[C3C] Rebalance Mod - Governments

[md4]

As the title suggests, I have been working on a government rebalance mod for C3C, with a couple of broad aims in mind:

• Nerf Republic slightly, so that, as an ancient-era government, there is at least some incentive to brave some more anarchy to go for Democracy when it becomes available.
• Redesign Feudalism so that it isn't completely useless outside of a single niche scenario.
• Early governments are tuned for less developed empires: fewer cities, smaller populations (higher % of 1–6 pop towns), so on, while late-game governments scale better.

I haven't had the chance to put all that many games through this mod to test balance, as my Civ-time is a bit limited at the moment, so I'd be keen to poll some feedback and ideas from peoples' own experiences (to clarify, definitely not expecting anybody to playtest this for me, just maybe pick up on obvious mistakes or bad ideas). I'm open to ideas from any perspective: historical accuracy, gameplay/balance, fun, anything at all.

One specific question I do have, based on my changes, is the AI is able to account for increased cost/unit or decreased free units and build fewer units accordingly, or will it go broke trying to build the same number as it normally would?

With that out of the way, here they go:

Despotism
Still the worst. No changes required.

Monarchy
War weariness is increased to Low, and free units per city is now 3/4/5. It's still (intended to be) a viable 1st choice if you intend to go down a more militaristic path (see Republic).

Republic
Corruption increased to Problematic (same as Monarchy) but the Trade Bonus is retained. War weariness is High. I have confirmed that this really hurts in the early era. Later on you have mitigation options (surplus population as entertainers, luxury slider, luxury trades, police stations, etc.).

Feudalism
Redesigned as a "better" Monarchy, forcing the player to choose between an early Monarchy switch or wait around a little longer for something with more future potential. Corruption is Nuisance, cost/unit is 1, and free units per city is 1/6/6, giving Monarchy a slight edge in small undeveloped empires, where Feudalism's decentralisation via vassals allows larger empires. I accentuated support costs to achieve this. Hurry method is also Paid, as with Monarchy.

Democracy
Democracy is unchanged, except it is no longer immune to propaganda.

Communism
Worker rate increased to 3, War Wariness is now Low (making Despotism & Fascism the only weariness-free governments), and the free units per city slightly modified to 5/6/7 (as much as I enjoy the thematic uniformity of 6/6/6).

This is positioned as slightly more of an all-rounder between peacetime and war, so I am still on the fence about the Hurry Method. The question is whether this Communism is more of the Soviet flavor of forced labor camps or a more modern hybridised Chinese version. (My limited understanding on this subject is that China is more a combination of free market and planned economics, and might in a very abstract sense be seen as "softer" than the post-WW2 Soviet model.)

Fascism
Draft limit increased to 3, and free units per city changed slightly to 5/7/9. A minor tweak, but changed to gear it towards getting more out of conquered cities (likely to have been battered down to size <7).

 

[Ozymandias]

First off, it looks as though you've begun some excellent work here

The shortest answer to your question is, YES. This tidbit is from the "FUBAR Editor Documentation" thread:

"These settings have effect on the AI government choice:

• [...]
  
• Unit Support (cost per unit, free units and support per city *combined* - the AI makes a calculation based on the number of units and settlements owned to determine the best option)"

Also, if it's any help, I just recently reposted the formula which the AI uses to select Governments.

- And, WELCOME ABOARD!

 

[md4]

I'm still a bit unsure about the unit upkeep question. I did know that it factored into their government choice, but once they've made that choice, what then? As phrased, it seems like they first decide to build x number of units, then find the government to best facilitate that. Is anything known about what guides that initial decision?

Mainly because in the last couple of games I've started (small world size, default 6 players, random everything else), the AI players have been almost consistently broke for most of the game. And I mean broke: 0 gold, and never willing to pay even 1gpt even for up to 5–6 of my techs.

I haven't made any other changes to AI behaviour (swapped some traits around, but maintained at least 1 of every 2-trait combo). I've added corruption reducing effect to Temples, cheapened Courthouses slightly, and that's about it.

Since my changes have resulted in every government (except Feudalism) having less military support, I wonder if I am breaking the AI, who has some fixed behaviour in their code that dictates how many units they maintain based on their size (or possibly every nth item in their queue will be a unit). I remember that the AI seems to struggle with research around the Industrial era, but now it's like they've just switched off completely. City investigation shows they're mostly setting 70/30 tax/science rates, and producing pitifil science levels (Greece (Feudalism) has a 12-sized city with Library + University producing just 6 beakers).

Although it's a slightly different topic, I'll ask here since I haven't seen it asked anywhere else, but has anyone tested the effect of removing all the "Build Often" checkboxes from the civilisations? Does it make AI play better? More balanced?

 

[Ozymandias]

I'm still a bit unsure about the unit upkeep question. I did know that it factored into their government choice, but once they've made that choice, what then? As phrased, it seems like they first decide to build x number of units, then find the government to best facilitate that. Is anything known about what guides that initial decision?

As a rule of thumb, the AI will, with standard rules, inevitably select the Gov with the lowest level of Corruption (frankly, how & when Communism is chosen - beyond it generally being a preferred choice for wartime - remains a bit of a mystery to me.)

Mainly because in the last couple of games I've started (small world size, default 6 players, random everything else), the AI players have been almost consistently broke for most of the game. And I mean broke: 0 gold, and never willing to pay even 1gpt even for up to 5–6 of my techs.

I've never run across the "0 Gold" AI issue (if, for perhaps no other reason, that the AI does "cheat.") I'm also confused by part of your your initial post, viz. the AI, "Go broke trying to build the same number as it normally would."

I think these are interesting enough, in and of themselves, that I'll give it a "shout-out" here to whom I believe to beone of our most experienced modders, regarding the deepest, darkest depths of the Beast,
@Civinator .

One question however - Have the AI Civs been at war?

I haven't made any other changes to AI behaviour (swapped some traits around, but maintained at least 1 of every 2-trait combo). I've added corruption reducing effect to Temples, cheapened Courthouses slightly, and that's about it.

As you are engaged in admirably methodical testing, I would recommend limiting the number of variables for each iteration, especially as you're focused on the one question - "Is the AI is able to account for increased cost/unit or decreased free units and build fewer units accordingly?" Personally, I would start each test with each AI Civ using the same Gov type, and changing only one, single variable for each Civ/Gov-modification. Also, use precisely as many different Civs as required for each test case, making sure that all Civs belong to the same Culture Group. As each Civ will have need a different Gov Type name, make sure that each Civ has their own Gov Type set for both its "Shunned/Favorite" Gov Types.

Similarly, for test cases, I would begin by giving the AI only one unit type to build: land, flagged both "Offensive" and "Defensive." The AI will automatically assign one - and only one - of those flags for each unit built.

[...] has anyone tested the effect of removing all the "Build Often" check-boxes from the civilizations? Does it make AI play better? More balanced?

All that is really known about the "Build Often" check-boxes is that (1) they only seem to modify the AI's behavior ~5% and (2) selecting more than 3 will "confuse" the AI, making the choices irrelevant. Once again, I highly recommend limiting the number of variables for each test type/iteration - unselect all "Build Often/Never" flags for all Civs.

One cautionary note: Do not choose "Rome" as one of your Civs. I'm uncertain how widely known this is, but Rome is hard-coded to build proportionally more Offensive than Defensive units (the usual O/D ratio is ~40%:60%.) and to behave accordingly.

 

[md4]

I think you are probably right about there being too many variables. I probably need to play a few games under the original conquests.biq for comparison, and start re-introducing the changes gradually.

 

[Ozymandias]

 

[Predator145]

From what I've seen the war wearing AI will switch to Communism when it has researched the tech but not Fascism yet. If it benefit from switching to Fascism it will do so once that tech is in. If not it will stay in Communism. The deity games I've lost almost always features a run-away Communist AI. That's why the moment I see an AI go for that gov I make sure to gang up on it.

But if Fascism is already in the AI will never go for Communism even though it's big enough. Fascism seems to always be the go to choice for the war-weary AI if it collapses from Dem/Rep.

But I've seen a long time Monarchy warring AI stay in that GOV instead of switching to Fascism. The AI seems to deem a Mon-Fa switch not worth the penalty. These long time warring AIs often have large sprawling empires that could benefit from Communism. But I've never seen them switch into that.

Communism is overpowered. Yet the huge AI seems to have an aversion towards it. I'm afraid that even after modding Fascism out it would still choose Monarchy instead. So right now I lean towards improving Fascism with even more unit support, no pop loss, no xenophobia and enable cash rushing plus getting the Police Headquarters as a single city glorified police station.

 

[md4]

I've gone back to the drawing board. I still need to get in some games with purely vanilla settings to observe whether the midgame tech stagnation and bankruptcy was something I introduced or if it was just a fluke of circumstances. (I'm not really sure how to do any more focused testing other than to just play full games because the problem doesn't seem to manifest until well after all the land-grabbing is done and the AI should switch focus to development.)

I have created another .biq in preparation that is now just a Monarchy/Feudalism rebalance with the same outcome as before: that Feudalism is basically inverted so that it's better than Monarchy when you get more cities to size 7–12, while Monarchy still edges it out when you mostly have low population cities, so there's still (hopefully) a good chance Feudalism will pay for itself versus those extra turns you spend in Despotism.

 

[md4]

Updated OP: my ambitions have been suitably reigned in from a sweeping overhaul to more a series of minor tweaks.

 

[Fergei]

The Civs get massive subsidies for unit maintenance on higher difficulties. Civs that might get bankrupt with your modified Republic in Regent or Monarchy, might be comfortably well off in Emperor+. Just a variable to throw in to the mix - apologies if you've allowed for that.

I am very interested in any work you do that results in a greater uptake of Fascism and Communism - e.g. if you do anything that encourages an AI to shift from Republic or Democracy with a little more regularity.

As an aside, your use of high war weariness for Republic makes me think you would create a niche for Imperialism as a new goverment. A sort of Monarchy/Republic hybrid around the time of Astronomy/Navigation (although that isn't that much before Communism in the science tree strangely). Imperialism is in the Rise & Fall of Rome Conquests that come packed with the game, but is very similar to democracy, so there isn't much to be learned there. I was thinking more of the combination of:

- commerce bonus
- low war weariness
- problematic corruption
- Monarchy type unit maintenance

I think there is a massive period between Monarchy and Communism and military viable government type might be required to plug the gap a bit in order to shift the AI out of its Republican torpor! I fully understand if you are keen to just stay with the default government options.

 

[md4]

I generally only play for fun at Regent/Monarch difficulties, so any testing I do will be limited to those settings (also what I observe from my son's games at Warlord, but I think he's ready to level up soon anyway).

My original settings (which, from what I recall from what factors influence the AI's government choice), which from memory had Democracy cost 2gpt and Republic 3gpt for unit support (Democracy got some free units to compensate), had one game where every single AI switched to Communism in the late game. Amusingly, even America. In truth I've never really been interested in Fascism for myself or for the AI, though I have seen them use it from time to time in both my revised settings. Right now, the AI's preferred choice seems to sit around 80:20 between Democracy:Communism, which feels about right. Prolonged wars see them switch out of Democracy when the weariness hits and go to Communism mostly.

As for Imperialism, I think my concept of Feudalism somewhat fills that niche in the timeline (insofar as bridging Monarchy and Communism) even though it does it rather differently (lower corruption, no trade bonus).

On a slightly random aside, I did think there is some potential in @need my speed's idea for Republic: drop commerce bonus, hurry with whips and boost worker rate, so it becomes more an infrastructure government compared to monarchy's higher unit support and lower weariness. It does leave Democracy as the only trade bonus government in the game (which might open the door for your Imperialism idea as a contender), but it did look interesting nonetheless.

 

[Fergei]

80/20 for Democracy to Communism seems good, with the odd sprinkling of Fascism. I suspect that less tinkering is needed in the modern era than in the ancient one.

 

[md4]

Feudalism
Redesigned as a "better" Monarchy, forcing the player to choose between an early Monarchy switch or wait around a little longer for something with more future potential.

I've just had the belated revelation that this is very ill-conceived. I originally thought "Feudalism is in the 1st round of advances in the middle age, so it shouldn't take much longer to get there if you skip Monarchy and Republic, which are both optional". I kind of forgot that you can beeline straight for them, meaning that, you can get:

• Monarchy in between 40–45 tech points (depending on starting techs)
• Republic in between 42–49 (if you're first to Philosophy, then subtract 10 for Code of Laws, or 28 for Republic itself)

Feudalism, needing all non-optional ancient era technologies, comes in at 155 (minus starting techs, trades, etc) so it's not even close, and you'd probably be bad at the game if you waited on Despotism for that long. It's probably only an option for civilisations with the religious trait. Otherwise I don't know what to do with Feudalism anymore.

 

[Fergei]

Yes, this is why I moved the vanilla Feudalism government waaay forward to Code of Laws (and made Monarchy and Republic appear later). About half of Civs will take it but almost all will dump it as soon as Monarchy or Republic come along. The AI will then never revert to vanilla Feudalism once other non-despotism options are available.

Despite appearing powerful to me, the AI seems to hate communal corruption, preferring even problematic corruption.

The way you have rejigged Feudalism I would have liked to think that AIs involved in war would consider changing to it from Republic or Monarchy (since you made Monarchy have low war weariness). So it should still be viable in some circumstances? Perhaps the AI refuses to shift from one low war weariness govt to another during wartime?

If you removed war weariness for your version of feudalism I'd be extremely confident that war mongering AI would give it serious consideration until the advent of Fascism / Communism. So that might be the easiest change.

 

[md4]

Perhaps you are right, that it might still be useful to revert to in times of war (before Communism/Fascism comes along), but it would never be your first choice out of Despotism. So it's still going to remain very much a niche government.

I wouldn't remove weariness entirely; that's a feature of a feudal system where you are granting your regional lords (from which you draw levies) more autonomy. It would in no way have less weariness than Monarchy, and if there were 4 levels of weariness in the game, I'd have given Monarchy low and Feudalism medium, but one makes do with what they have.

 

[Fergei]

Even keeping low war weariness I think the AI will use your Feudalism quite a bit given Republic has high war weariness. I honestly think you are close to having it as a viable option with its superior unit support. What is niche for the human player and niche for the AI can be different - as the AI anarchy is less harsh and it is seemingly obsessed by unit support calculations.

I'd suggest doing AI tests with Mongol (shuns Republic?) and Japan (favours feudalism allegedly) and other random AIs and see what your results are starting from the middle ages. They will typically all stay peaceful until all the available land has been populated. Once that happens I'd be hopeful that a Republic shunning AI will have a strong chance of going to Feudalism and building up a bit of a military. This will then increase the chances of war and, consequently, of other AIs considering Monarchy or Feudalism.

The most basic/crude way to increase Feudalism uptake may be to increase the number of Civs which shun Republic.