The Social Model

[Lord God Jinnai]

I'll wait and see cuz i still think the option to push an EG religion (a specific one) over others and a GWR should be allowable.

[Yoav Sissman]

From the Social model description:

A GRW spreading rate is then computed for the EG as (1-Traditionalism)*(K1-D)/K2, where K1 and K2 are scale factors allowing to use D in the calculation so the more similar attributes are, the fastest people will take the religion. A GRW spreading rate is then computed for the EG as (1-Traditionalism)*(K1-D)/K2, where K1 and K2 are scale factors allowing to use D in the calculation so the more similar attributes are, the fastest people will take the religion.

If a GRW is at the government (i.e. it's official), it has more power to convert people. So, instead of comparing D directly with the threshold, we'll really use D*(1-G*RC%), where G is a dummy being "1" if the GRW being checked happens to be the official one and "0" if not. RC% is the political power of the Religious Class, from the government model.

Second: If PR% is greater than 0%, then PR% is decreased by the magnitude of GRW's spreading rate.

It seems that unless there are sudden severe changes in the people traditionalism level, government endorsement of the GRW, or the political power of the religious class, the PR% will decrease in a pretty much constant rate (i.e. the PR% may be 100% in one turn, 97% in the next, 94% in the third, and so on).

We might as well leave it for playtesting, but still I would like to suggest a simple change to this system, that will hopefully benefit its realism. I believe it should include some random factor, to make the spreading rate less expected.

True, we haven't so far included random factors in most of the other social variables, but the changes in them are computed in a manner that is in itself complicated enough to prevent a long-term constant growth.

Also the nature of conversion is IMHO such, that its success rate isn't steady, in contrast to other, gradual changes in the social fabric. One's religion is a binary variable, not a real one, and this has some affect even when we are dealing with communities.

I believe that the normal distribution, with standard deviation equal to the mean, might suffice here. In this way the GRW spread rate could decrease occasionally as well as increase (but not go bellow 0) as long as the new religion hasn't taken over completely.

3) This sub-model is by far the one that needs more computational resources. I guess we can make the model just apply steps 1 and 4 each turn using stored TV, which is really cheap. TV would then be updated less frequently, like twice a century or so.

I disagree.

I think that full computational procedure can be easily preformed every turn, and is negligible when compared, for example, with the clock-ticks cost of the AI.

[Yoav Sissman]

This is officially the government thread, so I'll post here my thoughts concerning the government model as well.

As it is now, the government directly negotiated policies are calculated as the weighted sum of the views of all of the political blocks, which are actually the aggregated views of all social classes from which they are composed. So the basic principle is that the government profile average all of the social classes (according to their political power).

After noticing that this enables the ruler to have influence that is grater then his supposed relative power (simply by inputting a policy that is more extreme then his real one) I understood that some mechanism was developed to prevent this. It wasn't fully described in the model and I haven't been able to understand it.

Does anyone care to enlighten me?

As an alternative I would like to suggest a fundamental change of approach to the directly negotiated policies, which IMHO, can benefit both realism and gameplay.

The government profile will be the median of all views (with each class having his opinion measured according to its relative power).

We'll use private property (PP) for the purpose of example. For the sake of simplicity we'll assume just 4 social classes, each having 20% of the total political influence, with the ruler having the additional 20%. The total political power a social class posses is the sum of powers it derives from its relative weight in each of the PBs he's a member in, and the political power held by the corresponding PB in the current regime. This is the same as in the current system; only here we have the social classes directly negotiate the policies without using PBs after we've calculated the strength of each social class (otherwise the median would have been affected).

Let’s assume that the stance that class A pushing for on the PP issue is 20, class Bs stance is 40, class Cs is 60, and class Ds is 80. The ruler will thus be able to implement any stance between 40 and 60 using his interface, as while he's standing there he has the 'deciding vote' and have all of the power.

When we use the original system each marginal power enjoys actual power that is proportional to the distance between its stance and that of the actual government profile. In other words you get more power by being extreme. If that is the case, why won't every political figure act extremely to pull the government closer to him (as the ruler might want to do using his interface)?

The decisions should fall in the point where half of the powers stand in one side of it and the other half station themselves on the other side.

A real life example of this can be found in modern parliaments:

Where there is a dual party system, extremists have to influence the government decisions by working through one of the great parties (which are mostly moderate and have to compete for the votes of 'floating center'). So the representatives of the extremists only have political power as possible members of one of the big parties (which negotiate a policy that is somewhere in between them).

Where every part of the electorate have its own independent representation (multi-parties system) still the ruler (i.e. the president) pushes bills to match his preferences; when he have to compromise to achieve the required majority he often shapes a bill that is just soft enough for most of the representatives to support (because they would like to see a change that is at least as radical as the bill if not more). In the game we can compare it to the ruler wanting a high PP; regardless of weather it's 60 or 100 that he wants, he will only be able to get the actual level as high as 60. My suggested system doesn't require him to 'experiment' with his interface, neither does it give him unwanted advantage over the other PBs, that take a stance before him.

It is common to define citizens as either 'right wing' or 'left wing', depending on which side of the political map they support. The question of how hard-line or tame they are is of less importance, as in any case they only have one vote.

Notice that how extreme one policies are will still have an effect in the riots model; the further he's away from the government profile, the bigger his discontent will be. (He may even use the threat he imposes to increase his de facto power by affecting the ruler's judgment).

In the example set above the ruler's exact power, be it 1% or 30%, doesn't affect his true ability to influence government policies. As long as he wants to set PP to something between 40 and 60 he shall have complete power; he have 0 power to push it to anything lower then 40 or higher then 60. Perhaps it's a flaw in my suggested system, but nevertheless this is often the situation in RL, where those that sit in the middle of the political map can decide on the exact policy, using their deciding vote.

Of course that the ruler's chance of having the deciding voice still increases with his relative power. Also in the game there will often be much more then 4 distinct opinions, so the ruler's margin of freedom will increase more gradually when his power grows.


Another small comment concerning the model:

2) There's no official religion (secular State): RCM values are the average of moral code's values of all present religions.

Shouldn't the RCM values be the weighted sum of all present religions?

It's possible that this is the meaning of the text and I simply misunderstood it. But if not, I just think that not all present religions should necessarily have equal influence on RCM values.

[LDiCesare]

QUOTE]median[/QUOTE]
I think this is a good idea. It means that a party with 50% or more PP will be able to do whatever they like, discounting anybody else's opinions. As long as this can lead to riots, it seems good to me.

One's religion is a binary variable, not a real one

I have to disagree. Many people are both shintoists and bhuddists for example, but bhuddists are mostly not shinto. Also, christianism gained slowly on "paganism" in Europe, so people still believe(d) in their local deities at the same time, simply turning them into spirits, faeries..., and actively having religious behaviour (like offerings made to prevent ghosts from entering the house rather than made to the gods, but still offerings). The fading of religions was, even for individuals, a continuous process, not a binary one.[

[Mark_Everson]

Hi Yoav, welcome to Clash!

First of all, this thread was listed as the current govt thread on the web page by accident. The govt thread these comments should really go to in the future is Govt Model v.2 (contd.) . I'll let Dom know about the bad link.

In that thread (and preceeding one, follow link at beginning of it) there was Extensive discussion of Yoav's idea on using the median. There its called things like 'voting' or the '51% rule'. I don't have anything to say on the topic that I haven't said already, but I saw problem with it in terms of instability that I think are dangerous in the game. But please go ahead and renew the discussion, and we'll see what the two of you can bring to the table. I am pretty sure that the discussion also includes what you (Yoav) were looking for in terms of the solution to the player misstating their preferences issue. If its not in Govt V2 (ctd.) try in the previous thread.

[roquijad]

Hi Yoav!

I appreciate your comments. It's clear you've given the models a very deep look.

I'm gonna answer here only the social model part: I'm not sure if I support adding little randomness to processes, but it'd be very easy to implement as you've said. I'd rather design them with randomness only where strictly necessary and later, when coded and tested (at least a little), decide if some parts may work better adding randomness. Sounds reasonable?

Mark and all: I'm beginning my come back to Clash after months away. I've actually started already, since I'm almost done with the new version of the social model (I'll release it next week). Don't expect major changes. I've included some things others proposed and have arranged some things more in the "objects" philosophy (for coding purposes).

Yoav, I agree with you that, at least for modeling purposes, it's better to handle religion in a discrete manner. In the new version EG's will have only one religion and conversion to other cults will be treated a little different. There're several reasons why I made this change, but one is because now it'll be possible for a given ethnicity (nationality) to have more than one religion (jewish american, catholic american, muslim american, etc).

Yoav: I'll post my answer to your comment on the govt model in the govt model thread.

[Lord God Jinnai]

wb roq..When do you think you'll have your model ready to view? yours is the one i've most anticipated

[Yoav Sissman]

Hi again all.

I'm not completely new to this forum; I've made several posts about a month ago under the name of 'Yoav_Sissman', and since then had to recreate my account due to technical reasons.

I'll comment on the government related issues in the government thread from now on, after I catch up on the matter there, of course.

Mark, it seems that not all threads can be accessed directly from the Clash Forum. For future reference, is it safe to assume that all of the interesting ones are linked to their appropriate model descriptions?

I'll hate to repeat stuff from old arguments, or to bother you with supplying the links again.

LDiCesare, I'm afraid I only partially accept your comments on religion.

Your example of Buddhists- Shintoists people (on a stable, permanent basis) is probably possible in real life, but it's my impression that in Clash we simplify things and don't allow religions to coexist for long in one EG. Of course that the new model may change all of that and make my suggestion irrelevant.

True, Christianity took over Paganism in a continuous process, but it was often ignited by a single decision (that in most cases came from above) to accept Christianity. That decision wasn't necessarily made on a personal basis, but often in the community level or higher. In the Early Medieval period the decision of a country to accept Christianity was often stimulated by the conversion of the king.

It is for those reasons that I believe that the conversion process of an EG shouldn't be smooth all the way.

roquijad, it's perfectly reasonable to leave the decision on the actual random factors to playtesting. I just wanted to bring the matter up for discussion.

Maybe I haven't made myself clear enough on this discrete/real issue.

I haven't complained about EGs having more then one religion, I only made my distinction about religions to support my suggestion that it should include a random factor.

I actually think it's preferable to allow one EG to have more then one GRW.

[Mark_Everson]

Hi Yoav, sorry I forgot you! .

I thought you had posted before, but I searched on your handle and found only the two posts from that day! So wanting to welcome each potential galley slave, er team member, I erred on the side of perhaps welcoming you twice .

The links are Supposed to be mostly up to date, but sometimes aren't. I put in a request in the web page thread at top for all the model leads to look over their respective pages for errors and omisions in the threads they point to. The misleading link was in fact my fault, I requested it several months ago. If you do find things that appear to be screwed up, let us know in the web page thread (if you're sure), or the relevant model thread (if maybe things aren't right).

I will now stop blathering about OT things here...

[roquijad]

Hi LGJ!
My intention is to release the new version this weekend...

[Simon Loverix]

On religion acceptance..
The Middle Ages are generally believed to be devotedly christian, but recently there have been several argumentations that that was for a large part wishful thinking from the side of the religious elite ( almost the only source of written information from that time). So the Middle Ages probably have been more pagan than the traditional picture tells us.
Also, when the Portugues colonized parts of Africa, the local elite accepted christianity as part of a trade deal (they didn't drop their other religion!), but the people didn't. When later, in the scramble the missionaries came to central Africa they only found remnants and elements of christianity, integrated with the local beliefs.
So some religions may be easier on coexistence with others. (Is there a tolerance rating in their profile?)

[Gary Thomas]

It seems to me that when someone believes in two religions simultaneously, then that is a religion in itself. So, from Laurent's example, there are three religions: Shintoism, Buddhism and Shintoism/Buddhism.

I still have a great deal of trouble understanding what, in Clash terms, an "ethnic group" is. I live in New Zealand, where the indigenous people (the Maori) form a distinct ethnic group comprising a significant proportion of the population (around 8-20%, depending on how you count those of mixed race). This "ethnicity" is completely unrelated to religion. Because of the way in which colonization took place they are probably mostly Christian. On the other hand I would hardly describe Christianity as a single religion - excluding splinter groups, it is at least four religions - and Maori are more or less evenly divided between three of them (there are probably no Orthodox Maori).

Cheers

[roquijad]

You two have made very good points on religion mixing. At least you convinced me.

Anyway, in the social model there was never an intention to model religion mixing and an EG sharing two religions was only a way to model transition (conversion).

Although I accept religion mixing exists, I don't think it's worth modeling. I believe the social model has to be limited to provide what is strictly needed. We all want religions to play a role in Clash, but the social model shouldn't go into further detail, IMO.


Gary: It's very important to forget about races when speaking about EG's. We decided some time ago to not model races, so watch your analysis on the australian aborigins. In real life the term "ethnic" is usually mixed up with race elements, but in the model race doesn't exist.


All: Social Model Version 2 has been released. At least a descriptive document is available (a technical doc (equations) will be available soon). I've sent the doc to Dom and he'll tell us when it's up in the web site. In the mean time, if any of you wants a copy of it, send me an email ( roquijad@cec.uchile.cl ).

[Gary Thomas]

Gary: It's very important to forget about races when speaking about EG's. We decided some time ago to not model races, so watch your analysis on the australian aborigins. In real life the term "ethnic" is usually mixed up with race elements, but in the model race doesn't exist.

The Maori are of Polynesian race, but there are many other Polynesians in New Zealand who are not Maori. The group is very much an ethnic one, with strong cultural patterns not shared by other Polynesians. Nor was the example intended to make any point related to race, it was directed at the fact some ethnic groups do not share a common religion.

I am afraid that I am still unclear on what a Clash ethnic group is.

I have never mentioned Australian aborigines.

Cheers

[roquijad]

oops: My mind took New Zealand and converted it to "Australia". Funny how the brain works.....

oops again: new version of social model is 3, not 2.

EG in Clash=A culturally homogeneous group of people, sharing the same nationality and religion

where "culturally homog." means people inside the EG are equally individualistic, equally traditionalistic, equally nationalistic, etc (all attributes that define a (Clash) culture).

All the people in NZ who consider themselves more Maori than New Zealander, would in Clash be modeled as having the Maori nationality (in Clash nationality denotes the feeling of belonging rather than a "legal" nationality). Since there're Maories with different religions, you'd have in Clash terms several EG's, all sharing the Maori nationality, but each with a different religion.

Don't get confused with EG's. The model simply groups people with things in common to simulate their behavior. An EG is nothing but a "population unit" whose members have a lot of things in common. I understand the word "ethnic" may confuse, but don't take it too rigidly. In the social model the EG may not be a precise representation of what in RL an ethnicity is, but it's close enough: common cultural attributes, common religion, common sense of tribal/national belonging.