Simulacra

[William Waugh]

I have started coding an interactive simulator for evaluating single-winner voting systems against each other for their degree of support of vs. resistance to two-party dominance (2PD). I haven't gotten anywhere at all really, just practiced putting up a slider. But the page tells in which direction I intend to be heading.

So my idea is that a user can move sliders to establish parameters of a simulated electorate, and the code will then simulate elections using some voting systems, and we can all see which systems will elect the right candidate under the least friendly electorate. I intend that the human partner in the interaction will titrate to determine that threshold for each voting system. Then it will be possible to argue that one is better than another, and by how much.

If I get slow in coding this, it would be awesome if someone or several were to clone this code (what there is of it) and advance ahead of me on the work.

The page is not only going to run the interaction and simulation, but also expose and explain its own code, using, I hope, a "literary" programming style. Once I get voting systems and strategies in there, I want others to be able to contribute additional or substitute strategies. I want the code to be very readable so as to minimize barriers to others adding to it.

Since the simulator will run in a browser, the code is in Javascript (JS).

I don't know how to style this kind of work for handheld electronics, so at least for now, this is aimed at desktop computers with normal full-sized monitors. If anyone wants to suggest improvements to the style (with the CSS etc. to do it), I'm open.

William Waugh

PS -- I have been calling myself "William Waugh" when advocating about US electoral politics. When I write code, I use the name "Jack Waugh". So the subject web page is coded by, and signed by, my coding persona, Jack Waugh.

[Warren D Smith]

If you want to simulate elections to deduce which voting methods yield
2-party dominance,
then

(a) 2PD does not happen in 1 election, it develops over a sequence of many
elections taking 50-100 years.
(b) The reasons for it are in part psychological. The psychology of the voters.
That means you somehow have to simulate psychology.
(c) How can you do that in any manner that is going to convince anybody your
results mean anything? I mean, does your computer understand my psychology?
(d) So I think the main obstacle is not the lack of a computer simulator.
The main obstacle is the lack of a mathematical understanding of, or
mathematical
model of, psychology -- one supported by enough real-world evidence,
that anybody reading your study (whenever you get done creating it)
will have confidence in its validity.

Abcd all occurred to me, and I never had a good answer,
which is why I never tried such a sim-project myself.


--
Warren D. Smith
http://RangeVoting.org <-- add your endorsement (by clicking
"endorse" as 1st step)

[William Waugh]

On Sunday, November 11, 2018 at 1:26:55 PM UTC-5, Warren D. Smith (CRV cofounder, http://RangeVoting.org) wrote:

If you want to simulate elections to deduce which voting methods yield
2-party dominance,
then

(a) 2PD does not happen in 1 election, it develops over a sequence of many
elections taking 50-100 years.

I don't see that. 

Definition: "Repugnicrat" -- anyone who supports the Republican Party or the DINO Party to the exclusion of independent candidates and to the exclusion of "third" parties.

Definition: DINO -- democrat In Name Only. Note that I use a lowercase "d" in the word "democrat" in my definition, and that's not just a slip of my fingers on the keyboard or the carelessness typical of writers who substitute "too" for "to" and vice versa; I mean something by writing that "d" in lowercase. DINO means the party that Orwellianly styles itself as the "Democratic" Party, and when I write that, I want you to read extra-extra horror into the horror quotes.

Definition: "democrat" -- someone who judges that political power should distribute approximately equally among the reasonably competent citizens, and definitely not in a way that depends on how much money the citizens have or whether they are officers in large corporations.

I'm suggesting that 2PD can be seen to operate in one hypothetical or model election to the degree that the anti-Repugnicrat faction are thwarted getting their way, which would be to elect one of their own if possible, and otherwise, to support the lesser evil. Of course different subfactions of the anti-Repugnicrat faction could have different opinions about who is the lesser evil.
 

(b) The reasons for it are in part psychological.  The psychology of the voters.
That means you somehow have to simulate psychology.

No, you do not. It does not take psychology to understand why pines grow tall. They compete for sunlight, even within their species. This only requires the most rudimentary introduction to theory of evolution to understand, which is a lot simpler than psychology. The same for voters. I contend that for purposes of comparing voting systems, we can abstract away from the details of psychology and do fine with an assumption that each voter will use the best strategy we know of for moving the needle of the election outcome in a direction favorable to that voter's valuation of the candidacies, under an assumption that that voter cannot count on any cooperation from any other voters (same conditions as the famous Prisoner's Dilemma in game theory).
 

(c) How can you do that in any manner that is going to convince anybody your
results mean anything? I mean, does your computer understand my psychology?
(d) So I think the main obstacle is not the lack of a computer simulator.
The main obstacle is the lack of a mathematical understanding of, or
mathematical
model of, psychology -- one supported by enough real-world evidence,
that anybody reading your study (whenever you get done creating it)
will have confidence in its validity.

I am trying to improve on the current state of the argument, not make it perfect. I am not going for real-world evidence. I am going for example hypothetical fake elections where one voting system seems to perform better than another. An accumulation of such examples may be convincing. The current state of the argument is there is a lot of theorizing going on without in my opinion sufficient examples to convince. Of course, when I become able to show the examples, they may still not convince and some IRV advocates will no doubt charge that the examples don't represent reality, as though we who hear their arguments are supposed to believe that there is some secret factor in the real world that will work to prevent our examples from coming true. But I do hope to improve the quality of the argument over where it is now.

For an example of the current state of the argument, I can cite my vague belief that if a system has balance, it probably gives everyone equal power, or at least that is evidence that it may give them approximately equal power, and I can point to the certainty that if a system does not meet balance there are definitely at least some hypothetical situations in which it denies the voters equality of power. And I can say, reflecting I think a point closely related to your point that you make along with Jennings and Shentrup in the "as it ought to be" post, that when systems do not provide equality of power (FPtP being a gross example), the voters receive an incentive to seek power, which works against the incentive to seek merit, and so as you point out with your co-authors, money wins because it buys the power in FPtP and I contend in all other unbalanced systems as well. OK, so that's my argument, and I believe that from the viewpoint of the people I try to convince with it, it is way too theoretical and abstract to draw their attention. So, I'm thinking that hypothetical examples even if they don't come from political reality, are still more concrete than that abstract argument. So I want to make the simulator so that I and those who will go to it when I publish it will be able to present many quasi-concrete examples to take the place of more-abstract argument.

Abcd [Above?] all occurred to me, and I never had a good answer,

which is why I never tried such a sim-project myself.

Nevertheless, I am grateful for some of the ones you have tried, and for your advocacy of the utilitarian position. 

[Warren D Smith]

On 11/21/18, William Waugh <2knuw...@snkmail.com> wrote:
> On Sunday, November 11, 2018 at 1:26:55 PM UTC-5, Warren D. Smith (CRV
> cofounder, http://RangeVoting.org) wrote:
>>
>> If you want to simulate elections to deduce which voting methods yield
>> 2-party dominance,
>> then
>>
>> (a) 2PD does not happen in 1 election, it develops over a sequence of many
>>
>> elections taking 50-100 years.
>>
>
> I don't see that.

--well, look at history and you will see it. E.g. in the USA, 2PD
solidified only
after about 80 or 90 years. In Canada and India, it arguably still has not
happened yet.

Plus, you can't have 2PD if nobody knows which are "the two" parties,
and that initially is the case. It only becomes clear because it becomes true
over time, and the two parties have to both become recognized as such,
and to strengthen in terms of wealth, candidates, donors, networks,...
and to evilly corrupt the laws to
try to artificially disfavor rivals, and the rivals have to become
weaker and less-wealthy,
etc etc -- all that usually takes time.

> Definition: "Repugnicrat" -- anyone who supports the Republican Party or
> the DINO Party to the exclusion of independent candidates and to the
> exclusion of "third" parties.
>
> Definition: DINO -- democrat In Name Only. Note that I use a lowercase "d"
> in the word "democrat" in my definition, and that's not just a slip of my
> fingers on the keyboard or the carelessness typical of writers who

> substitute "too" for "to" and *vice versa*; I mean something by writing

--good luck with all that. It doesn't sound like science to me. It
sounds like your
personal ranting about how the USA currently is messed up.

>
>> (c) How can you do that in any manner that is going to convince anybody
>> your
>> results mean anything? I mean, does your computer understand my
>> psychology?
>> (d) So I think the main obstacle is not the lack of a computer simulator.
>>
>> The main obstacle is the lack of a mathematical understanding of, or
>> mathematical
>> model of, psychology -- one supported by enough real-world evidence,
>> that anybody reading your study (whenever you get done creating it)
>> will have confidence in its validity.
>>
>
> I am trying to improve on the current state of the argument, not make it
> perfect. I am not going for real-world evidence. I am going for example
> hypothetical fake elections where one voting system seems to perform better
>
> than another. An accumulation of such examples may be convincing.

--well, I hope you succeed, but I suspect whatever you do it'll be met
with a lot of skepticism.

> The
> current state of the argument is there is a lot of theorizing going on
> without in my opinion sufficient examples to convince.

--but you just said you were going to disdain real-world evidence!

> Of course, when I
> become able to show the examples, they may still not convince and some IRV
> advocates will no doubt charge that the examples don't represent reality,
> as though we who hear their arguments are supposed to believe that there is
>
> some secret factor in the real world that will work to prevent our examples
>
> from coming true. But I do hope to improve the quality of the argument over
>
> where it is now.
>
> For an example of the current state of the argument, I can cite my vague
> belief that if a system has balance, it probably gives everyone equal
> power, or at least that is evidence that it may give them approximately

> equal power, and I can point to the certainty that if a system does *not*