by, for instance, the difficulty of
walking to a voting booth, etc.

A voter is considered "disinterested" between outcomes only when she
applies the very same utility to both of them.

On Sat, 4 Mar 2000 DEMOREP1 at aol.com wrote:

> Back to elections 101 type material (in connection with the current Approval 
> Voting related postings)---
> 
> There is a plus N to minus N scale (N scale) for N choices.
> 
> +N            0             -N
> 
> The zero means neutral (do not care).
> 
> Example--
> 4  3  2   1      -1 -2 -3 -4
>     B  C        0   A             E
> 
> The scale can be made all positive by having 2N as the maximum but having N 
> or less be deemed negative (requiring some education for average citizens).
> 
> Such N scale is connected to the plus 100 Percentage to minus 100 Percentage 
> scale (P scale) for multiple choices.
> 
> +100         0             -100
> 
> The zero means neutral (do not care).
> 
> Example--
> 100   90     70               -20         -100
>           B      C             0     A             E
> 
> The scale can be made all positive by having a 100 maximum but having less 
> than 50 be deemed negative (requiring some education for average citizens).
> 
> There is also the possibility of tie votes on either scale by each voter for 
> 2 or more choices.
> 
> Various election methods have rather obvious defects when tested on such N 
> and P scales.
> 
> Simple Plurality assumes a vote (i.e. the 1 and only vote of a voter) is a 
> plus 1 on the N scale (having only 1 and zero) and/or a plus 100 on the P 
> scale (having only 100 and zero).
> 
> Simple Approval assumes all votes are only 1's or zeros on the N scale and/or 
> 100 or zero on the P scale.
> 
> In political reality some choices are obviously more critical than others on 
> either scales- 
> 
> War or Peace ?
> Total government tax/spending levels ?
> Shall act/omission such- and- such be a crime ?
> Who gets elected by the voters to be members of legislative bodies, executive 
> officers or judicial officers ?
> 
> Various criteria floating around apparently pick and choose a part of an N or 
> P scale to look at (but disregard the other parts) which causes much of the 
> criteria confusion (such as which or what majority in truncated vote cases).
> 
> The mere existance of the positive and negative parts of the N/P scales is 
> why I have repeatedly suggested having simple YES/NO votes on candidates in 
> public office elections (especially for executive and judicial offices).
> 
> Of course, computerized voting with a simple computer mouse can permit much 
> quicker scale voting.
> 
> 

--------------------------------------------------------------------
Atheists aren't vagrants- but we have no invisible means of support.