Settlers

[etj4Eagle]

I have decided to give thought to a discussion on ways to improve the way settlers are handled in Civ3, including the AI routines. I know this is one of Zouve's favorite subjects, so maybe he will also enlighten us with his suggestions on how to fix the problem without crippling the AI. But here are my ideas

They fall into two areas: AI settler routines and general changes to settlers for all players so that the map is not instantly all settled. (Both elements are really needed to reign in the AI "settler diehrea").

Currently, at least to me, all AI civ's are quite similar to how they settle. If there is unclaimed land they will send a settler to it, not caring about anything.

First thing is to develop a routine to valuate new territory. Tundra squares with no hope of growth would be valued low, while a resource tile would be valued high. Coastal tiles would be valued high if the civ doesn't already have a city on that coast. An isthmus would also be valued high if it would save a fair bit of sea travel time. This valuation routine would not take into acount distance to territory (that comes later).

Second is to develop AI settlement patterns. At one extreme you would have the current extensive sprawl, where the civ does not care where the territory is located, just that it gets that land. At the other end is a more perfection minded behavoir. In this case the civ will only plant a city if it is adjanct to another one, it will avoid isolated cities. In this case the values from above would be heavily discounted for each tile more than a city radius the new site is from the border.

One step above that ultra perfection behavoir would be the allowance for overseas expansion (ie there is no discounting of territory value if there is no other city on that land mass).

Now each civ would have sliders enabling you to select how severe its sprawl/adversion to sprawl is. Thereby allowing you to have some AI civ's that will go everywhere in the map and others at the same time that will stay very close in.


While this addresses the AI city in that culture gap in your empire and the worthless cities, it does not address the shear numbers of settlers the AI will produce, which from the term Zouve uses is his real complaint. Now there are many players you play this way as well, as it is a valid strategy. Since you maximize score (territory) and the number of resources that you have under your control. However, I do agree that a slower growth might be better.

The game here does not match history, since we are playing for the end run, not the imeadiate. Also we don't really have the trouble of managing a large empire in the early days (there is corruption, but who cares when you are just claiming land for the end game). What is needed is something stronger to hold expansion down.

There are two ways to do this a hard limit or an increasing risk factor. In the hard limit you would be restricted to having only x cities before you discover a certain tech. And then this would increase to y and maybe even z before being lifted.

The other way to do it is similar but instead of saying you can't build that x+1 city, you can build it, but you might not want to. This can be either done by a chance of population dieing out, or a chance of the city going rogue, or a combination of both.

The chance of population dieing out represents that the fringes of your empire don't get all the funding and are not as well kept up as the areas around the capital. Hence as a result of that neglect, each turn there is a chance that some of the population may die to disease.

The rogue chance represents the weakness of your authority out there. They are essentially with you only because it suits their fancy right now. If the dies come up bad at the beginning of the turn, the city will go rogue on you. When this happens it becomes equivelent to a barbarian camp (however if it had gotten any defense bonus due to walls or being a city it keeps those). The garrison will become barbarian warriors of the same type. And each civilian will become a barbarian. Essentially you are getting a barbarian uprising there. And finally some of your treasurey will be lost, equal to if that city had been captured by a foreign power.

These two options will instill a risk in expanding out too fast. Since the closest cities to your capital under the limit will be okay. But then each additional city further out will have an increasing chance of the die coming up bad. Now there is reason to limit your expansion and it will also make colonies more useful. As why risk loosing a city and the hard work that went into building improvements in it, when you could have just built a colony to get that resource. The AI would of course need to be coded to recognize this limit and be programmed to only go after high value sites when other options will not work.



Any other ideas, improvements on how founding cities can be improved? Of course all values that I mention above would be exposed in the editor.

 

[Lt. 'Killer' M.]

Essentially, you are saying we need a penalty other than rampant corruption for sprawling empires - great!
But sady I do not see how this could work, since more towns will always mean more culture and more money and more production.... If all that, how culutre is evaluated and so on, can be changed into somthig that makes more4 snes, then we micht also see somthing done about the settlers.

How can a 20 city country with a temple in each city be judged more cultured than a singel city land with a cathedral, a temple, a library and a wonder?????? If you can't get away from that problem, reducing expansion will not help

 

[Dark Sheer]

A reply to Killer's question: Perhaps giving more cultural value to cathedral and wonders will help solve this problem?

For example, a temple has 2 culture while cathedral have 4 and a wonder might have something like 8-12? Of course if this is to be implemented then the 100,000 culture point for victory needs to be toggle to something much higher.

In fact, I have always wonder why a palace that is build in 4000BC have a much lower cultural value than a stupid temple or cathedral that is build much later? Comon, a 5000 year old palace is worth less than a 2000 year old temple?? Even with cultural rate double every 1000 years in Civ3, a palace's culture can never catch up with the temple.

 

[Qitai]

I thought the AI is playing it right to expand that fast. Civ2 was much less a challenge exactly because the AI always expand so slowly.

And if some lands are worthless, why would it bother you that the AI expand there??!!

 

[etj4Eagle]

Originally posted by Qitai
I thought the AI is playing it right to expand that fast. Civ2 was much less a challenge exactly because the AI always expand so slowly.

And if some lands are worthless, why would it bother you that the AI expand there??!!

That is my thinking on the subject too, however I can see other people's point. That is why also the primary restraint factor on growth is an element that would apply to all players (the idea here actually being developed when I was thinking of how to give value to colonies).

And that is also why I suggested a slider for AI settling behavoir. Since you could leave it as it is and let the AI rush like you in trying to claim as much territory as possible. It might not be worth anything for a few hundred years, but again we are playing for the end game. Or if you are so inclined, you could have the AI take a much more structured build approach. Or have a mixture of the two.

 

[alver]

Hi,

I'm not sure if this is exactly to the point, but an early problem in the game was the AI building cities close to other civs' lands, that were doomed to culture flip.

In the first patch this was changed, and I for one have noticed that this now happens less often.

It may be because of my culture-rich strategy, filling up holes in my territory, but I've never had a big problem with the AI's settlers. I find it a big improvement over Civ2, as Qitai said. The last thing I want is a change that makes the AI build less cities!

When it comes to making possible bigger differences between AI civs, though, I agree with the points suggested. It would be nice to see the civs get more of a personality.

 

[cracker]

I think the city placement routine could be substantially improved as well.

Currently, the AI does not seem to do a very good job of placing their cities for the long term big picture. When I am forced to go to war and capture all the cities from a nearby AI, then I end up abandoning a number of the cities to reposition them for the longterm maximum territory control.

The two most common problems I see are:
1) the placement of coastal cities that do not extend the maximum territory reach.
2) the isthmus/chokepoint placement thing. I use towns to fully block almost every choke point and I build the equivalent of the Panama or Suez canal whenever I have a position that will support it.

One thing I learned in GOTM7 is that blocking the choke points may not always be the best idea. In GOTM7, the Russians built a city next to a choke point and did not canal through the isthmus. I later figured out it was to get controll of a Uranium that only the AI's could see 4000 years before it was discovered. But the key point was the placement of the city next to the choke point allowed the choke point to be opened and closed by the movement of units through the choke point. It is a great way to waste lots of AI time and energy by just blockading the chokepoint with units for s few turns and then opening it up for a few turns. When the chokepoint is open, the AI will send 100 units marching toward it. When the chokepoint is closed, the AI units turn around and march back the other way. You can just keep them marching back and forth for turns on end if you watch this closely.

 

[etj4Eagle]

Originally posted by alver

The last thing I want is a change that makes the AI build less cities!

Very true, which is why I said that any real restraint on building new cities has to be a rule that applies to everybody. Something that would instill a more controlled growth. Of course the limits would be settable in the editor, so you could always set them to say 256 and have the change applied.

Or have it to default off (making this an alternate play mode), as the main reason why I am liking this change is because of the added strategy. Knowing that for now you can only build say 5 cities, where do you want to place them for best results?

WRT Cracker's comments. I wonder if the valuation table for the land is at all updated during play or takes into account surrounding land. I am guessing that it does not. Since if it did you would not have AI's placing cities on undiscovered resources (as the table would be revaluated once the resource was discovered). And the evaluation of surrounding land would move AI cities off resources and place them nearby instead.

Improving this valuation table would in of itself improve the AI's settling behavoir substantially. You would still get some worthless cities, but many more of the cities would be of great value.

 

[Scientia]

I like your ideas in the first part of your post on changing the settler AI logic. The second parts may require some deeper thought, however. I'm not all that favorable towards having a limit on the number of cities, or additional penalties for exceeding x number of cities (corruption is supposed to handle that, but it just ends up being an annoyance)

I definitely agree that the AI "land rush" needs to be toned down quite a bit. Ultimately, I think this is linked to the scoring system, where more land = higher score. Perhaps if there were provisions thrown in such that more city improvements = higher score, this would slow expansion down slightly.

 

[Zouave]

We first of all have AI cheating regarding Settler Diarrhea, which includes the AI seeing open tiles and probably getting freebie settlers as soon as some tiles open up (due to razing, or a destroyed town) which they instantly head for. They do this even when at war with me.

I destroyed three enemy settlers boldly marching right past my troops to that open tile. I destroyed them again and again, and it got rather boring after awhile. Can't the AI even learn you do NOT send settlers to open tiles when you have to march them past the military units of the civ you're at war with?! This is sloppy programming, and it should have been fixed in a patch, as I posted it as a bug long ago.

I am convinced the AI see those open tiles as recently a rival AI civ suddenly appeared on my coast near some tiles open due to a war in the early Industrial period, and its transport dropped a load of military and settlers there. BUT. . . when I went to autosave and went back several turns and left units on those coastal tiles. . . the AI transport NEVER SHOWED UP!! Case closed. It cheats. And that was on Regent.

In Civ 2 we had no problem with Settler Diarrhea; expansion came at the correct ratio between infrastructure and the need for new land. Some civs expanded differently as well we recall - the Aztecs and Chinese very little as they worked on infrastructure; the Mongols expanded a lot even though at the expense of infrastructure. OK, Chinese and Aztecs expanded somewhat too little, but the point is civs acted differently.

This Civ 2 VARIATION between different civs's modes of expansion was very good. But in Civ 3 they ALL have Settler Diarrhea, apparently at the same rate. This, besides its other problems, makes for a "samey" feel to the game.

Settler Diarrhea is ILLOGICAL at its core. Owing to the massive corruption (another problem) these far-flung distant towns, often built by the AI in awful terrain, will never produce much, and, during a war, be almost impossible to defend as thet are so hard to reinforce, and produce so little. No real leader would WASTE his time and shields pewking out crappy little towns like chicken pox all over the map. The only exception would be a military base preceding an invasion, and I don't see the AI doing that.

Borders are a good idea. And Culture should have some influence on them, although forget flipping cities and borders flipping over your resources. BUT, due to Settler Diarrhea, these settlers just wander through your territory, and, when they FINALLY, leave, magically teleprt themselves to the OTHER side of your civ to some open tiles they should not know exist!! Another cheat.

WHY SETTLER DIARRHEA??

The same reason Firaxis gave us such ridiculously low values for post-gunpowder military units: TOO RARE RESOURCES.

The scarcity of resources is as poorly implemented in the game as other ideas, such as Flipping. Their ARTIFICIAL and non-historical rarity, especially vis a vis iron and coal, is one reason settlers go tramping all over the board. It is an inherent design flaw that needs fixing.

WHAT TO DO:

1.Borders must mean something more. Wandering Settlers MUST LEAVE IMMEDIATELY BY THE SAME ROUTE THEY ENTERED. I should not have to spread out warriors and workers to try to block them.

2. Each civ should expand at different rates which will be converse to their infrastructure and improvement progress: a civ that expands little will concentrate on a well-developed society first, but one that seeks expansion will do so at the expense of infrastructure.

3.Colonies must NOT be permitted to be "overrun" by newly built towns of rival civs. Occupying even an ungarrisoned colony should be an act war..

4. Resource rates must be appropriately inceased, as many of us did long ago in the Editor. Trade also should be facillitated - a civ should not have to be robbed, or go to war, to get a strategic resource, at least not with iron, coal or saltpeter. No war in history was every fought over an iron mine; it wasn't that rare.

5. The degree to which a given civ continues to expand should be a factor of its original propensity to expand (as with Civ 2's Aztecs or Mongols), AND its current size and infrastructure. Also, if it is "scientific" it should emphasis techs development; commercial it should buold more roads; etc.

6. Terrain must be considered if the only open tiles known to a civ are desert and tundra FAR from the capital. Settlers should not go marching off there. If there's a nice piece of grassland or floodplain rather distant from the capital, well, then it might be worth it; same for a gem cluster. But what the AI does with settlers in awful terrain is absurd, and wasteful.

So, that's what to do, for starters. Don't hold your breath it will get done.

 

[Alphidius]

Zouave:

1) Completely agree here. Furthermore, Settlers/Workers without any accompany of military units in other civ's borders should have a chance of being assimilated. I mean come on, they're after all civilians and civilians inside the cultural borders after each turn which is like what, 5 years or more should have a chance of converting to the culture of the land they're in.

2) Agreed here too. It's not right watching the Indians expand so madly. Some civs should focus on building up their infrastructure, some building up on their city improvements and others in their military prowess.

3) Don't really agree here. If the AI keeps building colonies then the player doesn't get them even after building a city next to it. A better solution is that once colonies are 'eaten' by the borders of a civ, the civ with the colony will immediately open up a diplomacy screen to demand compensation, if talks fail, war may be declared without any reputation penalty for both sides.

4) Agreed. Important resources, like Iron and Coal vital to the construction of Railroads should be made available to the players easily. If trade is needed to obtain these resources, the AI should just give in cheaply, say 20, 30+ gold per turn? Otherwise the player could be handicapped and have to go to war to obtain a resource which is 'untradable'.

5) Agreed. Not only should it affect expansion but moves of military units and diplomacy too. Scientific will focus on building research generating improvements, more likely to offer trading of techs and strategic placement of units (eg. fortiying a defensive unit instead of attacking like they always do when under attacked), commercial will focus on overseas expansion? Maybe more likely to offer trading of resources and build a better naval force perhaps?

6) Agreed. Settler diahorrea is really sickening at times but more importantly seeing them build a city at the tip of a tundra surrounded area makes me