AU mod: Balancing the Governments

[nbarclay]

Actually from what you say, my bet is that AIs would almost always choose Feudalism (with its minimal corruption) in peacetime. I've already seen AIs use Feudalism an awful lot in the AU Mod when I would have expected them to use Republic. (And that has continued to some extent even after the tweaks to make it choose Feudalism less often.)

[joncnunn]

This could be a side affect of changes made between 1.0 and 1.22.

In 1.0, the AI loved cash rushing so much that in Fall of Rome, the Sassidens overthrew their Tribil Council govt in favor of Imperalism upon aqurining that tech!

Perhaps now the AI always tries to pick the lowest corruption governement in peace time.

On the primise that peace time govts should always have higher corruption than war time govts, I disagree. Instead the earlier govts of all types should have higher corruption than the modern govts of all types.

Perhaps something like:
Anarchy : Total : no change
Despotism : Rampent
Republic : Problemic
Monarchy : Problemic
Fedualsism : Nusiance
Democracy : Minimal
Fasicsm : Minimal
Communism : Communial

The AI should be able to handle that and it does very marginally weaken Republic.

[nbarclay]

Originally posted by joncnunn

On the primise that peace time govts should always have higher corruption than war time govts, I disagree.

It's not a question of "should." It's a question of, "What options does the editor give us to improve the balance between governments?" Giving governments that have a commerce bonus worse corruption is one of very few tools we have available to make the commerce bonus less thoroughly overpowering from a research perspective. The real problem is, our options are hemmed in on one side by what the editor makes possible and on the other by what the AI can handle without doing stupid things as a result.

[Dominae]

Re: AU Governments

Originally posted by pvzh
At first, I think we need to isolate what exactly the problem is. This was done perfectly by Dominae: you can do anything with republic without much penalty associated and if there are any gain from any switch anarchy will eat any benefit. The reason that you can do anything in Republic is that you have trade bonus, second best corruption and low WW (government cannot collaps by itself). Trade bonus multiplied through markets, banks and stock exchanges can pay in the most cases for large enough millitary to do effective conquest, and second best corruption will help you with that too.
If we want to redisign government, we need to ask ourselves several questions:
1. What kind of government choices should be presented to player?
2. How each goverment is supposed to play, so what features should they have?
3. How many switches supposed to be?
Right now we have Despotism, Monarchy, Republic, Feudalism, Democracy, Fascism and Communism. I do not think we should create any more, we could consider ditching the Fascism.

Let us keep it simple and make any government either war or peace. War Government (no Trade Bonus, no WW, MP, free units) -- Peace Government (Trade Bonus, WW, no MP, no/little free units). Since, we cannot scale Trade Bonus and it is BIG (I mean really big increase in commerce we should work through corruption, and WW to offset its advantages).
War Government is world domination wanna-be, so it should have some kind low corruption to have lots of cities and have high shield production for units in them. Thus, all of them should have communal/minimal/nuisance corruption extra FP and so forth.
Peace government should be geared for relatively small empires whose objective to tech through by their high commerce and FEW highly developed cities. Thus, peace goverment should have HIGH corruption (rampant/problematic), to make conquest fairly unprofitable and developing your small core more important.

Since switching to advanced forms of government is extrimely painful if you are not religious we should consider another Anarchy.

Based on this I propose such changes to goverments

...

Redraw tech tree around Nationalism (since nationalism cost sky high due to attacker/defender assosiated with it and all other goodies) we should keep it optional, so it will not be even more expensive to trade. Government techs should have Espionage as pre-request, since Espionage have Industrialisation as pre-request. Move Police Stations to Espionage and increase its cost to 120 from 90, to lower chance that AI will choose it right after industrialization. Lower cost of both Communism and Fascism to 90 from 120, so AI will not spend too much on them time after Espionage or bulk them in one tech Totalitarism at cost of 120. Another option to get rid of Fascism and not to worry too much about it.

Good post.

[joncnunn]

There's a couple of ways I can think of to decrease the differerance between War Govts and Peace Govts without adjusting govt corruption:

1. Give roads a +2 bonus instead of +1 on Grassland & plains OR
2. Give grassland & plains 1 commerce.

I'd prefer the 1st, but the AI might be able to handle the 2nd case better.

Tech costs would probably have to be increased to compensate.

But this would decrease the commerce bonsus aspect of Republic/Democracy.

[Dominae]

Originally posted by nbarclay
Despotism - 790 - 211 - 26.7%
Monarchy - 857 - 187 - 21.8%
Republic - 1333 - 304 - 22.8%
Feudalism - 861 - 173 - 20%

You really think a 4% increase in Republic's Corruption will solve the problem we're discussing here? Republic would only be 5% more Corrupt than Monarchy; it's still producing ~55% more Commerce!

[nbarclay]

Dominae, the Republic figure is skewed by peculiarities of how the commerce bonus works and/or fits together with the particular scenario I was experimenting with. Republic's actual level of corruption is somewhere in between those of Monarchy and Feudalism. In the absence of quirks caused by the commerce bonus, it would be somewhere in the neighborhood of 21%, give or take a little depending on whether "Nuisance" is closer to "Minimal" or to "Problematic." The fact that corruption shows up as 22.8% instead reflects a phenomenon that results from the commerce bonus, not from the difference in corruption levels.

That same phenomenon resulting from the commerce bonus would almost certainly affect Republic about the same way with Rampant corruption that it does with Nuisance corruption. So corruption under Republic with Rampant corruption in this scenario would probably be somewhere in the neighborhood of 28.5%, versus probably somewhere in the neighborhood of 21% for Monarchy with Nuisance corruption. (Alexman, do you have anything to contribute regarding whether Nuisance would be closer to Minimal or to Problematic?) That puts the actual advantage for Monarchy at more like a 7.5%, give or take a bit depending on exactly where Nuisance corruption stands between Minimal and Problematic. And remember that that difference would affect production, not just commerce.

But that's still not the whole story. The difference in production would affect how quickly outlying cities can build courthouses, libraries, marketplaces, harbors, aqueducts, and various other improvements that have a direct or indirect impact on commerce. So the corruption change would close the commerce gap more than the difference in percentages alone would indicate.

Then factor in the difference in unit upkeep costs and the potential impact of military police on the luxury slider. Republic loses some of its gold paying for units. Monarchy does not or, in its worst situations, loses a whole lot less. And if military police allow Monarchy to run the luxury slider a notch lower, that's about a 10% advantage for Monarchy right there.

Republic would still have a commerce advantage over Monarchy, but it would be both significantly smaller and counterbalanced by a disadvantage in production. I think that could make for a much more interesting choice than we have now - if the AIs could cope with the change without making worse choices as a result. But if, as Alexman suspects, such a change would cause AIs to make worse choices of what government to use, this whole discussion is probably academic.

[nbarclay]

Originally posted by joncnunn
There's a couple of ways I can think of to decrease the differerance between War Govts and Peace Govts without adjusting govt corruption:

1. Give roads a +2 bonus instead of +1 on Grassland & plains OR
2. Give grassland & plains 1 commerce.

There is no way to do it without changing the feel of the game too much. Remember that the "standard tile penalty" affects any tile that produces more than two commerce. Under the stock rules, that leaves room for a +1 bonus to commerce (for example, from rivers) without having it eaten by corruption. Increase the amount of commerce generated and a lot of commerce bonuses would disappear under Despotism. Also, unless theere's a setting we could adjust that I don't know of, the luxury slider would give a bigger boost compared with the percent of their commerce players would have to give up to use it - something that would definitely work in humans' favor.

[joncnunn]

Looking at what the stats of the govts currently are in the AU mod, perhaps compared to the AU mod:

Add a Communist Version of the SPHQ. But instead of acting like a second FP, have it act like a Police Station in every city. Also move this Communist version of the SPHQ to the Communism tech.

Republic : Reduce unit support from 0/1/1 + X free units to 0/1/0 + X free units

Democracy : Reduce units support from 0/1/1 + X free to 0/0/1 + X free. Add Gunpowder as a prereq.

The idea would then be if you have more medium size towns than Metros, you get more free units in Republic while if you have more Metros than medium size towns, you get more free units as a Democracy.

[joncnunn]

On AI wasting time picking any tech allowing a govt, I think rather than rewriting the Industrial Tech tree to strongly encourage the humans to go thru Nationalism and/or Espioage that it would be simpler to trim the tree down:

1. Move the Commie & Fascist govts to Nationalism.
2. Move Police Station to Nationlism.
3. Remove the Commie & Fascist techs.

[Modo44]

Originally posted by joncnunn
Add a Communist Version of the SPHQ. But instead of acting like a second FP, have it act like a Police Station in every city. Also move this Communist version of the SPHQ to the Communism tech.

This would make Communism a no-brainer, even for peaceful games. It already dramatically reduces corruption, even without the SPHQ.

Originally posted by aksully
Republic : Reduce unit support from 0/1/1 + X free units to 0/1/0 + X free units

Either that, or improve Monarchy, Democracy, and Fascism...

Originally posted by aksully
Democracy : Reduce units support from 0/1/1 + X free to 0/0/1 + X free. Add Gunpowder as a prereq.

Why? Weakening Democracy has little value, as it is already a neglected government form.

[joncnunn]

Communism does not give extra trade arrows.

Democracy's extra trade arrows multipled from the core areas with cities would outnumber the number of trade arrows a communist civ that had been peaceful all game long would have, even if communism had corruption of none.

The counter move on Democracy to 0/0/1 +X is because I think it would be too much of a no brainer to switch from Republic to Democracy if Republic was reduced to 0/1/0 +X free while Democracy remained 0/1/1 + X.

Originally posted by Modo44

This would make Communism a no-brainer, even for peaceful games. It already dramatically reduces corruption, even without the SPHQ.

Either that, or improve Monarchy, Democracy, and Fascism...

Why? Weakening Democracy has little value, as it is already a neglected government form.

[nbarclay]

The idea of punishing players for growing cities into metros seems a bit perverse to me. A number of highly experiecned players, myself included, already use strategies that build our cores in ways that leave little room for cities to grow past size 12. (The goal of that approach is that we don't want to have to wait for hospitals before we can use virtually all the available tiles). Hospitals already have a relatively low value for players who pursue such a strategy, and having them cost free unit support would undercut their value even more.

I don't agree that giving Communism a small wonder that provides free police stations would make it a no-brainer even for peaceful games. Even with free police stations, it would take a bigger civ than players can get with REXing alone for Communism's lower corruption to offset Democracy's trade bonus.

The real danger to making Communism more powerful is that if we make it viable for small civs, we make it ridiculously overpowered for big ones. Under the stock rules, a communist warmonger can have it all: massive production, massive free unit support, no war weariness, and about as much commerce as Democracy could provide - probably even more in some situations. That's why we took the SPHQ away from Communism.

[alexman]

One last attempt to balance the Governments

After some thought, here is a new attempt to balance the governments, inspired by some of the posts above.

In order to encourage more frequent government choices without seriously weakening the Religious trait, the corruption of Anarchy is changed to be Rampant, like Despotism. Since it now gets an income, Anarchy’s free unit and building support is also removed. This is a major change, but very much worth it, if we are going to see a variety of governments in the game.

Despotism is unchanged.

Monarchy is unchanged. It is the wartime government of the ancient age.

The Republic needs to be seriously changed, as it is the major reason for the imbalance of governments in C3C. This new Republic is intended for specific situations: high-commerce, peaceful empires with very light military. No more general purpose good-for-everything governments! Therefore, all free unit support is removed. This change by itself would overly penalize empires with too many Workers. Therefore, the Worker rate for Republic is increased to the level of Democracy. That way Republics can cut down on Worker production to deal with the high support costs, without falling behind in tile improvement. Note that the AI will be able to deal quite nicely with the reduction in free unit support, as this is a factor (unlike corruption) that gets evaluated correctly by the AI when selecting a government. (On a side-note, this change will probably slow down the tech rate in the late ancient and early medieval periods, which are exactly the times of most interesting warfare, and most interesting Wonder choices, in my opinion).

Feudalism is unchanged from stock, and remains an alternative government to Monarchy. UnOrthOdOx’s campaign for Feudalism awareness has convinced me that the reduced corruption of the current version of the AU mod is not necessary, especially in this current proposal, where the Republic has been weakened.

Democracy is also unchanged from stock, but its strength relative to the Republic is greatly improved compared to stock. Peaceful players that still need an army to defend themselves now have a great incentive to research Democracy as soon as possible. That in turn makes the unpopular Printing Press and Free Artistry technologies more likely to get researched. I do not believe that the changes to Anarchy will motivate players to warmonger in Democracy. The penalties for a government getting overthrown may be reduced, but the war weariness penalties in happiness are still significant.

Communism loses the SPHQ just as in the current version of the mod.

Fascism keeps the SPHQ and the increased corruption, as in the current version of the mod. We also keep the adjustment so that the AI rarely chooses this government. Since the AI does not know how to choose the better wartime government based on empire size, it’s better to just choose the wartime government that’s better most of the time, which is Communism. Fascism remains a human-only toy, good only for very specific and rare situations.

So in summary, compared to the current version of the mod, I propose:

• Anarchy: Reduce corruption to Rampant. Remove free unit and building support.
• Republic: Remove all free unit support. Increase worker rate to 3. Restore AI favored government choices as in stock (they no longer all favor the Republic).
• Feudalism: Revert back to stock (problematic corruption).
• Democracy: Revert back to stock (no free unit support).

[nbarclay]

Am I the only one who would seriously consider no longer using the AU Mod if Republic were nerfed that horribly? I wish Republic weren't as unambiguously superior under as wide a range of conditions as it is, but creating a situation in which a Republic has to give up its economic advantage to maintain even a halfway decent defensive military - especially in the early stages, before city improvements start multiplying commerce - is not my idea of fun.

Even though Republic as it now exists in the AU Mod is pretty much a no-brainer under many conditions, the AU Mod makes choices of how to use Republic interesting to me. I have to seriously consider whether it's worth keeping outdated military units around and paying upkeep on them, and I have to weigh the advantages of conquest against the upkeep cost of a military capable of conquest. I don't view the advantages of making the choice of governments more interesting as worth making Republic no longer fun.

The question is, am I alone (or at least nearly alone) in feeling that way, or are there others who would also view such a change to Republic as seriously damaging their fun? If there are others, the harm done to those of us who like Republic's versitility would vastly outweigh whatever minor advantages making the choice of governments more challenging might offer.

I'm also concerned about what eliminating all free unit support for Republic would do to the AIs, especially on difficulty levels where AIs don't get massive bonus free unit support. Would AIs still use Republic, and would it be a good idea if they do? What happens if an AI switches to Republic at a time when it's relatively low on units (for example, right after fighting a major war) and then builds up a significantly higher number of units? I'm not convinced that the proposed change to Republic is something we could count on having work properly without some significant play-testing on a variety of different difficulty levels.