We need immediate help in the 'bug reports' thread on this exact issue.

I don't know what you mean by "successful result"
---------
"Another thing I really didn't understand why things must alwasys be negotiated with for a medium, why it can't be yes/no on changing with no middle ground."

quote: If I want to raise a value in a negotiated policy, I have to 'high-ball' the pref choice -- ask for way more than I want to get what I want. It's kinda wierd, and wasn't much fun. It kinda seemed like a loophole that allows you to manipulate the system.

In our initial discussions, I had pointed out the problem. The quotation is from the part of the model referring to CNPs.

quote: Rodrigo: For every variable discussed above we have what each class wants. The ruler/player has given already his opinion putting the values he thinks are best in a proper interface (ruler's govt profile), so the PC has values for each variable too. The final policies will be a weighted sum of the desires of each class, in which weights are their pol.power shares. axi: There is a MAJOR problem here. Because of the weighted sum, the player will be compelled to support more extreme policies than these that he intends, in order to tip the scales to his favor. You might call it a feature - I call it a bug, because such dishonesty would never work in real politics. Your example below is most enlightening: ------------------------------------------------------------------- Example: Suppose the political structure is RC-10%, MC-5%, PC-25%, UC-40%, LC-20%. Suppose values classes want are: UC:2 LC:2 PC:1 RC:0 MC:2 Now each class use its pol.power to try to make the govt. policy equal to its desire. The calculation is: (RC) 10%*0 + (MC) 5%*2 + (UC) 40%*2 + (LC) 20%*2 + (PC) 25%*1 + =1.55. Since the variable can take only integer values, rounded we get "2", so negotiations at the govt have given a Religious Tolerance equal to "2" (all religions allowed). ------------------------------------------------------------------- axi: The PC here voted "1" and he got "2". If he were smarter, he would have voted "0", so his share would be 25%*0 and the final outcome 1.30 rounded to 1. Then he would get what he actually wanted. One way to adress this is to penalise unpopular suggestions with anti feeling too and not just policies, but this would complicate things too much. Another is to modify the decision making system, introducing two rounds, (like in the municipal elections): Round 1: Distribute the political power between the choices. The best two get to the second round. Round 2: Throw each of the other percentages to the closest finalist. If one is in the middle, split. Majority wins. Rodrigo: You got me here. I knew this effect was in there and in fact we discussed it with Mark. Mark didn't care much about it and gave me a "we'll see", so I kept going, but I do dislike it. There's really no way in which we can force the player to put exactly what he wants instead of strategically choosing values. My negotiation procedure cannot fix that nor can yours. In this particular example your procedure might seem better, but in the general case the player will always be able to choose values he doesn't really want no matter the procedure used. I think the best way to solve it is, as you said above (and which is also what I proposed to Mark), including the ruler's values in the Anti feeling. In that way the player cannot go for extreme values without risking to face problems. As for the procedure to use, I prefer mine. The reason is your 2-rounds election mechanism is an election procedure, not a negotiation one. In an election procedure, the final value for any policy is one of the initial values in the competition since the procedure only chooses one of them. If we are to simulate the politics within a govt, then the final value should be the outcome of negotiations and so this final value is not necessarily one of the initial ones but a new negotiated one. If we simulate elections instead of negotiations, we'll be strongly reducing the possible values for each policy and this is a big problem for continuos variables like Foreign Affairs. The other important problem with your method is that if the ruler has more than 50% pol.power, then the final policy value will always be exactly what he wants, so it's irrelevant how much power the other classes have. If the rest classes have a total of 0% pol.power or 49% pol.power is unimportant and in both cases the govt policy takes the ruler's value EXACTLY, which I find very unrealistic. If the ruler has 51% pol.power and wants a policy value equal to 1 and the rest classes (with 49% pol.power) want a value of 3, I think it's way more realistic the final value being around 2, than the ruler imposing his view. Having an election procedure destroys any attempt to simulate actual negotiations, which I believe are extremely important to exist in the model. So, IMO, we should keep my negotiation procedure and just see how to push the player to be honest. MS! axi: I only suggested an election system here because I did the same for the INPs (although I actually said the opposite), but there there are some more reasons for it. I aknowledge that here it would probably do more harm than good. If the "penalising the profile" solution is feasible, I' m all for it. Solved!

In the govt model thread, in my first comment, I had said:

quote: 7) The Macchiavelian ruler: I have already pointed out that the weakness of the negotiation procedure is that it lets the ruler counterbalance the other classes preferences by taking some extreme decisions for the ruler's gov't profile. The things are bad enough for the DNPs, but for the INPs they are worse because of the multiple loops of the procedure. The first thing I did with Rodrigo's worksheet is to put 100% for Ruler's pol_power in the Ruler's gov't profile and the macro gave me exactly what I wanted. If this happened in the game, it would be exactly like playing civ2 after this. The obvious solution for this is to make sure that the player will not dare to give extreme RGP's, because that would cause an immediate raise to all the related PAFs and an immediate events check and would normally cost him his head (at least). This is something to be set at the riots model of course, but I have outlined it here because otherwise the whole negotiation is useless. It could also be possible for us to restrict the negotiation to a few or only one loop per turn, so that any governmental change wouldn't take place immediately. Rodrigo insists that an equilibrium must always be reached, but if this prooves to be in any way buggy or unnatural, this could save our efforts.

Rodrigo's reply:

quote: The Macchiavelian ruler: As you perfectly said it, Axi, the ruler can face a lot of problems in the Riots model if he chooses extreme values in Ruler's Govt Profile. Even more, there's a specific Pro-Action Feeling for this, the "Replace Ruler Feeling". Concerning the equilibrium point of the Negotiation Procedure, if you didn't notice, I actually chose your idea! Whenever nagotiations are called, the equilibrium will be found, but not applied immediately. It will be stored and each game turn, the Govt Profile will slowly move toward eq. I think these two elements will prevent the player from being dishonest.

quote: Rodrigo1: Your system is very interesting, but I think it's biased towards a modern democracy where elections happen and the ruler has to try to keep himself in charge (be re-elected). The real problem with it is explained in the next paragraph, but let me give you a comment before that: I understand why you took this road. As you said it, it should be a struggle between ideologies instead of classes and also merging ideologies should be more limited to ideologies really similar to one and other. About the first thing and as I said, there's no real difference between both views. However, I admit it's easier to understand the model seen as an ideologies struggle, so let's keep that view instead of mine for explaining purposes. About the second, the main explanation for allowing merging is given in the next paragraph, but maybe we can be more exigent with ideologies and impose some similarity level at least, but we have to be very clear about what each variable represent. I say this because it seems to me you are not understanding pol.power as I do. In your system an ideology having pol.power>0 can be left out of the final govt computation for policies, while I think this is fundamentally incorrect. If you have pol.power, then it means you can take part in govt decisions (you DO have power to do so, by definition). My guess is you treated pol.power more like "political support" or "political strength", so being low can let the ideology out of the govt, but that isn't the role of this variable. If your calculations for defining the winning ideology would have used only the Representation matrix and DIAC (but not pol.power), then at least all variables are being used in their right sense. We have to respect each variable role. The main problem with your system is that you're simulating only the "executive" power, that is, what ideology (or ideologies coalition) wins "the presidency" (or equivalent for other govt types), while I was trying to simulate the behavior and interactions between the executive power and "the senate" or any other equivalent for less representative govts. When I say "govt" I mean the whole political scenario and not only who's in charge of administration. That's way you feel is unrealistic for different ideologies to be merged to determine govt policies and I guess that's why you saw pol.power not as power to change govt numbers but as something else. I agree there's little room for merging ideologies in an election competing for "the presidency", FE, but this is different because I don't want to know which ideology won the elections (became the executive power), but the complete struggle between the executive power and the rest powers. If we were simulating which ideology was going to be elected as the executive power, then I could have agreed with you, but that's not the case and I still believe that shouldn't be the case anyway. Clinton and his ideology, once elected, doesn't decide all govt policies, but he and his party has to deal with other parties in the senate for any fundamental policy change (law, if you will). So, policies are determined even by those not elected in the executive power. So, what I think we need is to simulate the whole political game including interactions between the executive power (PC) and the "senate". And the senate can and should be a mix between all ideologies present and the power each ideology has over final numbers is the total pol.power the ideology supporters have (mod by Representation). No elections are needed and any coalition effect is included indirectly in the negotiation procedure. If you think we should ALSO simulate the election of the executive power (PC), then that's another matter. However, you have to realize that doing so is equivalent to define PC ideology, so that's like taking away any control from the player and therefore we have to be really careful about it. MS! --------------------------------------------- axi1: I am obliged to accept your reasoning, although different from mine, because of the different definitions that each of us (subconsciously at first) gave to political power, political struggle and the PC. Some points need to be clarified: 1. Under your definition, where does political power derive from, if not from the capability of each class to control the legislative and executive power? How can a class alter state policies like the amount of welfare, if not from within the government? For me, executive power and policies are the same thing, while the PC can be regarded as the incorporeal "Kratos", or more conveniently as the personal will of the person or personae moving the threads of power, directly or through the people. --------------------------------------------- Rodrigo2: I really don't know what to say here. Your questions are as far as I can see answered already. I'm afraid we're just having problems with definitions. The capability to change govt policies (at least to influence their value) is called political power in this model. That's the answer for the first question. If a class has pol.power then it's within the govt. That's the answer for the second question. "Government" in this model refers to all political institutions that take part in govt policies definition. PC is nothing but the ruler or if you prefer, the ruler plus a set of fellows who work with him to try making his ideology applied. --------------------------------------------- axi1: 2. It is commonplace among the pluralistic types of government, that, although everyone recognised as a political agent has the right to get his view represented in the government, not all of them get represented simultaneously, because, as it is claimed, a) there are differences that are generally considered unbridgeable and b) a political synthesis can sometimes be far worse than a pure ideology. So what happens is some variation of the majority rule, otherwise referred at as "democratic concentrisation" (directly translated). I will not discuss here if this is right or wrong, only that it is common practice and that it should be represented. In autocratic regimes of course, all of this is irrelevant. ---------------------------------------------- Rodrigo2: The question is how you include within one model these types of expressions of democracy with the negotiations that takes place in ancient regimes between the great military leader and the chief priest. For almost all history no democratic mechanism like those you mention existed, so I rather simulate a negotiation procedure than an election one. Since I don't want to include an "exception rule" for democratic regimes, I'll keep the same procedure for it. If the final values turned out to be so bad for everybody, then the Anti feeling will come to our aid pushing the govt to a more radical change. I prefer to make the Anti feeling more sensitive to the difference between what's desired for policies and what exist than to try to simulate individually each type of political struggle for each type of govt form. In that way you can ensure very little mixing will exist because the Anti feeling won't let it happen without suffering revolutions and other nasty events. ---------------------------------------------- axi1: 3. From a totally technocratic viewpoint, "majority rule" is the most natural way of avoiding the merging and so giving a chance to extreme policies of ever being practiced. --------------------------------------------- Rodrigo2: But totally unrealistic. The illusion of democratic elections is just that: an illusion. You don't just step aside if you're capitalistic and a communist govt wins the elections or vice versa. No extreme ideologies have ever reach govt by normal elections and without blood in the process (including the attempt to actually rule). There's a simple reason for that: people with extreme ideologies are too few. If a extreme ideology is pursued by the player, he'll do exactly what you mention below that's also historically accurate. ---------------------------------------------- axi1: We should not fool ourselves; the player is bound to do anything in order to get his civ under control, so we wouldn't want to force him do things that he normally shouldn't. Because to achieve this with total merging, you'd have to suppress the opposers political power to zip, giving it all to the conceding parties and this everytime you wish a change in policy, while this is definitely not the meaning of political power. (This reminds of some political bureaus of socialist regimes, where the consensus had to be unanimus, so no opposing voice was practically allowed.) ---------------------------------------------- Rodrigo2: You have just outlined the steps to follow for a player with the ambition of having an extreme ideology in the govt. You just missed some killings in between. Of course that's not the meaning of pol.power and that's fine. If you're afraid the player will do all that always to keep his civ as he wants it, you're (again) forgetting about the Anti-feeling. He just cannot do that all the time. He'll have to face too many riots and such. In no way I've trusted the players to be politically correct and use this model as good human beings. The temptation to be dictator is huge, I know. Just let the player do whatever he thinks is best, but let's see how he deals with the mess. ----------------------------------------------- axi1: Majority rule also allows for more and easier fluctuations in policy, while rendering the player willing to let two or more opposing ideologies to coexist. I have the general feeling that it provides better gameplay. However, as you have already pointed out, we must respect the variables, so my approach can only be accepted if we are willing to redifine them. I am afraid that this is a rather philosophical matter, about which everybody has extremely vague ideas. I am also afraid that a discussion in the forum about it would really get chaotic. Feeling that there are other, way more important issues to settle, I will not (I am not entitled to anyway) veto your approach. I am only concerned about the implications of #3, which will be practical. Solved? ---------------------------------------------- Rodrigo2: You have to give more attention to the Anti-feeling. Two antagonistic ideologies won't coexist peacefully if they both have a large support. They can if one has very little support, like the communist party in any first world country today. And IMO it's reasonable to say that this minority ideologies can affect policies (which what my negotiation procedure allows). Just look how these minority parties conform alliances to be able to change laws to what they consider better. They find a way to make some little changes at least, which is well modeled through little pol.power (derived from little population support). The world doesn't work changing from one govt form to the next through majority rules. Such radical changes occur only through violence. History is my witness. The player will need to be despotic several times and allow representative govt forms only if his people is homogeneous enough, which in time can be achieved through propaganda, murders, forced migrations, etc. It's not my fault. It's just the way it is. This is why the role of the Anti-feeling is very important and you shouldn't oversee it. It's a part of the political process.

These are the four competing profiles. To make things easy for me, I suppose that the current values match the de facto values for the power structure.

code:

---

VARIABLE	             Ideo1     Ideo2     PCIdeo    Current Values
PC pol.power 	20%	40%	20%	15%
UC pol.power	25%	10%	25%	10%
LC pol.power	30%	5%	10%	25%
WC pol.power	15%	15%	15%	30%
RC pol.power	10%	30%	30%	20%
Private Property	60%	40%	60%	50%
Economic Planning	60%	30%	70%	50%
Social Policies	80%	20%	60%	50%

---

code:

---

CLASS    Ideology 1            Ideology 2            PC Ideology          Current Values
PC         0%x15%               0%x15%              100%x15%           0%x15%
UC         26%x10%             38% x10%            22% x10%           14% x10%
LC         81% x25%            8%x25%               11% x25%           0% x25%
WC        72% x30%            13% x30%            14% x30%           1% x30%
RC         54%x20%             23%x20%              20%x20%           3%x20%
SUM       55,25%x25%        14,30%x100%       28,15%x110%       3,30%x100%
FINAL     22,14%                22,92%               49,65%                5,29%

---

With these final percentages, the profile variables will be negotiated. These resemble very much the results of the real world elections and should IMO be displayed on the interface, in this format. In this case, Ideology 1 has by far the largest support, but it is vastly misrepresented, which, along with a little propaganda, makes the PC ideology dominant.

Using these results, negotiations can be done in the usual way, through a weighted average, the negotiations procedure or whatever. Even in this case though, the problem that we encountered with the Culturally Negotiated Policies persists: the player is driven to dishonesty, concerning the ruler's govt profile. Of course in the INPs the PC Ideology loses support as it goes more extreme. For this reason and also because of the merging effect, I am proposing an alternative: we can use an election - type system like the one I proposed for the Culturally Negotiated Policies. In this case, the choices are the 4 (or more) Ideologies and coalitions occur with ideological distance as a criterion.

***From now on, things are different from the govt model.***

The formula that computes ideologic distance between two ideologies is simple, since we have to compare vectors with 8 elements, which have the same range (0-1):
D(Ia, Ib)=sqrt(2)*sqrt[(Ia1-Ib1)^2+(Ia2-Ib2)^2+…+(Ia8-Ib8)^2], so that 0So, for the 4 ideologies above, we would have the folowing (symmetric) matrix:

code:

---

IDEOLOGIES	Ideo1	Ideo2	PCIdeo	Current Values
Ideology 1	0	1,179	0,510	0,583
Ideology 2	1,179	0	0,922	0,742
PC Ideology	0,510	0,922	0	0,529
Current Values	0,583	0,742	0,529	0

---

the 51% method doesn't need anything to tell you when it's safe. Whenever you pass a law that people don't like, it'll make them somewhat unhappy.

Too much unhappiness, and *then* you have a problem.

Axi was worried about two model's features:
1) As a result of the process of negotiations between classes, some of the govt policies end up with values that are a mix between the values of different ideologies. Axi was worried about this mixing of ideologies in the govt, specially in cases when ideologies are too different. He thought (thinks still?) it was unrealistic to mix ideologies to find a final govt policy. As an example, you can think people supporting a capitalist system and other people supporting a communist system. All this people now "negotiate" to define what system will be implemented. The negotiation procedure in the model would give us as a result something in between these systems, where each prevails depending on the relative pol.powers of each "party". If they're both equally powerful, FE, the final system is "in the middle" of the two "pure" systems. I think Axi didn't like this because he saw it as a "friendly" conclusion of a negotiation between antagonic forces that would unlikely work together. My argument was that this analysis was running short. The negotiation procedure only reflects what happens within the govt and only what has to do with the "legal" part of politics. Since no party has total power, then it's fair to say the opponents have to give away things to obtain others, so the final setting is in fact not one of the original positions, but something in between. This, however, doesn't mean people will live happily ever after. Since the final setting is different to both initial positions, people of the two parties will be upset when comparing what they wanted to what was achieved. This is made in the riots model and chances of revolts, revolutions and such would be higher. So this antagonic forces are indeed antagonic. While politicians in the govt are bounded by legal activities and negotiations, what happens in the street can be totally different. In no way mixing ideologies implies a friendly cooperation. If antagonic ideologies exist at the same time, the most probable is that wouldn't last.

2)The ruler can choose other values than those he really wants to achieve what he wants. Usually this implies a more "extreme" position than the one the ruler'd normally take. This is what Axi calls "the machievelean ruler". This is in fact what I consider the worst part of the model. To solve it, I proposed this: People will look constantly at what the ruler offers (ruler's preferences), so if he takes really extremes positions, he'd face protests or things like that. I'm not entirely satisfied with this, but I don't see another good way to proceed.

Before I gave my arguments to Axi (the ones exposed here), he developed an alternative system based on the majority rule, which is, as far as I understand, pretty much the same thing F_Smith was proposing to do. With that model he doesn't allow for ideology mixing (his preocupation) and in part eliminates the machiavelean ruler effect. I didn't like the idea because I felt it was useful only for democracies (it was an election-type method) and even in those cases it had outcomes I considered unrealistic. Even more, election-type systems take away a lot of flexibility. In an election system, as its name says it, one of the competing choices wins (is elected). Therefore, the govt takes exactly the form of one of the ideologies in competition, which in time means it's impossible to model smooth changes in the political system because changes occur in a sudden way from one ideology to another. We couldn't get things like "an ancient republic, yet a little bit monarchic" regime or that type of things that the current model does allow. As for realism, the majority rule makes, FE, a scenario where the ruler has 51% equal to one where he has 100%. He's able to impose his will exactly in the same way in both scenarios because he has the majority pol.power. I find this absurd. Majority systems also leave aside any other political agents. It'd be like saying the elected republican administration doesn't have to care about the democrats and can impose its will completely, while we know this isn't real. The majority rule systems and election systems tend to be IMO all-or-nothing systems with lack of flexibility.

Part of the discussion quoted by Axi referred to what is exactly meant with the word "government". This is important because it defines what is possible and realistic in the model and what not. In order to model all the political games in any type of regime, the definition I used includes in govt all persons and institutions who have the capability to participate in decision making regarding govt laws (govt policies). So it can be just a king, a king and religious leaders or a whole senate plus the elected administration in a democracy, FE.

I am ready to move on with coding more of your model. I want to add the 'Ideologically Negotiated Policies'. First I'm going to have to define 'Ideology' objects.

So please help me make sure I understand --

• An 'ideology' holds the people's prefs on 'politcial structure', Private Property, Social Policy and economic planning?

• There will be a 'game level' list of all ideologies?

• Each Civ will have a list of 'known' ideologies?

• Each EG will have a list of 'chosen' ideologies?

• 'Ideologies' are going to be selected by the 'Ethnic Groups', and it is necessary to know what % of the EG supports each 'Ideology'?

• The 'Religious Leaders' will *not* have an ideology?

• Neither will the 'Generals'?
  

Is this all correct?

P.S. -- if you ever get the inclination to try out the beast again, your 'directly negotiated policies' are all in there. As before, I used the equations from the Govt Model v.2 thread.

Q:An 'ideology' holds the people's prefs on 'politcial structure', Private Property, Social Policy and economic planning?
A:Almost correct. An ideology holds that set of variables, but it's incorrect to say they're the people's preferences. An ideology is independent from who support it, so an ideology cannot be associated with a particular class per se. Ideologies are not class-specific or ethnic-specific or whatever. They are stand alone ideas anyone can choose or let go.

Q:There will be a 'game level' list of all ideologies?
A: yes

Q:Each Civ will have a list of 'known' ideologies?
A:yes

Q:Each EG will have a list of 'chosen' ideologies?
A:Social classes have a list of chosen ideologies. We could do it on a EG base, but that'd be ambiguous. Since within an ethnicity you have different social classes with different behaviors and needs, what's important is to know what each class prefers.

Q:'Ideologies' are going to be selected by the 'Ethnic Groups', and it is necessary to know what % of the EG supports each 'Ideology'?
A: yes... but for social classes. In my mind I imagined a "support matrix" like this:

Class_______Ideo1______Ideo2______IdeoN
RC___________10%________40%________50%
UC___________25%________5%_________70%
LC___________60%________20%________20%
WC___________75%________15%________10%

where each row must sum 100%. This matrix says, FE, that the 40% of the RC supports ideology2. Of course, you don't have to code it as matrix but in the way you think is best. But that's the idea.

Q:The 'Religious Leaders' will *not* have an ideology?
A:Wrong. RC will have a list of supported ideologies like any other class.

Q:Neither will the 'Generals'?
A:Wrong. WC will have a list of supported ideologies like any other class.

A couple of comments:

1)The list of supported ideologies should include by default the "govt profile" and the "ruler's govt profile". So, if the list of available ideologies has 10, then each class can support 10+2=12 different ideologies with different support shares. This is because the govt profile (that is, the current values the govt has for the political structure and PP, SP and EP) can be seen as an ideology and therefore people is allowed to support the current govt setting. The same happens with the "ruler's govt profile", but in this case with the ruler's prefs for the same values, so the people is allowed to support ruler's "vision" like they do with normal ideologies.

2) The one class that doesn't have an ideology is the Bureaucratic Elite. To be more precise, it supports ALWAYS the "govt profile" ideology with 100%.

3) What political structure are you using? The one from the pre-updated model or the one from post-update?
The pre-update:
ruler pol.power
RC pp
LC pp
UC pp
WC pp

The post-update:
ruler pp
capitalists pp
people's pp
WC pp
RC pp

1) Will each social class select it's Ideology choice at the Civ level?

I have put forward an idea for the govt player interface that I think is important and would like to hear your opinion on it. I put it in my post "posted August 21, 2000 21:00" in the Beast Prototype thread. If you could please read it and the subsequent discussion I'd appreciate it.

If it gains enough support I will try to transfer it over here.

Thanks,

In order to let you all (even those who don't know the math in the model) comment this proposal and analyze if it's better or worse than the current system, I'll explain briefly what's good and bad with the current system:

The negotiation procedure (current system) takes all the desires for a given govt policy from the different social classes and mix them to produce a single output (the negotiated final value). In this mix, the greater political power a class has, the closer the final value is to what the class desired and vise versa. That's how all govt policies are managed.

What's good:
1)It's very simple in the mathematical side.

2)It isn't biased to any particular type of govt. The "negotiation" can be interpreted according to the regime because it's simply an interaction of units with power and desires.

3)It provides a lot of flexibility. Since the final negotiated value is a mix of others, the govt profile can take a lot of different forms, so we can have, FE, several democratic regimes with slight differences between them.

4)Most important, it makes social classes "active" (as well as reactive). In reactive-type models, computer-handled political entities can only react to what the player does, like producing riots if they don't like what the player is doing. With an active political entity, it doesn't have to wait for ruler intervention to make a change in the govt. It doesn't need the ruler to try to pass laws or any such thing to express itself. The class uses the power it posseses to make a change in the govt. If a class has power, this is what you'd expect.

What's bad:
1)The player is encouraged to put as his preferences other values than the ones he/she really wants. He/she "lies" to the game. If FE he/she wants a CivilRights policy of 45%, then probably would have to tell the interface he/she wants 15% in order to achieve it. This is the "machieavelean ruler effect" Axi has pointed out, and means the player is forced to take more extreme positions than he/she would. The player is like cheating here because he/she understands how the procedure works, but social classes cannot counterbalance this action because they're not as smart as the human is.

2)You don't know as a player what the final outcome is going to be. You put in values, but they're gonna be mixed with others (what the classes want) so you don't know what's gonna be at the end. Maybe this can be fun for some players, but the trial&error this system encourages may frustrate others. It's also important to note here that although in F_Smith's beast changes ocurr immediatedly (one turn), it's mandatory for realism that changes ocurr slowly in several turns, so players will see the effect of their decisions only after a while, making the trial&error strategy less effective. However, this problem can be solved if we add something like a "Political Advisor" who could tell the ruler what's going to happen. This is easily implementable.

It's hard to classify the following in good or bad. I consider it good, but...:
The negotiation procedure actually makes the ruler less powerful in more representative regimes. The model was meant to avoid the player having total control of govt policies in this situations for more realism. Therefore, in a regime with low ruler's pol.power, saying "I wanted to have a CivilRights level of 34% and the procedure gave me 75%" and then complaining for that is against the overall idea of the model. It has been a goal all along to throw away the ever-all-mighty-player approach one is used to in other games. Realism in this matter has been a must. I know Mark and Axi knew (and agreed) about this. I wonder about the rest...

So what's new? The "machiavelean ruler effect" has always bothered me, but I wasn't able to create another system capable of erradicating this without losing the good stuff of the system. On the other side, even though I don't like F_Smith's system based on votes (for the tax rate in the beast), I must admit it has one very nice characteristic: you know exactly what value the policy will take (given the change is approved), which makes you feel more in control of what's going on.

So what I came up with yesterday is a system where the mach. ruler effect doesn't exist anymore and where the player can manage his decisions more directly. All without losing the good things the current system has.

The proposal is this: For a given policy, the negotiation procedure is used normally, but the ruler is NOT counted. This means the negotiation takes place only between social classes. A negotiated value comes as the outcome. Instead of saying that the final negotiated value is the one to be implemented in the civ, we transform it into a range where the ruler has to choose a particular value. For example, if the negotiated value between classes is X, then we say classes have decided the policy value must lie in the range [X-d,X+d]. "d" is a value that depends on ruler's pol.power. The more despotic, the higher "d" is. The ruler now has to choose the exact value within the range. The value the player picks becomes the official govt policy.

What does this imply?
The interpretation is classes define like a global "framework" for govt policies. Within the framework the ruler can define policies precisely.

If the ruler has little pol.power, then he has few options (small range). A policy like CivilRight would have FE a range like [66%,71%], so mainly it's the rest classes that are deciding its value, which is correct and realistic. But if the ruler has a large pol.power, the negotiated value just becomes a lousy reference. CivilRights would have now FE a range like [23%,78%], so it's mainly a ruler's decision completely.

In this situation the player doesn't have to "lie" to the interface. The values he chooses are those he wants, at least those the regime allows him. If he's still not satisfied (his really desired value is outside the legal range) then he has two options: a)simply accept he's not all mighty and live with that. b)try to get more power becoming more despotic. Of course here we're dealing with legal actions. He in fact has more alternatives to get what he wants with some questionable actions.

Once the ruler doesn't have to lie, the mach.ruler effect disappears. And on the other side the player can see directly how his decisions take place instead of going through the negotiation process. (the two goals I was seeking with this proposal)

Does this idea sound better to using simply the negotiation procedure (with the ruler counted in)? I don't have personally an inclination to what's better, so I want to know if you guys consider this proposal interesting.

It's important to say, however, that the proposal only works for policies, but not for the political structure. Normally the negotiation procedure is used too to alter this structure of power, representing negotiation between actors to define how much power each should have. For the time being I can't provide with a solution for that part. The proposal here doesn't work for pol.power variables because if the ruler cannot participate in the negotiation, it will be way too easy to steal him his power (assuming classes want that, which is not necessarily true always, but it'll be of course common many cases). And this, of course, is not admisible!

As you will soon figure out, your idea two posts above is very similar to the proposal I was trying to get you to look at . I think it is much better for the player than the way we were handling it.

The range listed is the "safe" range, right? The ruler can choose any number in there without causing problems, and it would be represented on the slider as the green area.

But the ruler should be able to override the desires of the other classes using "foul play", martial law, or whatever. Choosing a policy outside the range would require such actions and would not be popular. These areas would be represented on the slider as yellow and red, depending on the negative consequences that would result from the action.

2> The variance seems not quite right to me. It seems the variance should also be not ness X-d thru X+d, but more based on the social classes deciding what it is. FE if a majority wants very low taxes then it might be more like X-d-p thru X+d-p where p is the political pressure put on by that class to keep the rates lower.

Now about why negotiations shouldn't always work: Generally i say they should, about 90-95% of the time, but nothing is guaranteed and the ruler shouldn't always have the luxury of knowing the policies will be settled for a medium.
FE its quite possible that 2 groups that are opposite extremes (ie 0 vs 100 on everything for this example) will not tolerate each others proposal. Take also the fact that say they are very fundimentalistic and won't budge to even close to a moderate solution (quite possible) and the other groups make up so small a percentage of political power that their voices aren't heard amist the other 2 groups (also quite possible) so no matter what there would be no compromise because neither group will agree perhaps to even meet with the other group. Such situations will if not already eventually lead into internal warfare, but if it isn't already no policy can be negotiated because neither side can stand the other and the other small groups are so small that anything they propose will be drowned out by the other two (which they might likely meet temporarily to make sure that such a proposal doesn't pass).

LGJ: I'm not gonna comment your points 1 and 2 because I'd prefer you all to look at my new idea (below). As for your thoughts about antagonistic parties negotiating policies, I think you have a good point there. I believe you're right when you say a very heterogeneous political arena can lead to a less effective govt when it comes to negotiate policies. To generalize your idea, I think what we can do is to measure how much diversity a govt has. This is easy to do. Then we use that measure and modify the time for achieving negotiated policies. The greater the diversity, the more game turns it takes to achieve final values. I think this would be a nice feature. Opposed visions would tend to "paralyze" the govt, but not completely, of course. How does it sound? What do the rest think?

Richard: I think we see things pretty much the same way regarding what the player can and cannot do, but there may be some slight differences of importance, so I'll be more precise in what I think: I consider very important to limit player's control over the govt according to his pol.power. I don't like the idea of "safe" and "unsafe" ranges where the ruler can impose govt values outside the legal framework facing risks. This is like saying the ruler is really all powerful all the time. This converts a govt just in a matter of how much risk you want to take, but not a matter of having or not the power to do things. That's not realistic IMO. The US president can't do a lot of things, FE. Many things are simply outside the reach of his power and they've nothing to do with how popular a decision would be or if the guy in charge has or not the nerves to take a risk.

A regime for me defines the limits of ruler's intervention. If the player wants to do things he's not entitled to, then he needs another regime with more power for himself. He may use Special Actions to get more power and THEN, with a new framework, take the decisions he couldn't in the past. So I'm seeing it as a two steps move. We're not saying to players "you can't have whatever you want for govt policies", but we're saying "if you want to impose your view, then FIRST you have to get yourself a dictatorship (of some sort)". Does this sound reasonable to you all?

Mark: I don't understand well your idea with colored slides. If you say it was similar to my proposal above, then... there isn't much to say!
Your comment about the need to know ruler's preferences to see if he's reactionary or not and, in general, knowing what he stands for, is very good. You're right. We really need that. Otherwise the concept of govt and ruler tend to become one and this shouldn't happen. The truth is my proposal in my previous post isn't compatible with keeping ruler's preferences in the system. It wouldn't be possible to eliminate the machiavellian ruler effect and then the proposal has very little value. Thinking about this triggered this new idea I have.

Here's my new proposal:

As before, the negotiation procedure is used only for social classes not counting ruler's preferences. Social classes end up with a negotiated value X for a given policy. Then the game takes X and modifies it according to what the ruler wants. If the ruler was pushing for a higher value, then X is increased. And vise versa. The magnitude of the increase/decrease is given by ruler's pol.power. Example: Classes alone negotiate a Civil Rights value equal to 45%. The ruler wants 66% and he has 30% of total power. Since 66%>45%, the ruler pushes for a greater value. Let's assume his 30% pol.power allows him to alter any negotiated value by 15% at most. Then, CivilRights ends up being 45%+15%=60%. In general the system is

Govt_Policy=X+sign(ruler_pref-X)*min(abs(ruler_pref-X),F(ruler_polpower))

where F is a function TBD transforming ruler's pol.power into the maximum modifier the player could apply to the classes' negotiated value.

It's similar to the "range" approach I proposed in the other post in the sense that classes take a "preliminary" decision and the ruler then affects it, where ruler's intervention level is given by his pol.power. But in this case the ruler has to express what he wants prior to the decision-making process (as it was planned in the original system) and it's possible to differentiate ruler's intention from actual govt. Classes will now see what his intentions are, which is very important as Mark pointed out.

There's no machiavellian ruler effect. The player gains nothing if he lies to the game. Using the equation above, the best the player can do in order to achieve his goals is put exactly what he wants in the interface. Doing otherwise would only harm him. Therefore, there's no trial and error and players won't have to waste time finding how to fool the system. And this new system still has the good things: it's simple, classes can use their pol.power to influence govt values and it's sensitive to how much power each actor has.

What about pol.powers? In my proposal with ranges in the other post this part was still unsolved. The best thing about this new system is that it is also compatible with pol.power shares. We can use the same equation (slightly altered to ensure pol.power shares will sum 100%) to simulate how political entities define the political structure with good results.

Finally, for policies (that is, all variables except pol.power shares), we can still get that feeling of "more control" each time the player changes his prefs. In the example above for CivilRights, if now suddenly the ruler changes his mind and prefers a value of, say, 50%, then the govt can make the change immediate. The old govt value, 60% would be changed to 50% at once. On the other side, if the ruler changes his preference to a value of 75%, the govt value would remain at the old level:66%. With this, the type of messages F_Smith has in the beast would fit nicely here. In the first case the message would be "Our leader has decided to lower the Civil Rights level from 60% to 50%" while in the second case it'd be something like "Our leader wanted to rise Civil Rights to 75%, but was rejected by the govt's high council. The law remains at 60%!".

While I don't like the interface proposed in the new system, it seems that the math is better. The two proposals are very similar, so would it be possible to put the new math behind the range interface?

By the way, I have one word to respond to Rodrigo's comments about the political power of US presidents and their inability to do certain things:

Happiness must be computed comparing what they want to what the govt ends up being. That's all other topic. The thing to solve here is how classes and the ruler interact to set policies (a final value for each) and not what makes people happier.

The range represents the legal limits and not what's popular and what not. Of course topics are related, but don't confuse!

I haven't read all this carefully, and need to reflect on your new idea. But I have just a quick note to clear up a misconception. On the green - red ranges idea -- for a democratic government even the red (extreme actions) range does not necc go from 0-100% of each parameter. Suppose the people in general want 20% for some govt. parameter.
For a very democratic state the ranges might be: Green 18%-22%; Yellow 16%-25%; Red 10%-32%. And that's it! Even Extreme actions in a very democratic govt only allow so much flexibility because the rest of society is pulling against the ruler. If the player wants more power then they have to go your route of changing the govt. Or if they get caught doing something in the red range the govt might collapse anyway...

FE Nixon and many other US presidents have gotten away with a lot of things that push beyond their constitutional power. But even Nixon used extreme actions to move civil rights just a small amount (maybe a few percent on average). In no way could he have reinstituted slavery, even with the most extreme covert actions you can think of!

Below, I am using the term "power groups", which refers to classes for the DNPs and to ideologies for the INPs.

On the policy-setting procedure: Rodrigo's idea sounds rather good, since:

• It gets rid of the "Macchiavelian Ruler Effect".
  
• It is independent of the Current Govt Profile. This way, power groups that have the will and the power to do so, can be "active" instead of "reactive". As Rodrigo wrote:

  quote: Most important, it makes social classes "active" (as well as reactive). In reactive-type models, computer-handled political entities can only react to what the player does, like producing riots if they don't like what the player is doing. With an active political entity, it doesn't have to wait for ruler intervention to make a change in the govt. It doesn't need the ruler to try to pass laws or any such thing to express itself. The class uses the power it posseses to make a change in the govt. If a class has power, this is what you'd expect.

  
  
• It allows the ruler to be judged for his preferences, while the govt will be judged for the actual policies. Things like the RRF should be determined according to the ruler's profile rather than the govt profile.
  
• It shows to be compatible with the rest of the model.

I tried to find a proper expression for F. I am using the following abbreviations:
P=govt_policy (0-1)
R=ruler_pref (0-1)
r=ruler_polpower (0-1)

Negotiation procedure: P=X(1-r)+Rr

New procedure:
IF R > X
THEN P=X+min[R-X,F(r)]
ELSE P=X-min[X-R,F(r)]

To be equivalent, it should be:
If R > X and F(r) < R-X:
X(1-r)+Rr=X+F(r)
or F(r)=r(R-X)
which is always less than R-X because r < 1
For R < X, we have F(r)=-r(R-X)
Here we see what's wrong with the negotiation procedure: F depends from R, hence the "Macchiavelian Ruler Effect".

F(r) should be such that for r=0, F=0 and for r=1, F=max. But maxF=1-X if R > X and maxF=X if R < X. The simplest possible implementation is:
IF R > X
THEN P=X+min[R-X,r(1-X)]
ELSE P=X-min[X-R,rX]
This is the equivalent of using the negotiation procedure with the ruler always highballing to the maximum degree (always choosing 0 or 1, according to what suits him best). In other words, this is a way of "automation" of the MRE, so that the player won't get into the trouble of trying out manually to find what he should bid for. Obviously this is not the best solution, since the ruler has a clear advantage over the other power groups. Other implementations like F(r)=r, or F(r)=r² or F(r)=ar where 0 < a < 1 do not give what is expected for despotic regimes, while they are no more representative of the ruler's power.

We could of course accept that the ruler, as the initiator of the negotiation, has a de facto advantage, so his "automatic highballing" would be accepted as a feature of the game. But then the ability to initiate negotiations should be granted to other power groups too (since we agree that power groups in Clash should be active). This means that we would have to provide our system with criteria for when a negotiation should be pursued by the power groups and when such a claim should be accepted. We should also have to devise a way for the ruler to participate in such a negotiation, from a different, non-priviledged position.

These preoccupations, in accord with F_Smith's repeated statements about having a "game within a game" (which I gather is the desire of all of us; that's why we are trying to devise a political system), have led me to another idea: if we want the player of Clash to really "play politics", shouldn't we then use a "Political AI"? I feel that the power groups shouldn't be just "dumb" bodies, with certain power and certain preferences, that act or react always in the same way. I feel it would be nice for the player to know that he is facing an "internal enemy", and not just a "flock of sheep". Using the traits of their respective character-leaders (ambition, fundamentalism, ego, venture, etc ,whichever there are), each power group should have the liberty and the brains to differentiate it's actions, for the purpose of better achieving it's goals in the current circumstances; exactly like the player is expected to do. This sort of AI (internal enemy) should of course be much simpler than the military AI or the AI of a rival civ, but it should be able, within certain (severe) restrictions, to:
- Differentiate it's preferences, in order to affect the negotiations. (old negotiation procedure)
- Initiate negotiations, when it suits it. (new negotiation procedure)
- Form a coalition with another