Government Model v. 3.1

[Lord God Jinnai]

Why would say, discovering "Republic" enable you to do more than if you didn't know it? Its just a term and you could set your ideals to be a republic without knowing what it is.

Now i can see it having an inpact for the populace as its hard to move towards an ideal when it has no name to embody it...

[roquijad]

Hi all,

Congratulations, Laurent, for all your progress!!

It's true: the current regime is just a group a variables taking different values. That's the beauty of it in terms of possibilities: any form of regime and anything in between is possible. But yes, it's also true that that's totally ugly and makes more difficult to know in a simple way, what regime I'm in.

My idea has always been (and I don't remember if I wrote it in any of the documents) something like Mark proposed. The game "analyzes" the current setting and produces coherent texts that, briefly (a phrase), tell the player what he has as form of government. Something like "The empire is under a form of Theocracy that also rests on a powerful bureacracy", for example.

I admit I haven't thought about that too much, but I think is doable. For now, I'd suggest just a comparison of current setting with coded regimes that provides a text "The civ is under a form of " [closest_regime], with "closest" derived from closest distance in hyper-space.

Anyway, I had my doubts too, like yours, about the player having too much freedom to set his preferences, and actually trying to have modern regimes in ancient times, which I assume we all consider bizarre. And I too thought at that time of limiting player's liberty to set values as a solution. I think it's probable that we'll have to do that. But I wouldn't worry about it yet.

Quickly:

1) Why do we use exactly 3 regimes in addition to ruler and conservative? Why not 4 or 2 or all the known regimes?
What if only 1 or 2 regimes are known at that tech level? If it is here to ease computation, I'll gladly get rid of it because it makes for more programming.

Yes, it was for easing computation, but also to simplify player's understanding of people's preferrence. If you feel is better to not limit the number, do it.

2) Attractiveness of conservative regime: Why bureaucracy + warfare? This means the military is always conservative. If so, why not ethics? Churches tend to be as conservative as the military IMO.

Just a modelling decision.

3.6)
S6=50*Exp(-5+5*(1-ES))*RULERPW
What's ES? Empire Stability I guess? Meaning again?

Yes, ES=Empire Stability. ES represents people's perception of how stable is the government and the empire.

[quote[
4)What's the exact role of the ruler regime? The conservative government regime seems to be used only in riots, but it and the ruler regime are often excluded from the list of regimes to be checked. Why?
[/quote]
Ruler's regime appears to let people choose it as their preferred regime just like any other. If people like what the ruler's pushing for, they'll support those ideas. The same with conservative. If people prefer what the govt has been, they'll push to keep it.

5)What's the regime banning factor? It is set by who, how?

It's a way to introduce banned regimes. If you want to prohibit democratic ideas in your evil dictatorship, you change the banning factor. For now, IIRC, the model allows the ruler to ban regimes, as long as he is powerful enough.

I'll give you answers for what you asked about equations when I have a little more time. I'm going home now!

[LDiCesare]

Hi Rodrigo!

Any idea how best to present the player what 'the people' and the various political blocks want?
Currently the only ways to know population desires are:
-when they riot, you will know why. That's good, but a bit late.
-when you change policies, you see you can't get farther than this or that, which means that the power blocks want a value that can be guessed. That's also efficient for tuning things a bit, but doesn't give details on which political block wants what.
Any ideas on how to show that to the player would be welcome (that's user interface, and I don't like designing these things, so ideas are welcome).

I'm coding the regimes. That's more or less done, but I've problems with the user interface (looks like the power structure sliders do not always add up to 1 all the time when they should, which kind of wrecks the equations afterwards). I still have to show where people are.

[LDiCesare]

Knowledge of people desiderata

How much knowledge should the player be given of what the population actually wants in terms of policies?
As I said, you can currently infer some information from the negotiation system, but you can't know which social class wants what.
The system is quite complex, as each social class of each ethnic group in each province or even square may have different feelings. Showing this level of detail would require an elaborate user interface, and wouldn't be realistic since IRL you'd need polls, statistics, and some psychology in order to gather the correct information, which would be expensive. I also think it would be pretty useless because it'd be too much micromanagement.

I think showing the preferred negotiated policies and regimes per social class empire-wide would be necessary (in order to know which regime allows the ruler to reach their ends best and to avoid riots more effectively). I could offer a filter per ethnicity too, probably as a filter over the social classes, empire-wide.
I don't think details per province or per square need be provided. Per province could make sense, but province-specific user interface would require some work in order to have a generic way of entering a 'province management mode' by opposition to 'empire management mode'.

[Gary Thomas]

you'd need polls, statistics, and some psychology in order to gather the correct information, which would be expensive

Rulers in the archaic pre-poll and pre-statistics era had ways of judging the mood of the population, probably more efficient than polls and statistics. Roman Emporers were known to mingle, icognito, with the population. Spanish Kings (and other rulers of that time, including Elizabeth of England) had a very well developed spy system, loyal to them personally. Variants of these methods were widely used. Any ruler who did not at least try to get some feedback on popular feeling tended to be toast when the crunch came.

Cheers

[roquijad]

Hi Laurent. Here are some answers:

S1=20*IND*(p+Sign(L)*PPW+A*BURPW)
Why Sign(L) when everything else usesvalue * powershare? Coded anyway.

S1 represents how much a social class likes the distribution of power offered by a given regime, particularly how much power the social class gets. Workers classes get power only through the People's Political Block. This block represents the democratic ideal of "one man, one vote", so any social class with L greater than zero will appreciate a regime proportionally to how large PPB's power is. But all those social classes with L greater than zero must appreciate the regime in the same way, regardless of how large or small their particular value L is, because they won't get more or less "votes" depending on their L value, simply because in PPB, one man is just one vote. Therefore, we only need to know if the social class is or isn't a worker class, which is the reason to use Sign(L).
Note that in the rest Political Blocks, social classes do get more power proportionally to their social role value (a "High Priests class" would get more "votes" than the "Regular Monks class"), so in that case we need to multiply for E (Ethics Value) in order to consider how much more the High Priests love a religious regime compared to the Regular Monks.

S22 can be worth up to 1.7. Is it normal? Is there a + instead of a minus sign somewhere?

S22 actually has a mistake. Glad you noticed it, Laurent! The problem is incorrect arrengement of parenthesis. S22 should be:
Abs(PP-(ECON*Rexp(2,-8,1-OB)*(0.3+0.7*IND)+(1-ECON)*Rexp(2,-8,1-OB)*(0.3+0.4*IND))

Brief explanation:
S22 represents the evaluation the social class does about the level of Private Property a regime offers.

The more individualistic and the less "obedient" (*) you are, the more you will embrace regimes with high Private Property (ie, regimes where few economic activities are in the hands of the king/State).
But the equation also acknowledges that even if you're very individualistic and very disobedient, you won't push it as far as demanding a PP=100%, unless the "concept" of a completely liberal/capitalistic economy is developed. So, prior to modern times, individualistic and disobedient persons (ethnic groups, really) will embrace regimes with PP up to 70%. Only when the knowledge about how economies function and the ideas of free market and capitalism are fully developed (ECON=1), ethnic groups will be able to support regimes with PP=100%.

That's why the equation has the form Abs(PP - (ECON*A + (1-ECON)*B). A can take a value up to 1 (100%) and B up to 0.7 (70%).

(*) "Obedience" (OB) is one of the Ethnic Groups descriptors and tells us how pyramidal people think society should be. High levels of obedience imply people believe there're a few "illuminated" persons that should run and decide everything because "they know better". In this equation OB is used to make people dislike high levels of Private Property, representing people's belief that the ruler and their staff will probably run the economy better than passing economic activities to a larger group of people (that sure will make a mess with them! ).

where p is combined ethics/warfare/capital power. What does this term mean?

"p" is a meassure of the "elitistic" power a social class has. Since it's computed as E*RPW+W*MPW+K*CPW and since those Political Blocks represent the gathering of social classes with the "special" influence of religion, military and/or capital (as opposed to a representative power), p reflects how "special" the social class is respect to the rest of the people. It helps manage better the different behavior of "elite" social classes.

S3: This one I think I get. If religion is important, ethics having more power is more attractive, same for military, and asceticism/capital (inverted).

Exactly.

I gather from other equations and bits of code that obediance = RExp(-4,14,human) where human is the regime's human(people) power share? Am I right about obedience here?

hmmm, I don't fully understand this....

S5=5*(HPCI/PCI)*SP*Sign(L)
What does PCI do here?

S5 represents how much a workers social class likes the Social Policies level the regime offers, considering the recent evolution of workers' income. PCI is social class' current Per Capita Income, while HPCI is social class' Historical PCI. If HPCI is higher than PCI, then it means the social class has become poorer during the years and then the workers are more enthusiastic about more welfare (higher SP). And vice versa.


I'm going home. Next time I'll comment on data available for the player about social classes preferences.

Cya!

Edit: bas use of italics

[LDiCesare]

Originally posted by Gary Thomas

Rulers in the archaic pre-poll and pre-statistics era had ways of judging the mood of the population, probably more efficient than polls and statistics. Roman Emporers were known to mingle, icognito, with the population. Spanish Kings (and other rulers of that time, including Elizabeth of England) had a very well developed spy system, loyal to them personally. Variants of these methods were widely used. Any ruler who did not at least try to get some feedback on popular feeling tended to be toast when the crunch came.

Cheers

Yes, as you point out they had means to get the information, but that information cost them money. I know there's an internal intelligence level somewhere in the models. What I'd like to know is how much information we want to give (up to what detail: square or province or ?) and how accurate is it, based on intelligence spending (probably a logarithmic funcrion or square root or some such with big diminishing returns).

Rodrigo, the obediance = RExp(-4,14,human) part is extrapolated from other parts of the model where that bit of equation (with maybe L instead of H) occurs here and there in various equations.

I'll be integrating the corrections in the code.

[LDiCesare]

Another point I don't get well:
We have several social roles:
warfare, ethics, capital, worker, human, bureaucracy.
However, in the political power blocks, we have all of these plus ruler minus worker. This means although I share the same object for both representations, I must weed out workers from the political power blocks. Why aren't workers be shown as a political power?
It looks like workers favor democracy rather than a proletariat dictatorship, because this one can't really be modelled if you skip a 'workers' political block (which IRL is represented by labour unions).

[roquijad]

Laurent, here are my thoughts on some of the issues you rose:

About info available to the player:
As you say, eventually we can choose to go as detailed as letting the player know what a social class with a given nationality and religious preference (ie from a given EG) living in a certain square of one province, wants and feels.... Although possible, I'm not sure if that could really be considered useful. The "must have" info IMO is (considering riots+govt+social models):
Civ-wide:
1) The society’s political profile, ie the aggregation of all political blocks. Here we’re telling the player what the social classes want, proportionally to their power shares. In other words, the direction where the govt will be pushed to if social classes would be left alone to decide the govt destiny.
2) The political profile of each political block. If I know what the military want, I can give them something and get’em off my back...
3) The aggregation of PAF values (unhappiness levels) per social class. what’s bothering social classes right now?
4) The aggregation of PAF values per EG (nationality X religion). Are all those ethnicities I’ve conquered calmed enough?
5) Demographic chart pies: by religion, ethnicity (nationality), social class and regimes supported. What religions are there in my empire? What’s the relative importance of each? What kind of govt the people want?

Province:
1) PAF values per social class. Am I going to face protests, independence attempts, etc in this province?
2) PAF values per ethnicity (nationality). Am I going to see Guerrilla Units any time soon in this province for nationalistic reasons?


About a “workers’ political block”
I can explain the absence of such block in two ways:
a) The already existent power blocks (with the exception of People’s PB) represent the privilege of some social classes to influence the govt above the masses. If your society has no privileges for anybody, then those PBs have all zero power and all power is given to People’s PB, where they’re all equally weighted. Introducing a workers PB would imply they could have some privilege over the rest, which hasn’t been the case in RL. (interpreting those powers as privileges is useful for practical terms - see below)

b) If we add a workers PB, then we have to get rid of the People’s PB because it becomes redundant (you’d have one PB for each social role). Although this could be a possible modelling strategy, it rises a problem: in representative forms of govt, each PB should get as much power as its demographic representation. Since the actual demographic distribution of social classes changes from civ to civ, and as time passes by (one civ may have a larger middle class than other, FE, and it could growth or shrink in time), then the power distribution among PBs in representative forms of govt will change, even if your political system remains stable. This will confuse players. If the Capitalist PB increases its power, is it because they increased their demographic share or are they just getting more power than they should and my civ is actually becoming some sort of oligarchy? Having a People’s PB instead of a workers PB simplifies the scenery a lot: as long as your govt is representative, you should see power shares of all other PBs at zero. If someone is getting some sort of privilege on govt, the corresponding PB will have a non-zero value. One look at your power structure and you’ll know exactly how representative your govt is and if it’s not, who’s having special powers.


See you!

[LDiCesare]

Tanks for the explanation on workers. Convincing enough.
I think the information detail you want are good, we will need some UI design before we can show all this.

One more point about the equation:
Tax Rate=(((p-K*CPW)/3)+0.4*Min(SP , (1-(IND*(1-p)+p)))+AGG*0.1)*(1-FE*FR)

This leads to huge figures like 0.18 or 0.35 in the Delenda scenario.
There's mostly no discrimination to consider here, so the last parameter is mostly 1.
For parameters around p = 0.6, SP = 0.35, IND = 0.2, AGG = 0.8, K or CPW = 0, you get desired tax rates of:
0.2 + 0.4 * Min(0.35,0.26)+0.08 = 0.2 + 0.4*0.26 + 0.08 = 0.384.
This is huge according to what Mark expects (we currently had tax rates <= 0.1).
Of course, you get a lower total value for groups with lower p, but given the regime, they tend to have lower political powers, so we end up with huge tax rates.

The equation looks a bit wrong to me because:
Since: 1 - (IND*(1-p)+p) = (1-p)*(1 -IND).
The formula can lead to max values when CPW = 0, IND = 0, FE = FR = 0, AGG = 1, SP = 1. This gives an equation of:
p/3 + 0.4* (1-p) + 0.1 = 0.5 - p*(0.067).
So the max value is 0.5. I think this should probably be scaled down. I'd like to hear from Marl about the max tax rate the econ model can allow.
We can either rescale it down (i.e. multiply the result 0.2 if we want a max tax of 0.1 whatever the regime) - I think it's a bit extreme as 0.5 looks good as an extreme.
Or we can apply a function to lower the values, like rescaling to a 0-1 scale, squaring, and rescaling to the correct range. This means a 0.38 value out of 0.5 would become only 0,29.

[Mark_Everson]

Hi All:

On the subject of tax rate determination and limitations. . .

The limitations in D7.2 and before that tax rate needed to be <=10% were due to current simplifications in the model. This was done because in classical antiquity production is mostly food for use in subsistence, and higher tax rates would end up starving the people. Tuning of the model may make higher tax rates possible. I suspect that if we put a PCI (per-capita income) contribution into the tax forumula in an appropriate way, it can ensure that reasonable tax rates result.

Rodrigo, if you want to attempt to make the changes, the number of units of food per 1000 population that are needed for subsitence is 4. The absolute max tax rate allowed should be (PCI-4)/PCI Minimum PCI usually starts at 5. This should result in a min tax rate of 20%. However there may be other limitations that make achieving this max level a problem for the player. I will think about this further at some point. Perhaps it would be safe to make the max value something like .75*(PCI-4)/PCI in the interim.

Another wrinkle is that the tax rates even in classical antiquity can be much higher than the max I cite provided that the govt then redistributes the taxed food back to the people. IIRC tax rates in ancient Egypt were close to 50%. I'm not sure how exactly we should handle this, but it should be workable without huge problems.

[roquijad]

I admit I'm ignorant to say if a tax rate of 35% is huge or not. I recall reading in the newspaper (when a debate on rising taxes here in Chile was taking place) that, considering all taxes from all sources, the average US citizen payed something more than 30% of his/her income. And that in several countries in Europe, the figure was something like 45%. If so, I imagine taxes in antiquity were even higher. But yes, I might be remembering it wrongly and also assuming nonsenses about antiquity.

Bottom line: I suggest we use our current equation now, with a scaling factor on front. Fix that factor's value in whatever number the econ model needs currently. We can change the scaling factor later if we manage to gather historical info about tax values. And even later we can change the equation if it doesn't give us reasonable results. Agreed?

[Mark_Everson]

Originally posted by roquijad
Bottom line: I suggest we use our current equation now, with a scaling factor on front. Fix that factor's value in whatever number the econ model needs currently. We can change the scaling factor later if we manage to gather historical info about tax values. And even later we can change the equation if it doesn't give us reasonable results. Agreed?

Works for me! Maybe starting with a multiplier of 1/3 or so would be good for now.

[LDiCesare]

I'll start by putting a scaling factor of 1/3.
I think taxes of about 30% are perfectly legit for modern day countries. I just don't know the effect of such high rates on older economies. Sparta had such a system that I wonder how you'd model taxes there anyway.

[LDiCesare]

Here's a new mock-up of user interface.
The bar on the bottom with placeholder art (obviously) would be a bar of buttons. The leftmost one is the ruler button, which lets the player modify the sliders he can act upon.
The other buttons would replace the sliders by a panel showing the military (selected button), ethics, capital, human and bureaucracy desired government policies. I might even add a 'total' profile for desited views.
This would let the player click the ruler button and browse from what he can change to what the various political power blocks want in the same window.
The placeholder art buttons could be icons or faces of advisors. If advisor faces, it would be easy to have them evolve depending on techs.
The panel shown on top would change to show the desired regime and negotiated policies, along with the support shares for various regimes. We might even add some room for 'advice' like the advisors did in civ (though of course, the advice was often poor, and with no ai to give it it wouldn't go vary far).