Government Model v. 3.1

[LDiCesare]

Okay so a bit of a coding status here. In a few words:

It mostly works.

I got the government model hooked in to the point it's usable. The only things I'm not sure about are the one equation which could yield < 0 result (I arbitrarily added 1 to prevent that) (civil rights). The results seem quite correct with that fix.
One thing I didn't code is the dominant religion stuff when there is no state religion. This means that without state religion, noone wants religious discrimination, which is obviously wrong, but I couldn't find how to compute simply the dominant religion and in the early game there won't be secular states anyway.
I didn't meddle with regimes for the moment.
Taking Rodrigo's govt technical doc, here's more or less what is coded:

I. Computing Real Power Distribution:
Not coded- It's a constant fixed in the scenario files.
II. Creating the Conservative Govt Profile:
Not coded (not yet used).
III. Computing the 5 Political Powers shares for a social class:
Mostly coded. I will be adding the discrimination factors today. Province isolation is not for the foreseeable future.
IV. Computing preferences for DNPs for a social class:
Done.
V. Creating the Support Regimes Shares for a social class:
No regime stuff coded for now.
VI. Creating the 5 Political Block Profiles:
Done.
VII. Creating the Society's Govt Profile:
Done.
VIII. Creating the Future Govt Profile:
Done
IX. Creating the Current Govt Profile:
Currently, the future govt profile is applied immediately. (I am not sure how we can show the information to the player correctly, or if we have to.)
X. Determining the demographic share for a social class:
Specified in the scenario files.
XI. Determining Knowledge and Development levels for a Regime:
Not coded nor used. It will require the tech model to show many more techs and applications before coding this becomes worthwhile.
XII. Determining the Govt-Regime Distance and the Ruler-Regime Distance for a Regime:
Regimes not done.
XIII. Determining Social Class' Historical PCI:
Not done.
XIV. Updating Empire's Stability:
Not done. Depends on the length of the game, which I don't feel confident with for the moment.
XV. Determining Optimum Institutional Infrastructure (OII) needed in a province:
Not done. I do'nt know what this one is for.
XVI. Determining the Number of Administrators Needed (NAN) in a province:
Not done. Again, don't know what it's for.
XVII. Determining Administration Effectiveness Level (AEL) for a province:
Not done. What for?
XVIII. Determining the Maximum Number of Feudal Units (MNFU) in a province:
Not done. This verges on the riots model, and more importantly it requires some stuff to handle feudal units in the diplomacy/military code, which is probably not top priority.
XIX. Determining Govt's Official Religion:
This is specified in the scenario files for now. This means the official religion can't change for the moment. I have to know what dominant religion is for these purposes, and since state religion and religion with the most followers aren't always the same (think Kashmir), I believe there are things to do when computing the dominant religion which may not be straightforward; Maybe multiplying the number of followers of nondominant religion by ReligiousDiscrimination when computing the total number of followers?

With that big list of 'not done', you may think I am talking nonsense when saying it mostly works, but I do say this. Regimes are a refinement which probably require tech to work, and there are hooks to the tech model and length of game I don't feel comfortable coding until we have a tech tree and some length of the game.

I must say that the fact that desired policies depend on the culture of each ethnic group and then on the social classes inside the group will make for a really fun game, as the various ethnic groups want different policies, and this results in a lot of different desires from different ethnic groups, which with ethnic riotting and bad policies riotting add some spice to the game.
Congrats Rodrigo.

[Lord God Jinnai]

Well any amount of progress is going in the right direction

[Mark_Everson]

Originally posted by LDiCesare
Okay so a bit of a coding status here. In a few words:

It mostly works.

Great, Laurent! Thanks for all the hard work! I agree that what you have so far is a big improvement in terms of the playing feel of the game.

XV. Determining Optimum Institutional Infrastructure (OII) needed in a province:
Not done. I do'nt know what this one is for.

Determines amount of govt infrastructure needed to run the province: local administrative buildings; court houses, etc. This is an input to the effectiveness level (AEL) I think, as is NAN.

XVII. Determining Administration Effectiveness Level (AEL) for a province:
Not done. What for?

Things that depend on central govt control in a province, like tax collection and riot likelihood, will be affected by AEL.

[LDiCesare]

Regimes

I'm tackling the regimes now, and I have a few questions about the equations and the model.

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.

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.

3) Details about the equations for support regimes:
3.1)
S1=20*IND*(p+Sign(L)*PPW+A*BURPW)
Why Sign(L) when everything else usesvalue * powershare? Coded anyway.

3.2)
S2=-10*(1-ASC)*(Abs(SP-(1-(IND*(1-p)+p))) + 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)))
Argg... I rewrite this as:
-10 * (1-ASC) * (abs(S21) + abs(s22)) where:
s21 = SP + (p-1) * (1-IND) where p is combined ethics/warfare/capital power. What does this term mean?
s22 = (Private Property) + 0.3 + 0.4*ind - econ*f(OB)*(ind*1.1 + 0.6)
It's exactly the same, but simplified a bit. Still a big murky thing for me. Can be worth up to 1.7. Is it normal? Is there a + instead of a minus sign somewhere?

3.3)
S3=-30*(Abs(Rexp(3,-4,IR)-RPW)+Abs(Rexp(6,-7,AGG)-MPW)+Abs(Rexp(6,-6,(1-ASC)*IND)-CPW))
This one I think I get. If religion is important, ethics having more power is more attractive, same for military, and asceticism/capital (inverted).

3.4)
S4=-30*Abs(Rexp(4,-6,1-OB)-PPW)
Same as above except obediance/ruler. 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?

3.5)
S5=5*(HPCI/PCI)*SP*Sign(L)
What does PCI do here?

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

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? What name can I put on functions that -retrieve all regimes and -retrieve all regimes except ruler and conservative?

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

[wormbyte]

Here is something completely different but relevant. Its called the "International Interaction Game". I post the follwing text under the "Dplomatic UI" thread but I feel it might be worth mentioning here.


Here is what I said on the Diplomatic UI thread.....

-------------------------------------------------------------------

Has anyone heard of the Internal Interaction Game. I believe is was first published in a book called "War & Peace".

Check out the attached GIF of an Extended Tree form of the International Interaction Game.

It works best under uncertainty. Basically you have to reach an end node to complete the game. Each end node yields a utility payoff, some are higher than others. The value of the expected utility is equal to "Benefit value - Cost Value". Naturally a player wants to get to the node that give the best utility. But at the same time the player has to consider what the other player will choose at the next step.

The other crucial factors are "Preferences" and "Perceptions".

Each player has their own preference of the preferred end result of the game. Every player has a perception of what the other player preferences are. But nobody knows for sure and thus uncertainty and risk kicks in.

For example...

Player A chooses to make a demand. He does not expect player B to acquiesce and fully expects player B to make a counter-demand. Player B does indeed make a counter demand. Player A chose this route because he expects player B to then negotiate.

But then he finds his perceptions on this were incorrect. Instead Player B attacks Player A because Player B atcually thinks that his negotiating powers would not yield a high utility return and he would come out alot worse on the negotiating table. Player B also believes that Player A does not want a war and therefore will capitulate to Player B's demands.

Not the case though. Because Player A has a higher utility associated with the "War Started by B" then to the "A Capitulates to B" so Player A retaliates.

End result WAR. A outcome that neither side wanted by ending getting because the misunderstood the preferences of its opponent.

-----------------------------------------------------------------------

The reason I have posted it here is because the utility payoffs would be directly linked to the Government model.

[LDiCesare]

I'm trying to find how to show the regimes to the player and how the choice of a regime works.
I have several gripes, beginning with 'how do we show this to the player' and ending with 'how should that be':

The current user interface is a bit misleading because regimes don't appear there, and half of the information shown is about regimes. If you look at the screenshot on the previous page (http://apolyton.net/forums/showthrea...0&pagenumber=2), you can see that preferred social roles (power distribution) is on the left while regime policies is on the right, below the negotiated policies. Although it seems consistent to show all policies in one column, they belong with the other regime stuff.
Now, this 'regime' stuff is shown and can be acted upon, but, while negotiated policies are negotiated one by one by the various power blocks, the regime thing is negotiated as a whole, and has meaningful names like 'Monarchy', 'Despotism', etc. plus the ruler's regime. But these names show nowhere. Worse, even if the player's regime matches exactly an existing regime, there's no way I can show the regime name, nor is there a chance for the player to actually know he has picked the 'Banana Republic' regime.

So the model lets people find out regimes based on known regimes (probably obtained through tech), but lets the ruler pick any regime he wants. The regime would tend to revert to a known regime, except that people can support the current regime. The flavor of regime names is lost if the player doesn't pick their regime out of a list.
All in all, I feel rather uncomfortable with the regime part of the model. If I can pick any figure, the existing regimes are just meaningless names that can't be shown. The player can't know what regime they're in. The only time when a regime name appears is if people revolt, when they will say 'Down with the ruler's regime, let's have a Monarchy'. Next turn, the Monarchy is tweaked again towards the ruler's preferences and is no longer a Monarchy.
So I'd rather have the player pick their regime among a list of existing regimes, and be allowed to tweak the default regime values by a set number depending on the ruler's power. For instance in a despotism, the player can choose any value for private property/social policy, and that's still a despotism. In a monarchy, they can tweak less. If you want to change the value more, you have to pick another regime. This way, regimes would be more meaningful because I could show the name of the current regime in the panel.

[Mark_Everson]

Hi Laurent:

Some good points. . .

Originally posted by LDiCesare
The current user interface is a bit misleading because regimes don't appear there
(snip)
the regime thing is negotiated as a whole, and has meaningful names like 'Monarchy', 'Despotism', etc. plus the ruler's regime. But these names show nowhere.

This is a very good criticism. I think we simply need to figure out how to show the "effective regime" to the player. FE if the state shares primarily the components of Monarchy and Democracy, we should call it a "Constitutional Monarchy". Or perhaps we should call it "Monarchy-Democracy". If the effective regime is within a certain "distance" (multi-dimensional space of course) of Monarchy we just call it that. As it slide we start using the hyphenated name. This may not be the best solution to the issue, but I think something like that will give the player the flavor we want.

We can also use a similar device to show the ruler's regime. If I am the player and want to wield all the power I have ruthlessly I can have a ruler regime of "Despotism" even in a democracy. If the democracy is stable that discordance should cause the ruler big problems, but the ruler should not be able to significantly get their way.

I think these proposals fix most of your complaints, but I could be misunderstanding you. Anyway that's all for now, until I hear yes or no.

[LDiCesare]

One of my gripes is:
How come in the 5th century BC you could be able to issue a ruler regime of 'Communism'?
I'm not sure letting the player total freedom from the start about regime policies is a good thing. Discovering new regimes will be of little direct interest to players in such a case, and they will feel that discovering a new regime just changes the amount of support they get from the population, but that doesn't feel like a revolution.

Dividing the space into names is feasible, but in a multidimensional space, some names may seem strange: Democracy-Despotism may be impossible, but even Democracy-Monarchy doesn't sound as good as Constitutional Monarchy, which would probably be a regime unto itself, so we could end up with a Democracy-Constitutional Monarchy or Despotism-Constitutional Monarchy, which sound weird.

[Gary Thomas]

Surely civlizations do not change regime unless a traumatic event occurs. The idea of some sort of over-riding civilization avatar (read "player") decides to switch a feudal kingdom to a fascist state (whatever that might be) is nothing less than bizarre, and could only happen in Civ3 or its predecessors.

At least Civ3 has the characteristic that when a new "regime" (they call it type of government) becomes available they can choose to change.

If the government model is trying to plan how to present this kind of thing to the player, then I cannot help but feel that it has drifted very far from reality.

A realistic model would present the kind of events that could lead to a change of government type, and present a possible change in those terms.

To use "regime" as a synonym for "type of government" does seem to me to be something of a contortion of common English usage. A change of regime is when Stalin dies and Khrushchev takes over. Is this a change in government type? A change in behaviour perhaps.

In essence, allowing the player to change this particular characteristic of a civilization is not a step forward.

On the other hand, it is not an area that I have given a great deal of thought.

Most of my thinking about the game relates to the coding, as I suppose might be evident. Whenever someone proposes somenew idea, my first reaction is "How can this be coded?".

Most of the remainder is related to how the game plays. "Does this have a measurable effect on the game?" "Will it confuse players?" and so forth.

The little bit that is left is devoted to the questions "Is this realistic?" or "Does it make sense?"

At that point the regime proposals grate. The idea of the player of Sumerian civilization making the decision to change the "regime" of the government to a hybrid of Tibetan Lamaisn and Classical Democracy with Chattel Slavery is so peculiar that I immediately get back to thinking about coding.

Cheers

[Max Sinister]

Hybrid governments... well, I think they could be relatively easily coded as tuples of numbers (90% democracy, 40% social state, 100% capitalism, 30% religious, 70% militaristic...), but the bigger question is: What does that mean for gameplay?

[Lord God Jinnai]

Nothing, other than if we want to assign a name..

[Max Sinister]

Maybe "Gameplay" wasn't the right word, but what does it mean during a game if I change "capitalism" from 60% to 70%? What does it change? More income from taxes / trade routes / luxuries?

[LDiCesare]

I'll give a brief explanation here of how the govt model works:
The player decides of a set of policies. He wants for instance that the power be divided evenly between the ruler (50%) and the people (50%), that slavery should be abolished (0%) and that private property should be illegal (0%). Meanwhile, the existing government structure says that the ruler only has 10% of the power and that 90 remaining % are at the hands of aristocrats, who want 100% slavery and the current policy is 100% private property.
There is a negotiation system that takes place, which will lead to slavery (being a directly negotiated policy) having a value of 90% (weighted average). The other values (power distribution and private property) are part of a "regime" (used to be called ideology). Instead of directly negotiating either one, each social class (here the aristocrats, who are the sole in power) decides it prefers a regime or another one, which means that distribution of powers, private property and social policy are viewed as a whole by the social classes in power.
So the player can try to alter the existing regime policies, but it depends on which regimes the people support. There may be a revolution in some cases (riots model), promoting a change of regime.

If this regime stuff showed something for the player, I would code it now. But as it is, I have a lot of trouble understanding what it gives in terms of game play. When a civ discovers the regime 'communism', the immediate effect is that maybe a civilization will support this regime instead of another one, resulting in a change of the power shares and policies. But the name doesn't show. When you add to this the fact that the current regime and the policies the ruler wants are also considered by each social class, it means that the result of a change of the player's desired policies will be very hard to predict other than by trial and error. The flavour given by the name 'Monarchy' or some such is also lost. The information is in there, but I can't see how to show it.

So the current system models quite properly the changes of the government inside a given regime, but the effects of a change of tech (discovery of a new regime) is quite weird. It can either promote a 'soft' revolution (people start supporting a new regime) or cause a revolution through riots. But knowing what the current government is is hard.
I'd like to be able to
1- know what kind of government this is without looking at a half dozen figures (not currently possible)
2- show what types of governments are supported by the population (possible)
3- show the effects that the player's desired policy will have on the actual policies and on the riot/revolution chances. (the former is possible, but hard to explain, the latter is very hard to show or even get an idea of except through point 2 above)
4- when a new regime/ideology is discovered, show the consequences of this discovery (possible but hard with the current model - the figures would be subtly altered and it's hard to track the root causes of the change, as knowledge of the regime grows over time and doesn't appear immediately in the whole population).

So my unadressed problems are points 1 and 4.
1 would lead to weird names with the curent system where the player can provide any input. Note the weird names can also appear during revolution if the ruler has little power and the people want to revolt towards the ruler's desired regime policies. I can bypass the problem by limiting the choices the player has to a window around any one of the known regimes.
4 is just a lack of effect when a new tech is found. Although the government model may be realistic here, the point is that the discovery doesn't provide anything to the player: no additional choices available, only different effects when some sliders are changed. That's again why I'd like to limit the available policies to those known, because it gives new choices to the player when a new regime type is discovered.

[Mark_Everson]

Hi All:

First, in responding to Gary, I believe the model prevents all the sillinesses in civ games that are cited in the post. It's certainly possible there are things that are a bit weird, but I'm confident that we can ID them in playtesting and figure ways around any that might exist. The only govt form where the player can change things at will is in moving from a state of near-total control -- basically a despotism. And that makes sense to me in RW terms.

On a side note, I can't remember why Rodrigo wanted to change the name from ideology to regime. BTW I've emailed him to let him know about this thread.

On Laurent's latest post. . .
I agree #1 is a potential issue, but think that my proposed solution above somewhat covers it. I know that you may not be able to get one or two regimes that encapsulate all the settings, but think it is a pretty good shorthand. If the player is that concerned with the details, then they can look at the specifics IMO. I really don't like the idea of arbitrarily restricting the player regime to a range around the defined ones. Such an approach would, I expect, hugely frustrate me as a player.

We need to do #4 anyway for the AI. I think it can be done with simulation based on a crude abstraction of what the civ looks like extrapolated say 20 turns into the future. I think players will get a hang of the effects of different regimes appearing after the first few games they play anyway, so I think this concern of yours may be less important than the others. YMMV.

[LDiCesare]

#4 is more of a gameplay issue. I would like the discovery of a new regime to enable the player to do more things. If you would find it frustrating, then I guess we have to code it some way and see how it feels.