Coding the "Society Model"

[Gary Thomas]

RULERPW (ruler power) is a complicated function of NRULERPW (current government ruler power), in Section I of the equations.

On the other hand, in section IX of the equations, NRULERPW is a complicated function of RULERPW.

Are these supposed to be solved as simultaneous equations, or what?

I still haven't got an answer to this question.

Cheers

[Gary Thomas]

It has just occurred to me that the equations in the social model may be sequential, that is, section I is calculated, then section II, section III, and so forth.

Is anyone able to tell me if this is so?

Cheers

[Mark_Everson]

Hi Gary:

Sequential is the way I'd always interpreted it. I'm pretty sure that's right, but I was not responding for fear I would get it wrong... I've emailed Rodrigo, and asked him to check in here, since there are some issues that need his comments. If he doesn't show shortly I'll re-read the doc and see if my interpretation is consistent with what's in there.

[roquijad]

Sorry, Gary. Here are a few answers for you:

On demographics: using fix shares is more than enough for the moment.

On initial values: the document I made defines values for regimes and for social classes. As for religions and EG's cultural values, random values will do for now. But your "scenario-builder" sounds pretty useful!

On ruler power: there's nothing simult. in the model, but the order of "procedures" I put in the doc isn't the best seq. order... For your particular question the answer is "compute real power distribution first, then nominal (for the next turn)". In general, the order of procedures should be:

*X. Handling Religions Grand Scheme
*I. Computing discrimination factors for an EG.
*V. Creating govt's Code of Law
XIX. Determining Govt's Official Religion.
XI. Determining Knowledge and Development levels for a Regime.
X. Determining the demographic share for a social class.
XII. Determining the Govt-Regime Distance and the Ruler-Regime Distance for a Regime.
XIII. Determining Social Class' Historical PCI
XIV. Updating Empire's Stability
XV. Determining Optimum Institutional Infrastructure (OII) needed in a province.
XVI. Determining the Number of Administrators Needed (NAN) in a province.
XVII. Determining Administration Effectiveness Level (AEL) for a province.
XVIII. Determining the Maximum Number of Feudal Units (MNFU) in a province.
I. Computing Real Power Distribution
II. Creating the Conservative Govt Profile
III. Computing the 5 Political Powers shares for a social class.
IV. Computing preferences for DNPs for a social class.
V. Creating the Support Regimes Shares for a social class.
VI. Creating the 5 Political Block Profiles.
VII. Creating the Society's Govt Profile.
VIII. Creating the Future Govt Profile.
*II. Computing Govt Nationality's Force for a province.
*III. Computing Tendency Values for an EG
*IV. Changing current EG's attributes (cultural evolution)
*VI. Computing Faith Points for a religion in a province.
*VII. Changing EG's populations in a province via religion (in-the-province religion spreading)
*VIII. Updating the list of present religions in a province because of trade (province-to-province religion spreading).
*IX. Changing EG's populations in a province via nationality (nationality shift)
IX. Creating the Current Govt Profile.


Note1: procedures with asterisk are from Social Model. Without asterisk is from Govt Model.

Note2: The last procedure listed should be, ideally, made at the very end of the game turn because that's where the next turn's current govt profile is created.

[Gary Thomas]

Thanks, that was just what I needed.

Cheers

[Gary Thomas]

I am (very) slowly assembling all the parts of the Social/Government/Riots model.

So far I have (in addition to a great many pages copied from the forum discussions):

1. The Social model, version 3, dated October 2001 (though the last update listed for it is 6/22/00, which I assume means 2000/6/22). This came from the web site. It is a general discussion of the model, without equations.

2. The Government model, version 3.2, dated October 2001. This also came from the web site, and is likewise a general discussion of the model, without equations.

3. Government Model version 3.1, Technical Document (equations), dated December 2001 (version 3.1 is later than version 3.2?). This has the equations for the government model.

4. The Riots model, version 2, dated October 2001 (also with the 2000/6/22 update value). A general discussion, without equations.

What I am lacking is a clear notion of the current version numbers, and which is the latest version, and, particularly, the equations for the Social and Riots models.

Social coding (which means, in this case, xml coding) is on hold until I find out where to get the missing equations.

Cheers

[roquijad]

Hi Gary.

Current (latest) versions for technical and descriptive documents are:

Social: 3.0
Govt: 3.1
Riots: 2.0

There has never been a version 3.2 of the govt model. One of the documents had that number in the title, but that was a typo. A similar problem may have ocurred with riots model. Probably somewhere there's a version 3 of that model, but latest really is version 2. Sorry for the mess.

Anyway, to put away any doubt, I just sent you via email (to thomasg@acc.co.nz) all documents available for latest versions. That is:
Govt Model -description (version 3.1)
Govt Model -technical (version 3.1)
Social Model -description (version 3)
Social Model -technical (version 3)
Riots Model -description (version 2)
Coding the Society Model (version 2)

(a technical version of the Riots Model isn't available yet)

Note: a couple of errors in equations of govt model are described in my post dated 12-02-2002 02:45 in this thread. Take a look at that. (second page of the thread)

[Gary Thomas]

Thanks. Mark sent me the mssing equations one, so I have a full set now.

I have a question about nationalities. It is quite unclear whether "nationality" is synonymous with "civilization".

Is it possible for two civilizations to have the same nationality? It is clear that an ethnic group can have a nationality which is not represented by a civilization - if the ethnic group becomes independent, a new civilizations is formed, according to the model. Presumably a ruler will need to be randomly generated, and the various government parameters somehow generated from the characteristics of the ethnic group.

Currently the nationality is merely recorded as a string, that is, the name of the nationality. This is no good, since I can envisage situations in which a civilization's name is changed (from Wessex to England, perhaps?). This would leave the nationality names out of whack.

If I understand the intent of nationality, then I feel the best way to deal with it is replace nationality with civilization, creating any, in the scenario xml, which are required. This enables all the data to be set at the time the scenario is input.

Some civilizations which exist purely to provide a home for a possible nationality will have the government structure in place, but control no provinces.

Cheers

[Gary Thomas]

roquijad:

*X. Handling Religions Grand Scheme

Unfortunately the social model only goes to VII (and VII is just changes from previous versions).

Did you mean "VI.4 Religions Grand Scheme sub-model" here? with similar adjustments for the others?

Cheers

[roquijad]

Gary, all procedures I listed ("Handling Religions Grand Shceme", etc) refer to the technical docs of the govt and social models. It's the descriptive version of the social model that goes up to VII, so there's your confusion.
Note numbering of procedures and numbering of different model topics do [b]not[/n] match. In other words, numbering in descriptive doc has nothing to do with numbering in technical doc.


About nationalities: Nationality represents the nation/tribe the EG feels it belongs to. (Mark proposed a better name was NationalAffinity). That nation or tribe may or may not exist as a civ/empire. FE, after Hernan Cortes' conquest of aztec empire, there'd be no aztec "civ", but there'd be EG's with aztec nationality, i.e. people who still called themselves aztecs.

It's important to avoid "connecting" an EG with a civ. Things like a "pointer" from the EG to a civ should not exist. You only need to know in what civ the EG lives. The model was built with that in mind. Only in very specific, infrequent situations (actually described in the social model descriptive doc, IIRC) you need to see if for a given EG there's somewhere in the world a civ with matching nationality.

There shouldn't be two civs (govts) with the same nationality. But that's something that should be controlled when a new civ is created and when the player wants to change the civ's nationality. In general, if the civ's nationality changes, there's no need to change also EG's nationality.

[Gary Thomas]

The equation for religious discrimination (Social Equations I) is:

RD * (1 - (1 - sign(FVE - FVG)) * 0.5)

which I interpret to mean:

if (fve < fvg) return 0
if (fve = fvg) return rd / 2
if (fve > fvg) return rd

Is this correct? I cannot see it providing useful results.

I don't see that family values is being treated correctly. If the state religion and the ethnic group's religion are the same, zero discrimination is returned.

Family values are only compared for different religions, so I cannot see that the expressed intention of providing a treatment for similar religions, withing a larger religious grouping (Catholic versus Protestant is your example) is going to work.

[edit]
I thought about it a bit more, and assume that the intent was for related religions to halve the discrimination factor. Is that the case?

Cheers

[Gary Thomas]

I am working my way through the sequential list.

*X I have deferred to later
*I is discussed in my previous post.
*V is the one I am consideringg now.

My question is, how can the central government have a "province's Administrative Effectiveness Level"?

Cheers

[Gary Thomas]

Gary, all procedures I listed ("Handling Religions Grand Shceme", etc) refer to the technical docs of the govt and social models. It's the descriptive version of the social model that goes up to VII, so there's your confusion.

Nevertheless my social technical document only goes as far as "VII Changing EG's populations in a province via religion". Am I missing some pages?

About nationalities: Nationality represents the nation/tribe the EG feels it belongs to. (Mark proposed a better name was NationalAffinity). That nation or tribe may or may not exist as a civ/empire. FE, after Hernan Cortes' conquest of aztec empire, there'd be no aztec "civ", but there'd be EG's with aztec nationality, i.e. people who still called themselves aztecs.

I do understand that. What I have difficulty accepting is that "nationality" exists as just a string. A nationality must have more characteristics than merely a string, a name in a single language. In any case, as long as there are people who can consider themselves aztecs, there is an aztec civilization, even if it has no political control of anything, the concept exists.

It's important to avoid "connecting" an EG with a civ. Things like a "pointer" from the EG to a civ should not exist.

Why is this? It seems to me that this is just what is needed.

Essentially, my problem is that if nationality is required, I have to attempt to match it to a civilization. One way I can see of doing that is to attempt to match the nationality string to a civilization name. If a match occurs, then that is the civilizatoion that matches. So, if the aztecs are restless, Cortez should change the name of his civilization from "Spanish" to "Aztec" which would have the nice effect of calming the natives, though it might upset some people in Madrid.

If, on the other hand, the civilization itself has a nationality, which cannot be changed by the player, then the match should be to that nationality (rather than the name of the civilization). It seems to me to be completely irrelevant whether the match is done at the start or when needed. In particular, when a scenario is input from an xml file is the time to do it.

Cheers

[roquijad]

I thought about it a bit more, and assume that the intent was for related religions to halve the discrimination factor. Is that the case?

That's the case. And clearly there's something bad in the equation. It should be:
RD*(1-(1-Abs(Sign(FVE-FVG)))*0.5)

My question is, how can the central government have a "province's Administrative Effectiveness Level"?

The central govt has not an AEL. Each province has an AEL. The central govt controls provinces through its control over "local govts". Some effects of the central govt over the provinces are "felt" by EGs through the local govts. Models, then, make reference to characteristics of those local govts. In particular, AEL.

Nevertheless my social technical document only goes as far as "VII Changing EG's populations in a province via religion". Am I missing some pages?

It seems so... Give me your email and I'll send you the docs (the adress I used, that I mentioned in the other post, failed).

In any case, as long as there are people who can consider themselves aztecs, there is an aztec civilization, even if it has no political control of anything, the concept exists.

Well, yes, but I don't think we need to encapsulate EGs in a "thing" called civilization. The question is why would you want to do that. What game experiences can derive from such encapsulation we aren't covering now?

This is beyond my competence, but I don't see any reason to model "civilizations" as groups of people with certain common characteristics. Even more, I think that's a bad idea. I think we need govts and EGs. Nothing more. The "civ" the player controls is, really, a govt that, in time, contains people of different ethnicities. The word "civilization" is actually very deceiving. In terms of a game like this, it's IMO useless because players control empires (political instances). They don't control civilizations (that is, they can't control the customs, religions and other social characteristics people have). If we agree players are constrained to affect the game only from the position of the ruler of one govt, then the "thing" called civilization isn't needed.

Unfortunately this is the type of topic that easily "booms" to abstract dicussions stopping coding progress... I firmly believe all relevant social behaviors/effects are well modeled w/o the need to unite EGs to an abstract thing called civilization.

Nationality is needed mostly to model rebellions. For that purpose, the only info I need is what nationality an EG has and what nationality the govt under which control the EG lives, has. Simply put, a mismatch implies unrest.

There's no need to describe a nationality in more detail:

A nationality must have more characteristics than merely a string, a name in a single language

You're thinking EGs here. The EG is described in detail, socially. In particular, the people of the EG call themselves parts of a tribe called [nationality].

Essentially, my problem is that if nationality is required, I have to attempt to match it to a civilization.

Why you "have" to attempt that? What happens if you don't?

[Gary Thomas]

It seems so... Give me your email and I'll send you the docs (the adress I used, that I mentioned in the other post, failed).

I have located the missing two pages, thanks.

Nationality is needed mostly to model rebellions. For that purpose, the only info I need is what nationality an EG has and what nationality the govt under which control the EG lives, has. Simply put, a mismatch implies unrest.

So it is only used to determine whether the owning civilization is the "right" one and for no other purpose?

You're thinking EGs here. The EG is described in detail, socially. In particular, the people of the EG call themselves parts of a tribe called [nationality].

I am actually thinking object-oriented code, where objects are better than strings.

Why you "have" to attempt that? What happens if you don't?

I perhaps haven't had the time to go over all parts of the model as thoroughly as I would like, but I was going from the statement at the start of section IV of the Riots Model, Pro-Action Feelings, under NRF where it states that a rebellious ethnic group might want to "join the civ of the same nationality if it exists". This means that I have to identify that civilization. The best way to do that is to have a pointer to it, or a null if it doesn't exist. Clearly, if the ethnic group has a null nationality, it will not match any civilization, which presumably is what is intended if there is no civilization matching the nationality.

Hence, by saving a pointer to the appropriate civilization, or null if no such exists, I do not need to do a search later. There doesn't appear to be anything in the game that can change this nationality or this mapping.

Cheers