Option to regulate AI Expansionist tendencies

[Ecthy]

Playing the game I have noticed that even on lower difficulty level, the IA civs are getting global players too early.

Corrpution and waste as a barrier against early expansion to an obscene extent are OK, but only work if the AI is hindered as well.

They build worldwide empires even in the BC times, which should be changed.

Tonight I dreamt that there is an option which you can use for regulating the AI expansionism, just ike world size and AI civ number, an option in the 'start new game' menu.

If we were allowed to turn it down, the game might be more balanced, at least for beginners as it is now. And on higher difficulty, the game should be so hard that the human player is forced to be sort of perfectionist in the beginning.

I just can't stand it when people are able to build up huge empires in the early game, over and over again. It's not realistic.

Don't argue about realism, tell me what you think about that idea in general?

[Ecthy]

could a moderator please edit the thread title?

[Ralf]

Originally posted by Ecthelion
Corrpution and waste as a barrier against early expansion to an obscene extent are OK, but only work if the AI is hindered as well.

The AI-civs suffers from corruption/waste, just as much as you (Regent level) - but they keep on spreading fast anyway. And since they do it; the winning-eager player is forced to do it also. I dont mind that at all - this should be the game-default setting.

BUT partly I agree with your idea anyway:

Just in order to make it possible to variate early expansion-strategies somewhat, they should add a optional global editor-rule that enforces domestic road-connection before you (or the AI) found any new cities.
In other words: you must prepare in advance by building city-connecting roads to any planned city-placements.
If a newly founded city doesnt find itself immediately road-connected with all the other, already established cities (your own cities, that is); then that city automatically reverts back to settler-unit status again. The first founded city on any uninhabited island/continent is excused obviously.

This would slow down the whole process of early expansion, both for the AI-civs and the human player.

Don't argue about realism, tell me what you think about that idea in general?

I think its a good idea that could (and should) be implemented in that flexible game-editor of theirs. And it adds realism as well.

[Code Monkey]

There is no conceivable way to do what you are asking short of making the AI a pushover by setting up a rule which says something to the effect: "You cannot settle greater than (foo) distance from a capital if number of cities is below (foo2)" with a chain of these rules to define a suitably slow expanding empire.

The flaw with this idea is if the AI were limited in this manner then all the sneaky human has to do is plant a couple of cities far out from his capital in the early game such that AI cannot expand at all because the human has blocked up the only valid territory. The human player then has all the time in the world to fill in the empty territory he left between his capital and these border cities because the AI simply can't expand to the empty territory.

The problem is not the AI's expansionist tendencies, it's your inability to expand, and that is a "operator error" issue. I always keep up or outexpand the AI and I've been doing that since the second game I played.

It's not hard and only takes a rudimentarly understanding of city placement and tile improvement along with adding in a few granaries for population producing cities. Since the AI only understands the first two, the advantage is all yours on any difficulty below Emperor.

[Ralf]

Originally posted by Code Monkey
There is no conceivable way to do what...

Oh yes, there is. Read my first response. One just have tackle the problem with a little imagination.

[xane]

The problem is the "expansionist" trait is only a moniker for starting status and goody huts, all AI civs use REX and unless you do too you'll get caught out.

City "top ten" is now regulated on culture, meaning more wonders, not city size. A small civilization with multiple mega-metropolii is not a winner in the game any more, quantity not quality, et al.

[Code Monkey]

Originally posted by Ralf
BUT partly I agree with your idea anyway:

Just in order to make it possible to variate early expansion-strategies somewhat, they should add a optional global editor-rule that enforces domestic road-connection before you (or the AI) found any new cities.
In other words: you must prepare in advance by building city-connecting roads to any planned city-placements. If a newly founded city doesnt found itself immediately road-connected with all the other, already established cities; then that city automatically reverts back to settler-unit status again. The first founded city on an uninhabited island/continent is excused obviously.

This would slow down the whole process of early expansion, both for the AI-civs and the human player.

Interesting but problematic. First off, you can't settle new islands/continents ever and technically you can't even found your first city without a special case rule.

Assuming that you implement the additional function to make the basic idea workable you then introduce the real reason why this wouldn't work: The AI does not use any sort of high level control which would make this possible.

There is a level of control which sets weights for what each civ does (produce expansion, produce military, etc.), there is the level of control with which each city decides what it's going to build based on those civ-wide weights, then there is the level of control where each unit decides what it's going to do based on those civ-wide weights.

For your idea to work, there would have to be the ability for the high level to not just say "favor expansion" but say "we need a city at map coord 11, 118" and then, one, workers would have to intelligently build the road there, a settler would have to be queued up somewhere to roughly correspond with the completion of the road, and then it would have to be sent to those precise coords by the high level AI. There just isn't the provision for that sort of behavior in the engine.

Even assuming you find a way to implement the idea, the AI is still going to be utterly hosed by the fact that a human will make better long term plans every single time. That, and we'd just plant military units at the end of all these roads built in neutral territory and stop the AI from ever getting anywhere.

[Ralf]

Originally posted by Code Monkey
Interesting but problematic. First off, you can't settle new islands/continents ever and technically you can't even found your first city without a special case rule.

OK - lets compare:

Some Wonders-benefits only applies to cities on the same continent as that Wonder-city. These benefits dont apply to cities that you have founded on independant islands/continents.

Now - if the software can calculate which & how many of your cities that dont belong to the same continent; then I think the team could programme the software to calculate a special case rule for that pioneer galley/caravel-hiking settler as well.

The only difference compared with that Wonder-related this-continent-only rule; is that this special case-rule is applied once to that newly discovered island-situated pioneer city only - not to any spawned cities after that.

The AI does not use any sort of high level control which would make this possible.

How do you know? Read above Wonder-related exeption.

There just isn't the provision for that sort of behavior in the engine.

Whenever any AI-settlers stumbles across nice possible city-placements according to certain rules, they found their cities.

Whenever any early scouting warriors stumbles across nice possible city-placements, they could mark out these places with flags invisible to the human player, and then move on. Later an AI-worker build a road to that marked out place, and a settler follows the road towards that place, and found his city.

Even assuming you find a way to implement the idea, the AI is still going to be utterly hosed by the fact that a human will make better long term plans every single time. That, and we'd just plant military units at the end of all these roads built in neutral territory and stop the AI from ever getting anywhere.

You can do in the present game, as well. Just place out some early combat-units between foreign land and probable expansion-directions into no-mans-land. Whenever any warrior/spearmen-guarded settlers moves out; you just tag along and prevent them from "ever getting anywhere".

Mostly one dont do that, because A: you are occupied enough as it is with your own planes, and B: you dont want to trigger any wars at that early stage.

[Panzer]

Interesting but problematic. First off, you can't settle new islands/continents ever

Couldn't you use harbors for that? A city can be founded across a body of water as long as there is a harbor on the other side.

[Code Monkey]

Originally posted by Panzer


Couldn't you use harbors for that? A city can be founded across a body of water as long as there is a harbor on the other side.

Let's say you're the first person on the island, where does that harbor come from?

Not that I agree with this, you would have to have a check for special case scenarios; I just wanted to point out the obvious flaw in the ruleset he proposed.

As for Ralf's "How do you know?" regarding the AI: go read Soren Johnson's article here on Apolyton. The civ AI functions like a bee hive, an emergent phenomenon that has been called "Smart hive, dumb bees". Your invisible flag structure would actually work under the existing system but it's still too vulnerable to exploits and would require recoding of worker and settler AI as well as additions to every other unit's AI.

As an example of additional exploit since it escaped you, it only takes one military unit (or even a spare worker) under your proposed system to stop the AI from expanding to a given area AND you get plenty of warning with the fact it needs to build a road to it, currently it takes at least 5 units to completely shut down the wandering settler (I do it all the time and it's never caused a war, so long as the chump is on my or neutral territory the AI does not regard it as aggression).

The existing system works fine and the only thing artificially slowing down expansion will do is make the game easier for the player. It does not make sense to recode so much of the AI and game mechanics because newbies can't figure out how to make settlers quickly.

[kazper]

While I agree that recoding the AI for this is a bad idea, I nevertheless feel that there is something right here. The problem isn't the AI expansionism, but the fact that this is the only viable strategy. Unlike CivII where you could play for both quality and quantity.

And this is a real problem. I throughly enjoyed playing both styles in CivII and I think it's a great loss that there are literally only one way to play CivIII if you want to have a chance of winning.... Ironically it was my impression that Firaxis meant for CivIII to be more diverse than CivII, but this appears to me to have failed completely.

[N. Machiavelli]

Originally posted by Ralf
The AI-civs suffers from corruption/waste, just as much as you (Regent level)...

That is a sort of red herring. Sure the AI recives more or less the same corruption as you do, BUT they recieve bonuses to production, research, gold, starting units, and combat vs barbarians. These little 'extras' are more than enough to numerically give them an advantage in expansion, the only way to compete is to either get REALLY lucky with goodie huts or be 10x more ruthless than they are.

All these traits are now subject to change in the editor thanks to the patch. Try taking away all these little bonuses they recieve; it confuses the hell out of the AI.

[Badtz Maru]

Since pillaging improvements that are outside of a civs borders is not an act of war, the human player could virtually halt the AI's expansion by sending a warrior to hang out just outside the computer's borders and destroy any roads that he tries to make outside of them. True, they could just build cities within their borders, slowly pushing them out, but that would lead to even more dense and inefficient AI empires, making the AI a pushover.

I'd consider making it so that you could only build cities within your borders, if they sped up considerably how fast borders expand. This would make culture even more important, but would also lead to a high density of cities on the fringe of an empire once it's borders meet with an opponent as they try to force their borders out.

I think the current system works pretty well, I don't like the AI to have any restrictions that I don't have, and I sometimes like to build a city in a distant and isolated position. Those far-flung AI cities are always easy to assimilate anyway as the corruption keeps him from building cultural improvements fast enough to compete.

[Frugal_Gourmet]

How about?

How about making settlers cost more and more population as time goes on?

[Code Monkey]

Originally posted by kazper
While I agree that recoding the AI for this is a bad idea, I nevertheless feel that there is something right here. The problem isn't the AI expansionism, but the fact that this is the only viable strategy. Unlike CivII where you could play for both quality and quantity.

And this is a real problem. I throughly enjoyed playing both styles in CivII and I think it's a great loss that there are literally only one way to play CivIII if you want to have a chance of winning.... Ironically it was my impression that Firaxis meant for CivIII to be more diverse than CivII, but this appears to me to have failed completely.

This is mostly true and I'll tell you where they blew it: the removal of SMAC's terraforming possibilities and rules (and even some of Civ2's).

In Civ3 you have 3 improvements: mine, irrigation, and road or rail. Forest is never an improvement and w/ the new rules it's hard to see why it's left in the game at all. Mines do nothing to affect food production. There are no rules in place for individual climate per terrain grid or sea level elevation affecting anything. The result, particularly combined with the elimination of caravans/supply crawlers, means that a given city's production is fixed and the only optimising possible can be done by a butt simple worker algorithm.

There is no room for quality in developing cities. In SMAC I didn't have to expand like lemmings because I could use supply crawlers to bring in food and minerals. I could use terraforming to maximise the potential output far above what was initially possible. I grew cities up to size 60 and above regularly. And in the process, I traded the number of production centers for more refined cities. That's all gone and you and the AI are left with one way to increase production, revenue, and population: build as many cities as you can afford to.

The corruption model does stop this from being a geometric expansion of production and revenue but the culture model hoses you in another way. As others have pointed out, it expands too slowly to be of any real use for expanding territory - anything depending upon a border larger than 2 is asking for an AI to slip a settler in when you're not looking. End result, you either accept the AI is going to "pollute" your kingdom with its suicidal cities or you expand everywhere you can as well.

In the end, though, it's just another game mechanic to play with and, at least, under the major mechanics of Civ3 there isn't much to do about it.