CRADLE 1.35: Thoughts on game balance

[MrBaggins]

I've been playing a contrived Cradle 1.35 Impossible game for some time now; I started with equivalent techs and 3 settlers- to see how well the AI could 'race'. I placed a couple of civs in a somewhat isolated situation, some in close proximity and some on closely neighbored islands (on a 200x100 map.)

These are my findings during this game...


A while after 3000BC the AI effectively faltered and fell behind.

Many of their cities were rioting. A couple of the civs were having a number of their cities fall under barbarian control.

I took a look, and happiness issues were pretty profound in their cities.

The big issue?

* empire size penalties (due to conquest and/or expansion without thought regarding this)

compounded by

* pollution via population (the AI does about enough to deal with this... just... it can't cope with much more than this unhappiness though)


The AI's were tremendously capable in terms of growth, production (at 600BC, the AI that i'm fighting has 500 units vs. my 400,) and pw expediture (for the most part)

The AI is at a disadvantage, because it cannot deal with the concepts of 'too much'. It grows both in number and size of cities without thought. This *should* be a good thing... but unfortunately... it hits a brick wall.

Solutions?

* Help the AI out with empire size unhappiness

It doesn't know any better, so its not immoral to give the AI a helping hand, neh?

What I was thinking was keeping track of what government the AI was using, and how many cities it had, and use slic (as per disasters code?) to add happiness in each of their cities to make up for the difference between their government limit, and how many cities they found/capture.

* Have a few more 'pollution solutions'

I like the fact that pollution is an issue early on. I think that its a little too difficult to overcome. I think the solution is to increase the effectiveness of the first few pop pollution buildings (perhaps this might need to be balanced by decreasing the later buildings.) At first glance the production pollution buildings in late game also look a little lightweight (I know that this part of the mod hasn't been concentrated on however)


Last, but certainly not least...

* Wonders

Wonders are a classic... 'the rich get richer' problem.

Should they be removed for the sake of playbalance?

It wouldn't be popular, but it probably would be helpful in terms of competitive gaming.

The idea for a solution, I came up with in Ctp1 days, was limited wonders.

Essentially each civ gets (gifted) their own unique techs, once per age, which allow them to build a wonder. That wonder is built, independant of other civs. I was thinking, for varieties sake, of making the wonder that is built generic; that it gets replaced by a random, age appropriate wonder, when built. Since you can define how the wonders are gifted, you can use this method to make the civs more unique.

It seems more than possible to do given the state of SLIC in CtP2.

The wonders get obsolete after 2 ages, so only 2 are in effect (less conquest) at any one time.

Having scarer wonders makes them more personal and special. It makes them more wonderful, just like they should be.

Wonder races never made sense anyway- did they really compete to build the Pyramids?

Thoughts?

MrBaggins

[HuangShang]

the wonder thing sound nice, but it doesnt do much for gameplay

[MrBaggins]

why do you say that?

[HuangShang]

well... have anyone previously used it... u can include it in mods, i'll still play it

[MrBaggins]

thanks for the extremely useful and thoughtful feedback, Huang

[hexagonian]

Originally posted by MrBaggins
...(on a 200x100 map.)
Many of their cities were rioting.

I have set up Cradle to operate best on a huge map with 8 civs. The larger the map, the more the AI will expand and surpass the city cap. And although Ultra-Gigantic is still part of Cradle 1.35 (mainly because players who download earlier versions of Cradle have a file that has to be manually deleted), there is a warning in the options readme that spell out potential problems with that setup. I do not officially suppport that setup anymore.

Did you alter the Basic Cradle 1.35 setup or the 1.35 Ultra-Gigantic Map setup?

If 1.35 Basic, then your govern.txt file is not set up to handle the additional cities. You need to go into CRA_govern.txt and boost the city cap numbers in all of the government entries. If you used 1.35 Ultra-Gigantic, then the file that needs to be altered is CRAB_govern.txt.

Originally posted by MrBaggins
A couple of the civs were having a number of their cities fall under barbarian control.

You are probably using the 1.35 'More Aggressive AI' files. There should be a disclaimer on my site (my fault...) that these files do also make the barbarians more willing to attack all cities, both human and AI. If you are using a low barbarian setting, then these files probably could be used without having multiple civs fall to the hordes.

This issue was discussed in a thread, and I expressed my reservations about the alteration, but went ahead and posted the files because players wanted them. I would suggest going back to the files that are in the 1.35 main download.

[Maquiladora]

Originally posted by MrBaggins
* Have a few more 'pollution solutions'

I like the fact that pollution is an issue early on. I think that its a little too difficult to overcome. I think the solution is to increase the effectiveness of the first few pop pollution buildings (perhaps this might need to be balanced by decreasing the later buildings.) At first glance the production pollution buildings in late game also look a little lightweight (I know that this part of the mod hasn't been concentrated on however)

The solution is not to provide more buildings to deal with pollution but to make pollution less harsh. Why? Because its easier for the AI to cope with the second. If the AI has to build more buildings at the same time as building units, it will get out of synch and lose focus if it has more tasks and switch about, not good.

The late production buildings arent that light weight. Consider the numbers they deal with later in the game. ie. 10% of 100 production in 10BC is alot less than 10% of 500 production in 1700AD. Its also cumulative with all other production buildings, so it comes out a pretty big bonus for the short investment through the game.

* Wonders
Essentially each civ gets (gifted) their own unique techs, once per age, which allow them to build a wonder. That wonder is built, independant of other civs. I was thinking, for varieties sake, of making the wonder that is built generic; that it gets replaced by a random, age appropriate wonder, when built. Since you can define how the wonders are gifted, you can use this method to make the civs more unique.

I dont really like the idea of anything unique, in some ways it just makes the game 'not civ' anymore. Even though this is civ3's direction, i prefer a totally openended atmosphere.

Thoughts?

Yes

On barbarians,
I remember in Civ2 it was easier playing with "Raging Hordes" than it was with "Ruins only" for barbarians, because they caused the AI civs no end of problems. The barbs strongest time is early on with surprise attacks, its also the weakest time for AI civs. Playing a cradle game its not unlikely to see 4 out of 12 of an AI's cities controlled by barbarians, maybe they should be toned down in spawning... less units but more aggressive sounds good.

Building choices,
One thing that often stalls the AI in building happiness buildings before the unhappiness hits is they build pointless buildings, and often at the wrong times. Building granaries when the city is only producing 35 food, building bazaars when the city is only producing 50 gold etc. This could probably be helped by increasing the priority of the happiness buildings queue, perhaps even maximum priority, especially with pollution effects so harsh.

[MrBaggins]

Thanks Hexagonian

I used the Basic Cradle 1.35 setup, and changed the gigantic map scale.

I understand the concept of what you are mentioning; the govern.txt file, but firmly believe that it is not the ultimate solution to the problem that is evident...

The AI's are ignoring the empire size cap... period. That would happen regardless of whether you set it to twice the values. Of course, the issue would be delayed... you could probably delay the issue to beyond the end of the game. The issue with all of this, is the effect would be made for the human player also. This removes choice (government choice, conquest, etc) ... and limitation, at least in terms of empire size. This gives the human less of a challenge, not more.

In my estimation, its better to keep the empire size caps down, and fudge in favor of the AI's, to deal with their blind nature.


----

The Barbarians were gaining control of the cities through riot unhappiness conversions, not barbarian conquest; somewhere between a couple and swaths of cities have converting to Barbarian control.

MrBaggins

[MrBaggins]

Maquiladora>

Whats your view of limited wonders in general- the concept is viable with generalized wonders or unique wonders?

What would be the ideal situation that you'd like?

MrBaggins

[Maquiladora]

Originally posted by MrBaggins
Maquiladora>

Whats your view of limited wonders in general- the concept is viable with generalized wonders or unique wonders?

What would be the ideal situation that you'd like?

MrBaggins

Hard to say. Firstly I dont have the wonder movies on anymore, so theyre just another building to me. I agree with the problem of larger civs getting stronger (gaining the enabling advance first), especially the human player taking advantage of this.

One thing you could do is provide many wonders with a similar advantage, but if one civ gains a certain wonder first, they cant build anymore wonders that give a similar advantage. So the Pyramids and Stonehenge give production bonuses, i build the Pyramids first, so i cant then go on and build Stonehenge, taking advantage of my new production bonus of course. Just a thought.

[hexagonian]

Originally posted by MrBaggins
I understand the concept of what you are mentioning; the govern.txt file, but firmly believe that it is not the ultimate solution to the problem that is evident...

The AI's are ignoring the empire size cap... period. That would happen regardless of whether you set it to twice the values. Of course, the issue would be delayed... you could probably delay the issue to beyond the end of the game. The issue with all of this, is the effect would be made for the human player also. This removes choice (government choice, conquest, etc) ... and limitation, at least in terms of empire size. This gives the human less of a challenge, not more.

Actually, in my games, as the files are set up, I never ran into the problem of the AI exceeding the caps. Generally, the AI was below the cap, and I was always bumping into the cap. Again this is on the Huge maps/8 civs option. Even on the normal Gigantic setting, the AI was closer, but rarely over the cap. I would check the AI from time to time to see how it was managing its cities during the course of my games.

On the flipside, if you play on a small map, then the cap is never a problem at all. So the numbers do have to be altered, based on map size.

Besides, Cradle does give you the means to ignore the cap anyhow via the Raze city option, so changing the numbers may not be as drastic a measure (unless you choose not to raze cities).

And as a sidenote, with the 'More Aggressive AI' files, the AI is also more likely to attack cities, so the city cap may have to be increased to absorb this new dynamic. The city caps in earlier versions of Cradle were established by what was happening during the course of actual games.

If it cannot be done via slic, there is a way to give the AI the ability to ignore the caps, but it would involve a lot of work in altering multiple txt files - something that I probably won't do for Cradle.

RE: Barbarians
There is also a way to change revolting cities from Barabarian-owned to actual new civs. Go into userprofile.txt and change Maxplayer to 24.

I still think that many of those cities initally fall into Barbarian hands via conquest, and once they get established, they act as an actual civ and send out troops to conquer other cities - because I tested to see what was happening. In previous games, Barbarians never had cities, but once I plugged in the new Aggressive numbers, all of a sudden, there were a lot of Barbarian cities. (there is a thread in this Forum talking about this issue). It should be noted that the government city caps have remained constant since Cradle 1.3 and before.

In versions 1.31-1.32, there was a problem with revolting cities, but it was due to the fact that a SLIC file was causing the AI to override movement limitations for militias, so once the AI started taking slaves, it did not have garrisons in some of the cities to monitor those slaves, and they revolted. This problem was fixed in 1.34 and was borne out in my last playtest too.

Couple Barbarian conquest with the AI losing some cities via revolt due to its inability to monitor the cap (because the caps are too low when you increase the map size without altering govern_txt), you end up with a snowball effect.

[hexagonian]

By the way, here is the link for the AI strategies alteration. My wrapup comment is on page 4

HERE

[MrBaggins]

thanks... but where is page 4??

I checked in the cheat editor... and the cities changing hands ARE due to unhappiness. I can see where a swath of cities change hands due to revolt. I get the revolt message too...


Regarding the empire size cap... In the game i'm in, the Israeli's are on the same largish land mass with the Arabs. They both hit about 15-20 cities before they got into a pretty serious war (250+ units on each side) The Israeli's beat back the Arabs, and built settlers while it took over a number of cities to put themselves to 25 cities... 5 over their cap. They then continued building settlers and had, and continued to have happiness issues; running themselves into the ground.

I don't think its a bad thing that the AI was agressive, or its reasonable to have 30-40 empire size caps for early governments. It effectively means that the human player is not limited by empire size; so can ICS.

I personally do not believe a one-size-fits all empire size cap change works, when a programmatic solution can be made that gives just the AI the advantage, rather than everyone, human included.

MrBaggins

[hexagonian]

oops, I meant page 3...

If somebody could simulate this by a SLIC file, they could post it. I do not write SLIC though, because I do not know how.

But here is a way to do it in a non-slic manner. This should work, but it is a bit of work.

What can be done to help the AI overcome the cap is to make a duplicate set of governments. So for every normal government like

GOVERNMENT_(GOVERNMENT_TYPE)

you would also have a

GOVERNMENT_(GOVERNMENT_TYPE)_A (for the AI only)

The only difference is that there would be larger caps for the AI governments.

This would be enabled by an advance called ADVANCE_(GOVERNMENT_TYPE)_A.

ADVANCE_(GOVERNMENT_TYPE)_A would be enabled by the existing Government advance (for Republic, it would be ADVANCE_REPUBLIC) as well as an advance that only the AI gets at the beginning of the game (call it ADVANCE_AI_GOV). This would have to be triggered in CRA_DiffDb.txt. so the AI is sure to get it, and it needs to be enabled by something like ADVANCE_SUBNEURAL_ADS, an advance that cannot be researched at all.

Make the cost of each ADVANCE_(GOVERNMENT TYPE)_A very cheap, so the AI would not be penalized for having to research it - you would want the AI to research it in 1 turn.

You also want to prevent the human player from getting that AI- enabling advance.

So, this new AI advance could not be able to be traded (in CRA_strategies.txt, there is a section of the file that allows you to block trades of certain advances, and you would also have to change CRA_Risks.txt to block all free advances to make sure that it cannot be popped from a goody hut.

You would have to change all the government entries in CRA_strategies.txt to reflect the new governments that the AI would be getting and to enable the AI to use those governments.

You would also have to add the new advance entries in CRA_AdvanceBuild.txt to make sure that the AI would research the advances it needs to get those governments.

And make sure that you have all new advance and government entries in CRA_uniticon.txt. and CRA_gl.str.txt.

Finally, you would have to make all the changes to all of the same files that may be linked into the various Cradle playing options. (for example, each Cradle playing option has its own (CRA..._)uniticon.txt file)

As I said, a bit of work. I may have missed something.

[Martin Gühmann]

It is possible to prevent the human player from getting the extra advances, by using the mod_CanPlayerHaveAdvance slic function. The advance for the new government could be granted to the AI so it wouldn't need to research it.

-Martin

[Immortal Wombat]

If you integrate Dale's method of finding what government you're in, you can SLIC either happiness bonuses for the AI to counteract penalties, or happiness hits for the human to simulate them.
But this means a change to units.txt, and yet more SLIC. It may be easier to do it Hex's way.

[MrBaggins]

I'm thinking of it for a new mod... basically... I'm not sure if I need it right now, but if I did, I'd probably go with Dale's method, and SLIC the happiness, to be done with it.

[HuangShang]

Originally posted by MrBaggins
thanks for the extremely useful and thoughtful feedback, Huang

is that sarcasm?

[MrBaggins]

Well... was your feedback useful and thought provoking ?

[Peter Triggs]

What I was thinking was keeping track of what government the AI was using, and how many cities it had, and use slic (as per disasters code?) to add happiness in each of their cities to make up for the difference between their government limit, and how many cities they found/capture.

I'm pretty sure this can be done exactly as you've suggested. In fact I've knocked off a handler that should do it but it keeps hanging on me. It may be a problem with Dale's 'WhatGovernmentAmI' when you reload SLIC, because I know it works fine if you start from scratch.

Give me a day or two.

[Peter Triggs]

OK, try this. I didn't actually try it with Cradle. I'm currently running SAP so I tried it there (it works) and then made some changes so that it should work with Cradle.

You'll have to add the necessary Government specific settler units to CRA_units.txt. Here's an example:

UNIT_SETTLER_ANARCHY {
   Description DESCRIPTION_UNIT_SETTLER
   DefaultIcon ICON_UNIT_SETTLER
   DefaultSprite SPRITE_SETTLER
   Category UNIT_CATEGORY_SETTLER
   Attack 0
   Defense 10
   ZBRangeAttack 0
   Firepower 1
   Armor 1
   MaxHP 10
   ShieldCost 0
   PowerPoints 0
   ShieldHunger 0
   FoodHunger 0
   MaxMovePoints 100
   VisionRange 1
   ActiveDefenseRange 0
   LossMoveToDmgNone
   MaxFuel 0
   NoZoc
   GovernmentType GOVERNMENT_ANARCHY
   SettleCityType UNIT_CITY
   SettleSize 1
   CanBeExpelled
   CantCaptureCity
   DeathEffectsHappy
   BuildingRemovesAPop
   OnlyBuildOne
   IsSpecialForces
   Civilian
   SoundSelect1 SOUND_SELECT1_SETTLER
   SoundSelect2 SOUND_SELECT2_SETTLER
   SoundMove SOUND_MOVE_SETTLER
   SoundAcknowledge SOUND_ACKNOWLEDGE_SETTLER
   SoundCantMove SOUND_CANTMOVE_SETTLER
   SoundAttack SOUND_ATTACK_SETTLER
   SoundWork SOUND_WORK_SETTLER
   SoundVictory SOUND_VICTORY_SETTLER
   SoundDeath SOUND_DEATH_SETTLER

   CanSee: Standard
   MovementType: Land
   MovementType: Mountain
   Settle: Land
   Settle: Mountain
   Size: Small
   VisionClass: Standard
}

You need one of these for each Cradle government type. See the handler 'GOVT_SetupArrays' for the names you have to use.

Also you have to go to CRA_gl_str.txt and, in the units section, for each of them include a line like:

UNIT_SETTLER_ANARCHY                            "Dummy Unit"

Finally, the happiness bonuses that this handler gives might be a bit excessive: it gives them across the board to every city in an AI civ that's exceeded it's TooManyCitiesThreshold. See what you think; they can be changed.

[child of Thor]

Nice suggestions recently.

I don't have a problem with the AI 'cheating' its number of city limits - to a certain degree. But i don't want it to sprawl ICS style, which IMHO is one of the worst traits of the tbs game.
If the unhappiness bonus could be fudged to help out the AI then that is good, but i think a production and growth penalty should count for the cities over the limit determined by goverment type.
As far as the player is concerned i sometimes think that the penalties for going over you city limit(and size for that matter - i mean size 30etc) should be harsher.
If you think about it how many pre/ancient civilisations could effectivly manage their empires once they had reached too(this is an x number of cities which varies from civ to civ) large a size. With the Mayan's it caused massive civil war and led to their downfall, the Greeks,Romans,Mongols did manage it to greater or lesser extents, but they had very advanced lines of communication and invested heavily in infrastructure to support it.
The Wonders are tricky, it wouldn't be a civ/ctp game without them.
Not so sure on civ specific but I like Maquiladora's suggestion,which would be a halfway house i guess, of having more wonders available at the same time with roughly the same bonus, but with some kind of limiting factor. Like only x amount per age/epoch, maybe with reduced power for most of them(keep a few good must build ones), maybe a higher upkeep cost(i think they are free at present) etc. Still this approach would mean a lot of other things would need to be tweaked to take it all into account(especially Wonder upkeep). Nice Idea's Mr Baggins

[Peter Triggs]

Mr Baggins,

Did you try this out?

I'm surprised none of the other SLICer's haven't had a look at it and found the Easter Egg.

[MrBaggins]

Thank you so much for the code, Peter... its much appretiated...

I am, at the starting point of a seriously big mod project for CtP2, as we speak, called 'Poor Get Richer'... linkified here. I've been testing minutae to see exactly how much is possible. I've not incorporated *this* code, yet... but will... as soon as the (new and fully modifiable) colonization stuff is done.

MrBaggins

[Immortal Wombat]

er... Easter Egg?
Did somebody shoot the easter bunny, or is 2am a bad time to be SLICing?

[Peter Triggs]

IW,

2AM is a bit late. I'm gonna hit the sack. But, ever seen this before,

        government[0]=GOVT_WhatGovtAmI[player[0].owner];
        maxCities=GovernmentDB(government[0].type).TooManyCitiesThreshold;

You know how we've been trying to access the fields in the Databases for ages, but all we got was syntax errors? Basically, MrOgre slipped up on the documentation and, although it's a bit hit and miss, I've been having some luck with getting through to those fields.

[MrBaggins]

I noticed the maxCities=GovernmentDB(government[0].type).TooManyCitiesThreshold; line... and thought itlooked a little... useful...

*chuckles*

MrBaggins

[Martin Gühmann]

Why government[0].type I thought you do it like it is done in the Great Library:

GovernmentDB(Government[0]).TooManyCitiesThreshold

-Martin

[Immortal Wombat]

Hey, thats really cool. Good discovery Peter

I'm sure there was something I was going to do once that was sorted, but I cannot for the life of me remember what it was.

[Peter Triggs]

Yes, that works too. The thing is that MrOgre wrote:

A record in any of these databases can be accessed by name or by index. ...

But it doesn't stop there. Many fields from each database are also accessible. ...

Examples:
tankAttack = UnitDB(UNIT_TANK).Attack;
costOfAdvInfantryTactics = AdvanceDB(ADVANCE_ADV_INFANTRY_TACTICS).Cost;
minimumHappinessForDefaultStrategy = StrategyDB(STRATEGY_DEFAULT).MinimumHappiness;

The problem is that these examples don't work: they give syntax errors. Although it's true that a record in any of these databases can be accessed by name or by index, evidently you can't get through to the fields of a record if you try to access them directly via the record's name. That's why I used "GovernmentDB(government[0].type)", the ".type" suffix always maps to the builtin's database index. But it seems, at least in this case, to be unnecessary.