[EM] A graphical simulation for your browser. 50+ methods

[Kevin Venzke]

Hi Chris,

Thanks for your comment.

> Kevin,
> 
>> A central concern is that when ballots are truncated, preference information is
>> lost, and that loss may be unevenly distributed. 
> 
> Given that voters are allowed to have equal-preferences, how do we know that 
> necessarily "preference information is lost"?
> 
> The "lost" or hidden preferences  (a) may not exist and (b) be of such a quality
> that the method is better served not knowing what they are, because if it did it
> could be confused into producing a lower social utility winner.
> I think it is better to assume that a lot of truncation (especially by the
> supporters of the major candidates) is natural and normal, and then assess the
> consequences of voters adding preferences.

My comment is specifically about the simulation itself, in which voters actually do
not possess any equal preferences, because the utilities are based on spatial
distances that aren't likely to produce real preference ties.

I think it's difficult to assess claims about utility independent of what truncation
logic was used, but my simulation presents utility stats anyway, so viewing through
that lens is available, if one prefers.

I certainly do think that lots of truncation by the supporters of major candidates
is "natural and normal," however I find it disturbing in any case if the ultimate
outcome of an election doesn't seem to conform to what we might expect given the
sincere preferences. A frontrunner can possibly be literally anyone, and there's no
guarantee that if all voters use a frontrunner-based strategy that something
positive will result.

Kevin
votingmethods.net



> Kevin Venzke stepjak at yahoo.fr
> Mon May 6 14:49:03 PDT 2024
> 
> Hello, here is a scenario generator to explore:
> 
> What if voters only use truncation strategies?
> votingmethods.net/trunc
> 
> Features: Over 50 methods, numerous settings, spatial or non-spatial preferences,
> ability to save/load scenarios, conversion to ballots, stats on sincere Condorcet
> efficiency and method similarity.
> 
> The truncation strategies include fixed or random thresholds, as well as strategies
> based on the voters having an awareness of two frontrunners, chosen in various ways.
> 
> A central concern is that when ballots are truncated, preference information is
> lost, and that loss may be unevenly distributed. I try to depict this graphically.
> You can observe often that a sincere Condorcet winner loses the CW status to someone
> else when the votes are cast.
> 
> I could not include STAR, since that is a ratings method, but I did implement "ATAR"
> and "BTAR" substitutes using implicit approval and an interpretation of Borda.
> 
> Kevin
> votingmethods.net