Coding the "Society Model"

[roquijad]

Laurent: In the package Mark sent you, you should find equations for social and govt models, but not for riots model. I checked my other PC and the document with riots model equations was nearly finished. Give a couple of days and I'll send a finished version.

[Chilean Presidentâ„¢]

[small_threadjack]hola Rodrigo, sólo pasaba por aqui para saludarte. He estado mirando el trabajo que están llevando aqui y es excelente. felicitaciones [/small_threadjack]

[LDiCesare]

Here's a small progress report:
I've managed to implement a not-omnipotent ruler. So the negociated policies thing is working. I must update the ruler user interface to show the result of what choices you take. Right now everything is static, so no tendencies or changing with time. I'll add that after, when the static stuff is working. IT might be worth delivering the static model before I start on the dynamic one.
I think I'll try to put most of the equations in the xml files so that it will be easy to tweak the figures. Many equation have figures like 22 or 8 which you probably worked a lot on in order to simulate, but if they don't work as intended, it will be easier to change it in the xml files than to have me recompile everything.

I'm still fighting with the existing code to find out what's totally unused, what's coded but in accordance with the current model etc.

The ruler panel needs some work, and the UI in general because it's totally heterogeneous and somewhat bugged. So I'll not be in the equations soon, but I hope I'll be able to make some progress next week.

[roquijad]

Hi All,

After more than a month (sorry, too much work in real life), I have the equations for the riots model. I've attached it to this post.

I see you're on the govt model now, Laurent. Good. And I saw a very nice interface in the other thread. Good too! You seem to have chosen a simple average for the negotiated system. I think that's a good decision to start with. We'll improve it later.

Tell me how I can be of help now.

And hey, presidente de Chile (chuaaaa!), thanks for your support!! We move slow, but we hope we'll get this game running in our life time!

[Mark_Everson]

Hi Rodrigo! That's quite a doc! I am swamped with RL at the moment too. I will check it out when I can.

Cya,

Mark

[yellowdaddy]

languages

Where do things stand on the (token) languages front?

(I remember some of the coders being indifferent to them, but some of the others and potential players being more keen...)

Can we have a poll on languages?

[LDiCesare]

Rodrigo, I downloaded the doc, but didn't look much into it today (I'm a bit sick this week end and the rest of the week quite busy with RL... okay, and playing nethack too). Right now, the social model scenario I made is over a short period of time. Looking at the model a bit more, I think it's a mistake. Science is a quite important part of the model in the sense it drives evolution of several parameters. I haven't coded the evolution over time of the various parties desired policies, but that should probably be in the scenario, as there are two ways to lower ethnic discontent: By using the government panel (saying you want no ethnic/religious discrimination) and, indirectly, by researching new applications which will help mentalities evolve.
About languages, that's not for the immediate future as I'm coding very slowly right now.

[wombat42]

Even though I've been very much for languages all along, I recognize that the crew has a lot to code right now, so it can wait a while. It is inevitable unless you want the Inca battling the Koreans on some randomly generated world in 5000 BC, but it can wait.

[yellowdaddy]

sure, i bring it up again becasue they were talking about a way distinguishing a Latin EG1 from a LatinEG2 (for instance).
I argue that language is the obvious way.

i posted a while back a scheme - a lunage tree i researched years ago, to break down humanity into 72 language family gruops or "peoples" which could/would split into 6 "tribes" and then each tribe into 6 "nations"

e.g.:
People - Tribe - Nation
Italics - Latins - Italians

I think is the best way for historical (ethnoliguistic) accuracy and for reflecting the stages of human civilisation evolution in a "from the dawn of time" game.
I did a lot of deep research, I think it'd be a shame not to ustilise it somehow, eventually.

but yeah, perhaps now isn't the time, unless programming tribal splitting is something that might be tricky to do lateeron.

[LDiCesare]

I've mostly finished coding the ethnic discrimination riots (beware ethnic discrimination - it can spawn a LOT of units if you control many unhappy provinces). I started coding the 'bad policy feeling' one, and have a few questions.
How do different riot causes stack?
If 3 social classes have a riotting chancd of 20%, do I sum these to get 60%, or do I roll 20% three times?

Also, it seems weird that 'bad policy feeling' and 'ethnic discrimination feeling' would both be computed based on social classes. So right now Ethnic Discrimination is based on ethnic groups as a whole, while Bad Policy is trying to be based on social classes (I must just put the social classes proportion in). But maybe I should instead add the probabilities of both events together in order to check what happens?

Last, I couldn't find a term in the probability of showing discontent based on the proportion of the population. It seems likely that a population will try to shout more if its population is more important? At least for ethnic discrimination.
A related point: The effects of revolts should vary based on the riotting social class. Kapital would probably be better off corrupting the economy than going to the streets with weapons, particularly if their numbers are low?

[LDiCesare]

I'll be answering my own questions since noone seems to want to do it:

How do different riot causes stack?
If 3 social classes have a riotting chancd of 20%, do I sum these to get 60%, or do I roll 20% three times?

The model says there's an addition of all PAFs, with a weighted factor, to compute the probability. So it's a sum and was written in the model. So long for me.

So right now Ethnic Discrimination is based on ethnic groups as a whole, while Bad Policy is trying to be based on social classes (I must just put the social classes proportion in). But maybe I should instead add the probabilities of both events together in order to check what happens?

Here I came to discover that there's a difference between and the model and the code. A rather big one. PAFs should be added to compute events, while the code computes PAFs inside the events, which means the addition of PAFs should be done outside. Since I haven't coded many events yet, it's no big deal, particularly since the ethnic discontent spawns units different from other events (civ based on EG).

The other points remain open.

An additional question:
I'd like to have the Bad Policies Feeling coded, but to that end I need to have each social class' prefered policy somewhere. So I'm busy coding that. When done, I think I'll have enough of a govt/riot model coded for it to be released.

[Mark_Everson]

Hi Laurent:

I didn't say anything because I didn't consider myself knowledgable enough anymore. Unless there is a substantial amount of effort in changing code to a different approach, I'd say just pick whichever way makes sense to you. We can email Rodrigo if you don't get feedback on something that is not easily changed later. Also I'll force myself to dig in and answer any question you think is really critical.

The progress sounds great!