Love for the AI: AIL(ove)-Mod

Update 7.5 from 2012-01-06:

http://www.civforum.de/attachment.php?attachmentid=660042&d=1325872516

Assault Infantry now has 12(6)-6-1 instead of 10(5)-5-1
Submarine now has Stealth.
Nuclear Submarine now has Stealth.

Removed the effects of global warming.

Update (7.4 from 2012-01-05, see post later in this thread for changes):

http://www.civforum.de/attachment.php?attachmentid=659402&d=1325718871

Fixed a bug with Leonardos Workshop and Copernicus Observatory to require 5 forges in total instead of 1 forge in the building city.

Duration of Knights Templar has been extended from Steam-Engine to Replaceable Parts.

Towns now can reach a size of 8.

Cities now can reach a size of 14.

Granaries now in addition to their previous effect give +1 Content-Face.

Temples now give 2 Content-Faces up from 1.

Theatres now give 3 Content-Faces up from 2.

Courthouses now give 2 Content-Faces up from 1.

Reason for these changes:
Making the settling-pattern the AI uses more attractive by making it easier to use the tiles in the city-radius more quickly.

I'd like to share my little mod with you:

To see if this could be for you I'm going to explain a little about it.

As the title states the Mods main purpose is to improve the AI.

This is by no means an epic mod that has a lot of stuff. No new media, only a mere biq-file consisting of nothing but rule-changes.

Development on this mod began back then in 2003 when I played Civ 3 a lot. I just recently picked it up again and made some further improvements.

So what did I focus on?

There's 5 fields in which I felt the AI to completely suck in the original game and those are what my mod focused on:

1. Defending rushes
2. Improving tiles
3. Managing corruption
4. Industrialization
5. Handling artillery-units

The measures undertaken to fix those problems are relatively drastic. As it was more important for me to get rid of them than doing it in a nice way.

Let me explain reason and my solution to the various problems:

1. Defending rushes

The biggest reason for the AI to be so bad at defending rushes and performing them on their own were wonders. They would have 4 of their 6 cities trying to build an early wonder even if the cities production and thus the chance of succeeding was very low.
Since wonders are too important for this game I could not simply remove them so I found following solution:
I integrated a new building, the forge. It is available right from Bronze-working, costs 80 shields to produce and 1 gold in maintainance.
Only cities that have built this are qualified to build any of the early wonders.
The AI reacts to that by, well not building it until it feels it's worth it. Which usually is at around 8 shields per turn.
In my games with this mod the first wonders usually weren't built before monarchy/republic. And then only by cities who can build them pretty fast.

To compensate for their later appearance most of the wonders got their expiry date removed or pushed back considerably.
Only one of the wonders was changed completely: The Great Libraby: It now gives 2 Civilization-Advances and a permanent Research-Bonus for the city it was built in.

This change leads to the AI producing way more units and actually usefull infrastructure like barracks and temples early on. Rushing is way harder and the AI with an abundancy of units also is more likely to rush itself. Beware!

2. Improving tiles

The way how the AI improved their tiles was very lackluster. They very often would just choose the bad option.
And how can I affect that? Well, by reducing these options.
Might be hard to get used to but now every tile can only be treated in one way.
This also affects how city-locations should be planned.

Here's how:
Grassland and Plains are limited to Irrigation.
Hills are limited to Mines.
Forrests can no longer be removed but instead also can be mined.
You now can plant forrests in Tundra right away to make these areas more usefull.

The AI will simply do what it can do with each tile and the player has no advantage here since he only can do the very same thing.

3. Managing corruption

Not building Courthouses and placing the Forbidden palace right next to the capital is what the AI did.

How to handle this? Well, I never thought of corruption as a fun game-concept so I removed it alltogether!

All governments were altered slightly to still differ from each other, now that corruption was gone.

I also removed Facism as it did not fill in a role anymore.

I won't go into too much detail here, it basically is like that:

If you want to wage war, you play in Monarchy->Communism.
If you want to build and research quickly you play in Republic->Democracy.

Also note that besides the governments you now no longer can build a new palace. This is to prevent you from cheating on the AI by "prebuilding" a wonder.

Warning: This change has a drastic consequence which should be accepted as it is for anyone playing this mod: Infinite City-Sprawl actually is great strategy. There's absolutely no reason to not do it. Filling in each gap with a city will never be disadvantageous.

4. Industrialization

For some reason the AI sometimes just would not build those damned factories despite the cities being actually quite productive already and would greatly benefit from a factory. This often caused the AIs to fall behind in the industrialization-age and made achieving a military victory in the later stages of the game far easier than it should be.

Here I actually found a fix without changing the rules for the player at all.

By using Flavors I just made the AI love building factories so much that they would just get them immediately in all but the most unproductive cities.

5. Handling artillery-units

The AI was totally inept to use land-artillery-units even remotely as effective as a human could.
I also have not found a way to achieve that.

So what did I do? I removed them completely.
A harsh step but I felt it had to be done.

How did I compensate for this:

The Koreans got a new UU, the Koryo Hwarag, wich is a Longbowman with the same stats as a MI but without requiring iron.

The attack-unit-gap between Cavalry and Tank was bridged by 2 new units:
Assault Rifleman 8(4)-4-1
Assault Infantry 10(5)-5-1
I swapped around a little between the models so it looked more consistent and fitting.

All naval and air-bombard-units now have leathal bombardment for land units!


So, while these 5 aspects covered most of the changes in this mod there were some more minor ones. I do not remember all of them but I'll try:

Marines 14(7)-8-1 and TOW-Infantry 12(6)-14-1 got buffed slightly to not make a player without oil completely chanceless against someone with.

Cruise Missile got buffed as well. With 16x4 It now can take out most units but the very strong ones like Battleship, Modern Armor or Mech-Infantry.

Colosseum now is called Theatre and comes with Philosophy.

Ironclads no longer have their own tech. Instead Railroads have their own tech.

The AIs split up their research via the use of Flavors. However, all of them get vital stuff like Iron-Working, Feudalism, Industrialization, Railroads and Sanitation with high priority.

Irrigating without fresh water now is possible with Construction.

It is no longer possible to rush production in a city by any means. (AI was so eager to kill their own pop and treasury for that even if not really necessary)

Conscription also was removed for the same reason. (Not the tech, which still allows mobilization and Riflemen/Assault Riflemen, only the ability)

Golden Ages were removed. Beeing able to trigger one with an early UU made AI-Boni look like nothing compared to it.

I think that's pretty much it.
Might have forgotton some minor-changes like +25% research required to compensate for the increased RP-output due to the removal of corruption.

My assumption on how this mod affects the difficulty-level of the game would be at around 2 steps. While I'm not the best player out there, I find it to be pretty challenging on Monarch and with a subpar starting-location even Regent was loseable. I've yet to succeed on Emperor and think anything above that might be almost unwinable. (All Stories and AARs I've seen to beat Demi-God and above based around exploiting techniques and AI-weaknesses no longer present in this mod)

So if someone actually dares to try my mod out, feedback would be much appreciated.
Of course the changes as described in this post can also be base for a discussion without having to try it out.

This is a good mod and I have been thinking of doing something like this as well.

You can change the rules, units, techs all you want but the AI is still gonna play the same unchanged game. That's why a mod that shows love to the AI is so needed.

You have made the game significantly more challenging just by removing some choices where the AI usually made the wrong choice (irrigating grassland, going into Fascism, conscripting units, etc). That's very smart.

Not so sure about the corruption change yet.

One thing I would like to see addressed is the different settling styles of the AI and the player. As we know ever since the discovery of ring city placement, the player likes to create rings of cities, even as tightly packed as 8 cities at a distance of 4.

The AI does NOT do R.C.P., instead they have a very spread out city placement, with minimal overlap. This puts them at a significant production disadvantage if the player gets his empire up and running without interference. Hence the advanced AI is programmed to be very aggressive, and get huge production bonuses and lots of starting units. Which, again, it ONLY NEEDS because the player outsettles and outproduces the AI.

I was thinking of a mod that changes the food, shield, and coin values of terrain so that the way the AI is hard-coded to settle, actually becomes the ideal, and the player is forced to adapt and settle like the AI.

Any thoughts?

I haven't followed the evolvement of strategies over the last um... decade. So I didn't know about that RCP and while I sure settled a little tighter than the AI, it did not make that much of a difference.

I really like your idea of making the AI's settling stlye the way to go.

It actually is suited to be the most efficient... in industial- and modern-age, when there is a lot of buildings and a city uses all of its tiles. Early on, when limited heavily by unhappiness and the mere lack of food it's just better to have more of them.

So the way to go would be increasing food production and reducing unhappiness alongside changing the neccessity of an aqueduct to grow beyond 6 (or make it earlier and cheaper).

Specialists could also be increased in effieciency so that maxing out on population becomes even more desireable.

A main problem I see with this approach is that specialist never can be unhappy and thus it's not possible to simply move the unhappiness to later... It would become pretty meaningless together with all concepts to fight it.

Another approach I can think of would be giving small cities severe disadvantages so that the bigger ones are what you strive for. Like making everything but the very basic buildings require a certain size so that you cannot build a barracks, marketplace or library in these small cities. Unfortunately there is no direct link to the size of a city as requirement for something.

I will put some more thoughts into what could be the most elegant way to accomplish this.

I have some ideas but I have to try them out in the Civ3 editor when I get home. I like the idea of forcing players to make large cities by making small cities pretty worthless. Right now the advantage to many small cities is the free square (free food + shields) provided by the city itself and that all else being equal many small cities can produce as many swordsmen as one large city and the small cities don't have to worry about unhappiness.

I've done some research about the RCP-issue.

It's main purpose is to reduce corruption through an exploit of a logical error in it's calculation.

See this for further explanation: http://www.civfanatics.com/civ3/strategy/ring_city.php

Since my Mod has no corruption anyways, this issue is more or less neglected and it doesn't matter if you place them at the same distance or just pretty tight in general.

This leaves the issue that tight colonizing gives you more use out of the same amount of land in earlier stages of the game. Mainly because unhappiness prevents cities from growing and using the same amount of land with less cities.

This early-game advantage will turn into a late-game disadvantage as in: you need more buildings to improve the usage of your land. A similar thing but to a far lesser extend goes for military-police.

Also the Governments with their amount of free units per city-size play a role in making more small cities less attractive.

So to sum up the things you can do to make AI-settling-pattern more attractive without even touching happiness and resources per tile:

Remove corruption.
Increase the amount unit support costs.
Decrease the free support for Towns.
Increase the free support for Cities and Metropolis'.
Increase the support-cost of vital buildings, that benefit most from such a strategy, like barracks.
Increase the effectiveness of these builings to make them more desireable at the same time like increasing HP on veteran-units.

Now when touching happiness comes into play, I guess the best bet would be increasing the effectiveness of happiness-buildings.

Thanks for creating the mod, AIL.
You added very cool solutions and I have to try them out in your mod and in my own too (how can you say you love Civ3 and don't create your own mod??? ), except of 1, half of 3 & 5 (according to the post #1).
I already got an idea of 1 and implemented it.
I already thought on the 3 and implemented it partially, but it's known idea.
About 5 -I'm still in love with arty, so I can't remove it. I can't, I don't want to.

Solution 2 is very interesting. What about jungle? IIRC forests don't provide a disease, but jungle do. So, probably, a disease should be removed?

4 interesting, need to try it out.

Finally, thank you once again for providing new ideas and creating a mod.

The CCM mod has very interesting concepts about what to do with the atrillery units. If it shows to be working well, I'll definately bring them back.

Yes, CCM has a lot of interesting concepts, but it doesn't fix "AI & land arty" problems. It just limits a human player in number of artillery units - not to have a crucial advantage using them. Anyway, I wanted to recommend you this mod, but I thought you already saw it - it's one of cult works.

Wolfshade said:

Yes, CCM has a lot of interesting concepts, but it ... just limits a human player in number of artillery units - not to have a crucial advantage using them.

Click to expand...

Thank you Wolfshade for your kind words about CCM, but about artillery in CCM I have a different opinion. In my eyes it´s the closest try to let landartillery work as it should and due to the stealth attack and retreat from infantry options it has a crucial advantage when using it. Of course it is not any longer an exploit for the human player as in normal C3C, but this is intended.

Civinator, thank you for clarifying things out. I know there's a lot I need to learn from CCM. So my words aren't kind only, they're true.

Wolfshade said:

Civinator, thank you for clarifying things out. I know there's a lot I need to learn from CCM. So my words aren't kind only, they're true.

Click to expand...

Wolfshade, I hope soon we will have the public betatest of SOE. Here you can see from start with masses of these units, how landarty with these settings works. When looking at SOE in DEBUG mod, it seems to see a film of WW 2. Duels of landarty against each other and of landartillerry softening up enemy defense positions are a big part of it.

The reason, why artillery in CCM and SOE is autoproduced is: Landartillery has by far the highest attack value of landunits and therefore the AI would produce artillery over and over and nearly no other landunits. The autoproduction-only for landartillery in CCM is for reasons of balance in gameplay and not to restrict the human player.

AIL said:

I've done some research about the RCP-issue.

It's main purpose is to reduce corruption through an exploit of a logical error in it's calculation.

See this for further explanation: http://www.civfanatics.com/civ3/strategy/ring_city.php

Click to expand...

The RCP bug was fixed for Conquests. I'm not sure whether it was fixed for PTW though.

This leaves the issue that tight colonizing gives you more use out of the same amount of land in earlier stages of the game. Mainly because unhappiness prevents cities from growing and using the same amount of land with less cities.

This early-game advantage will turn into a late-game disadvantage as in: you need more buildings to improve the usage of your land. A similar thing but to a far lesser extend goes for military-police.

Also the Governments with their amount of free units per city-size play a role in making more small cities less attractive.

So to sum up the things you can do to make AI-settling-pattern more attractive without even touching happiness and resources per tile:

Remove corruption.
Increase the amount unit support costs.
Decrease the free support for Towns.
Increase the free support for Cities and Metropolis'.
Increase the support-cost of vital buildings, that benefit most from such a strategy, like barracks.
Increase the effectiveness of these builings to make them more desireable at the same time like increasing HP on veteran-units.

Now when touching happiness comes into play, I guess the best bet would be increasing the effectiveness of happiness-buildings.

Click to expand...

Another few things to make ICS less attractive would be:
* Increase shield costs for units
* Add a cheap happiness building: +2 or 3 content faces for 10 shields, no support cost.
* Add a cheap shield multiplier building - maybe 20 shields for +50% shields.
* Make Hospitals available in the AA.

If possible, require that the buildings be built in a City, not a town -- make size 7+ more attractive, and you'll reduce ICS rather naturally.

I did not read all of the first post, about 3/4. That said I think this mod is geared towards below emp play. The wonders and factory and unit counts are not issues in Sid games ( donot address deity as I seldom played it). Rushes are not really a human tool as they will have more units than you will.

Improving tiles issue is a function of lack of workers, more than lack of skillful use, which they do not have either. Most weask players have the same problem, they do not have enough workers, nor do they use them effectively. That puts them on par with the AI.

Removing corruption is going to be too strong for the AI on Sid as they get such cheap settlers, they will grow far too large for the human. This was less of an issue in C3C as most of those towns will be too corrupt to make much of a difference.

Artillery, well this one I am not sure. I tend to not make much use of them in any level or play style as they are a pain to drag around and shoot. I tend to have a few on any given landmass to ping ships and invaders. Not to attack towns.

I do not care about removing or not, but players such as you feel it is too strong, just put in the control that you will not build them. Like we do in AW games, we self impose restrictions. No need to break a lot of things by removing Gov and adding units.

BTW do not want this to be taken as a slam on the mod, just another perspective. You want to slow down the human player, remove armies and make settler/workers harder to come by. This is what is done in CCM.

I am not sure how one is to enjoy playing the same as the Ai, which is what I see as the results of these changes.

I forgot to mention that I and some others put up the FP close to the palace as well as it does not have enough impact to anything other than the OCN to justify taking longer to build farther away.

Prebuilds are not hard to come by at lower levels and at Sid they are not a concern for me. I only use one for the GLB and it often is only a unit as I do not know Masonary at that point.

Civinator, OK, I'll definitely follow it, if it'll be possible - take a participation in it.

@Elephantium:
Oh, since I just learned about RCP, I did not know that it already was fixed. Or maybe I once knew it and have forgotten.

As I said, I have enough ideas how to reduce the attractivity of many small towns against fewer big ones. And they are less drastic.

However, I still have to first try it out, how much of an issue it actually is. I mean you already really are forced to make a huge benefit out of this strategy before AI cities use most of their area.

@vmxa:
Nice to still see you around after almost a decade. I remember having read stuff from you before. ^^

Of course the Mod is "geared towards below emp play". It's whole purpose is to make the game challenging without having to give the AI absolutely ridiculous boni.

On Regent/Monarch the AI will always die to your rush when it currently is or was trying to get wonders.
On Regent/Monarch the AI would always fall way behind in production once in Industrial age as it just did not build the factories.

In my opionin I think that Sid should not be possible to win at under even remotely even starting-locations. And I'm pretty sure it ain't in my Mod. Of course removing corruption is too strong for the AI on Sid. It's helping the AI on Regent/Monarch!

Infact this Mod is kind of a self-restriciton since I mainly made it for myself.

I do not want to slow down the human player or make settlers/workers harder to come by. As you said: That's what CCM does. I have a different approach but the same basic goal: Making the game harder on lower difficulty levels than those where the AI-cheating becomes absolutely ridiculous.

You are right: I'm taking away options only the player used correctly, which more or less makes you play the same as the AI. You still can decide your unit-composition, the order in which you build buildings, in which you research technologies, how to distribute your units, how, whom and where to attack and so on.

Yes, it is a little dumbed down but not so much that I'd lose any joy in playing it.

I play on Emperor/Demigod usually, but I dislike the higher difficulty levels in principle because they aren't making the AI better, they are just giving it freebies. I think the idea of this mod is better. You will play on Monarch (the "equal" level) but the game will be harder than a normal Monarch game, by far, because the rules will be changed to advantage the AI's natural style of play.

I would suggest two changes to tilt the game towards few, big cities (aka AI-Style). These are drastic changes though.

1. Larger Cities. Keep the aqueduct and hospital where they are, but move the city size caps from 6/12 to 8/16, or even 9/18 maybe. Thus you can have large productive cities in the Ancient & Medieval ages.

2. Replace Corruption With Empire-Wide Unhappiness. There is no more corruption. Instead the Palace places a Town Hall in every city. The Town Hall can't be sold (because it is generated by the Palace small wonder).

The Palace makes 3 happy faces in EVERY city, BUT,
EACH Town Hall makes 1 unhappy face in EVERY city!

So for example if you had 4 cities (including your capital) you would have 4 Town Halls which would generate - in each of your towns - 4 unhappy faces.

Under this system it is MUCH better to have few, large cities because each new city contributes unhappiness to ALL of your cities.

Remember that this is supposed to be played on Monarch, standard map. This system would REPLACE the AI bonus where you start with fewer content citizens in Emperor and beyond.

This large change would have to be paired with some of the small changes AIL suggested like increasing the unit support.

Would really like to hear feedback on this! A lot of great thoughts in this thread so far.

TheMapDownloade said:

2. Replace Corruption With Empire-Wide Unhappiness. There is no more corruption. Instead the Palace places a Town Hall in every city. The Town Hall can't be sold (because it is generated by the Palace small wonder). Would really like to hear feedback on this! A lot of great thoughts in this thread so far.

Click to expand...

A small wonder cannot give virtual buildings to other cities.

AIL I am always glad to see people coming out with mods and it should fit the bill for a number of players.

Increasing City-sizes to 9/18 ain't that big of a change afterall but the 2nd one is really, really drastic.

The Palace ain't a small wonder, it's an ordinary building. The thing that makes it unique to each civ is the "Center of Empire" flag.
And you can easily change it to a wonder so it could work as described.

However, this has no counter and the AI, should it ever have too many cities would certainly not benefit from it.

Really, I think making it a little easier to get cities a little bigger faster by small changes like giving the granary a happy face, giving the temple 2 and the colosseum 3 happy-faces, increasing town-cap to 9 and voila colonizing tighter than the 2-3 distance would hardly be an advantage anymore. Especially if you give Towns a poor supply-number and cities a way better one. Nothing too drastic required, really.

AIL said:

The Palace ain't a small wonder, it's an ordinary building. The thing that makes it unique to each civ is the "Center of Empire" flag.
And you can easily change it to a wonder so it could work as described.

Click to expand...

Sorry it´s not so easy. A Great Wonder can only exist in one location, but the game gives a palace, even if it´s a wonder to any civ. Here you have for all but one civ a virtual building, that gives a lot of other virtual buildings. This is possible but sideeffects are not researched yet. I made an attempt in this direction with my multiple palace thread many years ago.