Era of Miracles fantasy mod - developer diary

[Ahriman]

The problem is, as always, the AI, which probably won't be able to put the elite units where it needs them the most...

Right, but the AI can somewhat compensate by just having lots and lots of units. The AI won't be efficient at combat, but it can have big enough military production bonuses to field a big army.

I'm thinking about changing the food requirement per citizen from 2 to 3

Its worth considering. Keep in mind that this kind of change means a whole economy overhaul, which isn't necessarily a bad thing.
A good goal would be to try to get food, production and gold to actually be equal in value. So for example, farms could give +2 food, +3 with fresh water and a midgame tech or with a late-game tech, mines could give +2 production, +3 with fresh water and a midgame tech or with a late-game tech, trading-post equivalents could give +2 gold, +3 with fresh water and a midgame tech or with a late-game tech.
Adjust buildings so that their maintenance costs are moderate.
Adjust trade income from cities downwards.
etc.

I hope the "half food per specialist" policy effect (I'm going to call it "Spiritual Nourishment") will work correctly with this setting, changing it to 1.5 food.

I think that this would probably be too strong as a policy, I would think that working as intended might require only -1 food per specialist. But that is the kind of balance tweaking you can't possibly assess without playing the mod.

Force civilizations to develop around high-production resource tiles.

One other way of doing this is to encourage different types of terrain to be high-yield for different civs. Some civs might like grasslands with fresh water, others might like forests, others might like hills.
I'm skeptical though that encouraging cities to be really spread out would necessarily work well. First, the AI city placement algorithm isn't designed to do that. The AI also won'd do a very good job of militarily defending a scattered territory. Road maintenance costs can get really high with a spread-out territory, and transport time can get long. I think it is find to hard-code 3 minimum tiles between cities, but I wouldn't do more than that.

 

[Pazyryk]

One other way of doing this is to encourage different types of terrain to be high-yield for different civs. Some civs might like grasslands with fresh water, others might like forests, others might like hills.

Having favored terrain is fine. My point though is more about land development density. In Civ4 and in FFH, you basically work very quickly to a sort of post-2050 suburban sprawl. Suburban sprawl as in Los Angeles area were one developed city runs into the next with no undeveloped land in between. (I say post-2050 because even today, in 2011, a lot of land is still undeveloped in places like the Amazon and the great boreal forests and taiga. Let's hope I'm not too optimistic in thinking it will last another 40 years.) I think Civ5, once they fixed the ICS problem, does a better job of holding off complete land conversion to the industrial age. I always thought it unfortunate that FFH so much paralleled the base Civ4 game in this, with most of the land looking more like LA than some kind of fantasy setting.

I'm skeptical though that encouraging cities to be really spread out would necessarily work well. First, the AI city placement algorithm isn't designed to do that. The AI also won'd do a very good job of militarily defending a scattered territory. Road maintenance costs can get really high with a spread-out territory, and transport time can get long. I think it is find to hard-code 3 minimum tiles between cities, but I wouldn't do more than that.

Placement is fine. The AI does a good job of spreading out cities early (some better than others). Then they fill in gaps very aggressively later. Actually, it's not city placement that makes a medieval countryside different than Los Angeles. It's the ability to use and develop the space between the cities. Once you can produce so much food (and everything else), it becomes inevitable that all land gets developed and cities grow to fill the space. However, if you got the land production correctly balanced for an ancient or medieval setting, you would not need any hard-coded distance minimum at all.

On military: That's all in the balance. If conquest is too easy, increase military unit costs and/or city defense. If too hard, do the opposite. [Edit: and anyway, maybe these outpost cities should have more turnover...]

On road costs: This is also in the balance, and a fairly easy area to make many adjustments (more gold from trade, reduced road costs through policies, etc.). And in any case, I would argue that more sea trade and less road trade would be:

1. more suited to a fantasy setting. Rome is famous for roads, but Rome and certainly other ancient Mediterranean civilizations realied on sea trade to an extent not emulated in the current Civ5 balance (where you always build a road unless it is a very remote or island city).
2. fine for the AI. It's pretty well inclined to building harbors already. I don't really know if it will bankrupt itself building roads (that's hard to tell in base Civ5 where connecting with roads always pays off and is usually cheaper than harbors).

 

[PawelS]

2011-10-17

About "less developed areas" - I agree that cities have too much food in vanilla civ5 (especially when you're allied with several maritime CS), changing it to 3 food per citizen should help (also the number of CS, renamed to Minor Civilizations, will be reduced because they will have multiple cities). But nothing prevents you from building improvements that are unused by the city... My method to ensure the existence of wild lands is to limit the number of cities using the happiness system, unhappiness per number of cities will be far more important than in normal game.

A good goal would be to try to get food, production and gold to actually be equal in value.

That would simplify things a lot, and I think I will treat them as equal, but how to tell if they are really equal or not?

So for example, farms could give +2 food, +3 with fresh water and a midgame tech or with a late-game tech, mines could give +2 production, +3 with fresh water and a midgame tech or with a late-game tech, trading-post equivalents could give +2 gold, +3 with fresh water and a midgame tech or with a late-game tech.

I'm not going to give any bonuses to mines and trading posts from fresh water, but there will be other factors, for example mines will get extra production from nearby mountains. I don't want all improvements to work the same way except giving different yields, I want more uniqueness.

About more sea trade: Harbor (or another building that requires Harbor) will increase gold from sea tiles by 1, this should increase the role of sea trade. But roads are not to be underestimated, landlocked civs shouldn't go bankrupt too easily

 

[Ahriman]

My point though is more about land development density.

I understand what you're saying from a realism perspective, but I'm not sure that having large areas of the game that aren't developed is good for gameplay, for the reasons I noted (AI in particular).

Basically, I would limit the "empty areas" to bad terrain; tundra, desert and jungle. [Jungle could use some nerfs so that it is actually an obstacle development; I'd probably put it back to 1 food, prevent trading posts being built in it, and increase the clearance time or have clearance require a high-end tech in a fantasy mod.] Other areas should be densely settled.
You could then easily have a desert civ, tundra civ, jungle civ that prefer those zones.

[Also, change some mapscripts to give more jungle; most standard scripts now hardly give any jungle.]

I agree that cities have too much food in vanilla civ5 (especially when you're allied with several maritime CS)

This is basically a Maritime CS issue.
I strongly recommend adopting many of Thalassicus's economy balance changes - in particular changing MarCSs so that they give a flat amount of food per era (spread across however many cities you have) rather than flat food per city per era.
I'm sure he'd be happy for you to use his code.

Also, if you're increasing food consumption to 3 per person, then you will need to increase food supply significantly.

but how to tell if they are really equal or not?

Playtest and balance tweaking is the only way. I'm happy to help with this.

I'm not going to give any bonuses to mines and trading posts from fresh water, but there will be other factors, for example mines will get extra production from nearby mountains. I don't want all improvements to work the same way except giving different yields, I want more uniqueness.

Be very careful how you do this. The gold should be to have the choice of which improvement to build be an interesting one. If you give mines bonuses near mountains and farms bonuses near fresh water, then you make the choice of which improvement to build an uninteresting no-brainer; every river-adjacent tile gets a farm, every mountain adjacent tile gets a mine.

This is why it is valuable to have every tile get the bonus from fresh water; that way there is still an interesting decision as to which to choose, because all of them have an opportunity cost; if I build an extra high-yield farm here, I give up the opportunity to have an extra high-yield trading post. The way VEM does it is that the bonuses come from different techs. So the trading post bonus comes from naval tech line, the farm bonus from growth tech line, the mine bonus from metals/military tech line.

Harbor (or another building that requires Harbor) will increase gold from sea tiles by 1

Yes, this and many other good economy changes are in VEM.
I think the idea should be to have coast tiles be better than non-river grassland tiles IF you have harbor and lighthouse. So for naval tiles you have to expend production on workboats and buildings, but you don't have to build workers and you get better yields than non-river tiles.

 

[Pazyryk]

I understand what you're saying from a realism perspective, but I'm not sure that having large areas of the game that aren't developed is good for gameplay, for the reasons I noted (AI in particular).

It's not a realism perspective. It's a flavor and gameplay perspective. The flavor perspective is that a fantasy world should be sparsely occupied, not look like the most productive areas of a modern "developed" country. The second part, gameplay, is very subjective and personal, I admit. My subjective and personal feeling is that Civ5 pre-renascence (when sparsely occupied) is more fun than the period of complete land development that comes after.

On AI, I could come up with arguments either way. I really don't think a strong prediction is possible. In base Civ5 at high difficulty levels (Immortal+), the AI is particularly threatening at the very beginning and then in late game. However, this is an effect of AI handicap bonuses (which can be adjusted), not actual AI performance. You may perhaps be right that the AI does worse with sparse development, and better in the "modern" age with full land development. I honestly can't tell. I certainly wouldn't take this as a given.

Basically, I would limit the "empty areas" to bad terrain; tundra, desert and jungle. [Jungle could use some nerfs so that it is actually an obstacle development; I'd probably put it back to 1 food, prevent trading posts being built in it, and increase the clearance time or have clearance require a high-end tech in a fantasy mod.] Other areas should be densely settled.
You could then easily have a desert civ, tundra civ, jungle civ that prefer those zones.

Agree that jungles should be awful. Actually, I know I'm proposing a rather radical view, but here is what I think should be the case:

grassland -awful
plains -awful
forest -really awful
jungle -really seriously awful
tundra -like jungle
ice -forget it
desert -forget it​

for jungle, desert, tundra+ice and forests, you have a few civilizations that specialize on these terrains. That doesn't make it paradise by any means. Just means that they can survive and grow a little on these terrains, where others can't. Really big cities (10+) are only found in the most favorable terrain (river flat lands) with outpost cities grabbing worthwhile resources elsewhere. The AI will inevitably build other cities, but these don't grow much and the game is carefully balanced so that these are neither a big benefit (encouraging ICS by human) nor a big burden (gimping the AI).

By the way, I'm mostly arguing here for my own sake, to think out some ideas for a different sort of approach for my mod. The back and forth is very helpful to me -- hopefully useful to PowelS also. I think that I can balance city development so that there is no need for hard-coded limits (min city distance) and not even much reliance on happiness limit. Even more important to me, I want to get the flavor right for a fantasy setting and avoid late game boredom, which is going to take a radical departure from everyone's normal expectations for Civ. Of course, this may be impossible.

 

[Ahriman]

The second part, gameplay, is very subjective and personal, I admit

Part of my problem is that in a big open world which is very sparsely occupied, you have a lot of downtime moving units around through open areas. To launch an invasion, you have to march many turns to just get to the enemy, and then many turns back again. It makes warfare kindof whack-a-mole if your army can take my city before my army can get there, and vice versa.
I think warfare is much more fun when it feels like your empire vs their empire, with your army vs their army, and this means that it needs to be reasonably easy for your army to get around your core provinces.

At a more macro level: interesting conflict between players happens when they are pressing up against each other. When they are racing to control the same areas. When they are in each other's way. When they can easily bring force to bear against each other. When you want to approach from particular directions because otherwise you might be getting shot from another neighboring city.

I think it is much better to have an empire of cities that are relatively close to each other (4-5 tiles, say) rather than rare and scattered.

Actually, I know I'm proposing a rather radical view, but here is what I think should be the case:
grassland -awful
plains -awful
forest -really awful
jungle -really seriously awful
tundra -like jungle
ice -forget it
desert -forget it

That sounds really, well, not-fun.
A bit part of why Civ is fun is in getting resources and building stuff.
If every terrain is bad, then that means that most cities will never amount to much, and won't be very valuable.

Also, if you have most terrain as being weak, then you have a situation where you either can't have many decent cities (which I don't think is fun), or where you have to have a huge map, which means it takes a very long time to get anywhere. It means that an early-game rush isn't feasible, because of the distances involved. It means that it will be a lot of time spent shuffling units around. It means that roads will be really long and will be impossible to defend from pillage (and the human will be much better than the AI at surgically cutting off all the enemy trade routes).

The AI is also likely to do a very bad job in such a system. The AI doesn't make smart settlement decisions, it has strong preferences for settling in nearby places. It doesn't make smart military decisions when cities are scattered, it can't concentrated its forces well and respond rapidly. The AI isn't very good at setting war goals, and prioritizing a resource-rich area as a target. Basically; a big scattered empire really makes planning ahead critically important when it comes to warfare - and AIs are nearly always very bad at planning ahead.

The AI performs best when city placement doesn't matter much. If you had a model with extreme resource differences where terrain was incredibly important, then the AI is going to be really dumb. Its going to build lots of weak, wasteful cities in the bad terrain, and cripple itself with unhappiness.

I've never seen any colonization/builder game that worked well with really scattered-far-apart cities. Some that do this, like Elemental, end up being not-fun in part because they try it.

So I think I'd stay away from it. Its fine to have a few zones that are very low value for most civs, but not too many, or there isn't enough conflict and tension.
If you wanted to try a radically different design in how the game feels, then I would suggest designing a mod around that, rather than trying to do do that on top of all the other changes from a fantasy mod.

*edit*
Apologies if this came off a bit too harshly. Its certainly an interesting idea. I just worry that it is too big a change to design a mod around.
You might want to check out the Fury Road mod for Civ4, which is post-apocalyptic, and comes the closest I can think of to what you have in mind. Most terrain is terrible, and there aren't many good city sites. But the only reason I think why that mod manages to get away with it (mostly - it does suffer from some of the issues I raised) is that the world starts with a pre-established network of highways. They're missing a few pieces here and there, but they link the main ruined areas, which is where most of the important resources are clustered and where cities will tend to go.

*edit2*
Maps with lots of empty space will also tend to spawn lots of barbarians, and humans tend to manage to fight barbarians much better than the AI can, and so the empty spaces tend to hurt AI development.

 

[Pazyryk]

@Ahriman, Many good points. Just a couple responses.

Part of my problem is that in a big open world which is very sparsely occupied, you have a lot of downtime moving units around through open areas. To launch an invasion, you have to march many turns to just get to the enemy, and then many turns back again.

I can see your point here. What's definitely not fun is moving a lot of units over a lot of land. However, moving a few units over distance is not particularly more tedious than moving a lot of units over a short distance. I find early warfare with sparse armies and small cities just as fun (and certainly quicker) than late warfare with many units and dense development. I certainly enjoy the latter, from time to time, but it takes a lot of time and is not essential for my game enjoyment.

At a more macro level: interesting conflict between players happens when they are pressing up against each other. When they are racing to control the same areas. When they are in each other's way.

The conflict is still there. However, it is conflict over a resource or grouping of resources, or perhaps a fertile river area. In base game, you do have this but you also have the fill-every-single-tile conflict later, which I don't enjoy at all.

When they can easily bring force to bear against each other. When you want to approach from particular directions because otherwise you might be getting shot from another neighboring city.

I think it is much better to have an empire of cities that are relatively close to each other (4-5 tiles, say) rather than rare and scattered.

Again, I'm not really talking about city placement. Rather, its about the proportion of tiles that can be fully developed and worked. Density of cities would be more or less the same as base (in part because it will be hard to convince the AI to do otherwise; but also because the current spacing seems fine to me and I do want neighboring civs to run up against each other).

That sounds really, well, not-fun.
A bit part of why Civ is fun is in getting resources and building stuff.
If every terrain is bad, then that means that most cities will never amount to much, and won't be very valuable.

I may be engaging in a bit of hyperbole to make my point. But really, what makes a city of 10 pathetic and a city of 30+ good? I used to think my old AD&D level 6 rogue was pretty advanced, but now you have to be level 20 or 40 or whatever to be "powerful". A big city used to be 10,000, now its 10,000,000. It's all relative.

Also, if you have most terrain as being weak, then you have a situation where you either can't have many decent cities

A few great cities (i.e., 10) and then villages and outposts (1-3). There is nothing wrong with these villages and outposts. They serve their function in claiming land or holding a strategic point. They just will never grow (or at least most of them won't grow) into giant core cities. (Note that I am reworking Social Policies so that building these has no plus or minus effect on social advancement.)

(which I don't think is fun), or where you have to have a huge map, which means it takes a very long time to get anywhere. It means that an early-game rush isn't feasible, because of the distances involved. It means that it will be a lot of time spent shuffling units around. It means that roads will be really long and will be impossible to defend from pillage (and the human will be much better than the AI at surgically cutting off all the enemy trade routes).

No. Again, I'm not (for the most part) advocating more distant cities (or if so, only just a little). In fact, if one doesn't lower unit costs considerably, then the changes I propose would lead to much less "time spent shuffling units around" because the number of units will be smaller (more like base Civ5 in early game than late game).

The AI is also likely to do a very bad job in such a system. The AI doesn't make smart settlement decisions, it has strong preferences for settling in nearby places.

I've seen it do both. Sometimes it goes out and very aggressively claims resources at distance. Sometimes settles low value areas close by. I don't know the exact flavor settings that control this, but it seems adjustable. In my system, you would want the AI to aggressively claim good areas, then fill in later (i.e., not really different than optimal AI in base Civ5).

It doesn't make smart military decisions when cities are scattered, it can't concentrated its forces well and respond rapidly. The AI isn't very good at setting war goals, and prioritizing a resource-rich area as a target. Basically; a big scattered empire really makes planning ahead critically important when it comes to warfare - and AIs are nearly always very bad at planning ahead.

This is all true. But I don't see how it affects the scenario I describe. City placement is more or less the same. You just have city sizes (and hence land exploitation) that is much reduced. (And consequently, smaller armies... though this could be altered if one wants large armies.)

The AI performs best when city placement doesn't matter much. If you had a model with extreme resource differences where terrain was incredibly important, then the AI is going to be really dumb. Its going to build lots of weak, wasteful cities in the bad terrain, and cripple itself with unhappiness.

Agree with the first sentence. The thing is, I want the same thing that the AI likes to do. That is, to have some great cities but then a lot of smallish villages scattered about the countryside. The AI does exactly what I want. What I hate is the rule set that makes this behavior crippling. Both the happiness system and (perhaps worse) the policy system. Keep the basic AI placement tendencies. Change the rule set in a way that villages don't kill social progression or growth of large cities (why should they?), nor provide unbalanced benefit (leading to ICS abuse).

I've never seen any colonization/builder game that worked well with really scattered-far-apart cities. Some that do this, like Elemental, end up being not-fun in part because they try it.

Haven't tried that yet. Actually, I'll confess that I haven't played much except Civ and BG/NWN/DA (and mods thereof) for the last 20 years.

So I think I'd stay away from it. Its fine to have a few zones that are very low value for most civs, but not too many, or there isn't enough conflict and tension.
If you wanted to try a radically different design in how the game feels, then I would suggest designing a mod around that, rather than trying to do do that on top of all the other changes from a fantasy mod.

I thought about this. Thing is, I think full land exploitation is perfectly appropriate for Civ5 post industrial age. I don't have any particular complaint about Civ5 general balance. (I do have complaints about tactical AI. Also, I think the designers like to encourage lots of metagaming with RAs, resource selling, and such. Since the AIs can't do this, they then have to add even more absurd handicap bonuses to AI, much more than would be needed to compensate for the (perhaps inevitable) poor strategic and tactical abilities.)

*edit*
Apologies if this came off a bit too harshly. Its certainly an interesting idea. I just worry that it is too big a change to design a mod around.

No hard feelings. You raise many valuable points.

You might want to check out the Fury Road mod for Civ4, which is post-apocalyptic, and comes the closest I can think of to what you have in mind. Most terrain is terrible, and there aren't many good city sites. But the only reason I think why that mod manages to get away with it (mostly - it does suffer from some of the issues I raised) is that the world starts with a pre-established network of highways. They're missing a few pieces here and there, but they link the main ruined areas, which is where most of the important resources are clustered and where cities will tend to go.

I wish I had tried this. I spent a lot of time with FFH and the Orbis and Wildmana modmods (and made a little modmodmod for Wildmana). My one complaint about FFH was that Kael never succeeded in implementing "wildlands." Perhaps I should take that as an indication of how difficult it might be.

Edit: I really hope PawelS is not offended by the thread hijacking...

 

[Ahriman]

What's definitely not fun is moving a lot of units over a lot of land. However, moving a few units over distance is not particularly more tedious than moving a lot of units over a short distance.

I'd say: its not just about tedium, its also about the strategic consequences. If it takes 15 turns for my invasion force to arrive at your borders, that is a huge time commitment. It means I won't be able to reinforce the army, it means that if I am attacked in the meantime I will be super-vulnerable, it means that you can build many new units (and possibly even get new techs and upgrade) between when I send my force off and when I arrive.
Also, just creating lots of bad zones won't in itself somehow make armies smaller.
I also think that at least moderate size armies are important to have, because it makes strategic management of congestion effects more important.

However, it is conflict over a resource or grouping of resources, or perhaps a fertile river area.

But the AI isn't really programmed to context a river valley or cluster of resources. It is programmed to contest land, and to care about people who are near them, and to particularly care about shared borders.
You'd have to rewrite a lot of the AI - settlement, warfare, diplomacy - to get satisfactory performance in a very different map layout, and I don't think we have sufficient code-access to do that.

Again, I'm not really talking about city placement. Rather, its about the proportion of tiles that can be fully developed and worked. Density of cities would be more or less the same as base (in part because it will be hard to convince the AI to do otherwise; but also because the current spacing seems fine to me and I do want neighboring civs to run up against each other)

I don't think I understand this. If good cities only exist on rivers or in resource clusters, and the rest of land is weak or useless, then how does this not reduce the average density of cities?

But really, what makes a city of 10 pathetic and a city of 30+ good?

I understand what you're saying, but absolute and relative levels matter.
For example, on absolute levels; in one design world A, cities are still worth having nearly everywhere (eg are worth their extra unhappiness and policy cost increase) but are extra-good in particular locations, that are worth prioritizing. In another design world B, cities aren't worth building except in particular special locations, and so most land would remain wilderness if civs were playing optimally. It sounded to me like B was the kind of design you were supporting, but I don't think this would work well. And if you have A, then you don't have the wilderness effect that you were talking about; A is basically just what we have now but with more resource clustering.
Relative levels matter too; the AI is more likely to just plonk down cities wherever, so the larger is the difference in value between a carefully selected city site and a randomly placed city site, the further the AI will fall behind. That isn't a sufficient reason to not have variation, but it does need to be kept in mind.

In fact, if one doesn't lower unit costs considerably, then the changes I propose would lead to much less "time spent shuffling units around" because the number of units will be smaller

Why? Because you're saying it would take forever to build a unit? I don't think that would be fun. I think build times are already too high in vanilla, this is one reason I prefer VEM.

You just have city sizes (and hence land exploitation) that is much reduced.

I don't see this as a desirable goal. I think it is too hard in vanilla to get big cities. This is another reason I prefer VEM.

That is, to have some great cities but then a lot of smallish villages scattered about the countryside. The AI does exactly what I want. What I hate is the rule set that makes this behavior crippling. Both the happiness system and (perhaps worse) the policy system.

It sounds like what you want is ICS.
I don't really like that.
I think it is great that there are mechanics that make you think really hard about whether to build a city or not. I don't think it is a good idea to make cities always worth building, even if they are weak and will never amount to anything much.
I also don't think "scattered cities everywhere but only a few that are any good" really sounds like the Fantasy Wildlands concept you had.

Haven't tried that yet.

Don't bother. Elemental stinks. I haven't tried it lately, I know they've made many changes, but it was just not at all fun.

I wish I had tried this.

Its not too late, Civ4 still works

My one complaint about FFH was that Kael never succeeded in implementing "wildlands." Perhaps I should take that as an indication of how difficult it might be.

Yes, I think we probably should. Its really hard to do. Fall Further tried it to some extent, and IMO ended up failing, because the other Civs couldn't handle it (the barbarians in particular) almost always crippled.
Creating Wildlands IMO always sounds better in theory than it actually works in practice, because there is just a fundamental problem with creating terrain that isn't really worth contesting.
Basically, I think it is better to think of the farms and trading posts as the rural areas/small towns/villages etc, and cities in Civ only represent the big cities that have decent yields. I think that is even easier to do in Civ5 than in previous civ, because 3-tile range city coverage can cover a lot of tiles.

*edit*
Just to clarify, I think that my preferred design would be that forest/plains/grassland are all reasonable, rivers and lakes are good extra bonuses, some areas have natural wonders or magical nodes or unique features or extra resources that make them more desirable, because cities there will have slightly higher yields and so will be more efficient. Some places will be better than others, but it won't be the case that a city in one zone can be size 25 while a city in another can only be size 9.
So, pretty close to how vanilla works, except with more jungle and jungle as an undesirable area.

 

[Pazyryk]

@Ahriman

Many of these arguments fall under the sort: "if I change A and not B, then C will happen." It's pretty much impossible for me to list all of the B's that need to change to prevent C if I do A. I'm juggling 5 different major changes but only talking about one or two. It's a sort of endless discussion deflecting "AI this", "ICS that". The detailed balancing is, well, very detailed. But to be frank, the math that drives ICS is pretty easy to understand. I don't know why it took the devs so long to grapple with this.

That's why I've tried more to communicate the desired result, rather than then specific balancing adjustments.The interesting point of discussion is not How will you defeat ICS?. Notwithstanding the fact that the devs struggled with this for so long (or weren't aware of the problem?), it is a trivial technical problem that isn't hard to solve in many different ways. The interesting question is: What kind of world do you want?. Now, of course, the world that one player wants is not necessarily pleasing to another player. Many of your points are along these lines and they are good points. Others are just technical problems that may be easier or harder to deal with case by case.

One reason I think modders should be bold in their changes is that, to a large extent, the AI doesn't cope well with the current system and often acts completely contrary to current game balance. For example, they have an AI that aggressively fills every spot on the map. Then they build a happiness and social policy system that punishes this behavior. Then they have to give absurd happiness bonus to AI to compensate. Perhaps I'm arrogant to think I can do better. But, indeed, I beleive that I can use the current AI and change the rules in a way where the AI does better and I get the desired effect.

I also don't think "scattered cities everywhere but only a few that are any good" really sounds like the Fantasy Wildlands concept you had.

Really just going for something like ancient or perhaps medieval level of development. Much undeveloped land (with villages here and there -- where a village is a smallish city). Some wide expanses of wilderness (as we even have today in the real world). I don't think modern LA is what I want in a fantasy.

Yes, I think we probably should. Its really hard to do. Fall Further tried it to some extent, and IMO ended up failing, because the other Civs couldn't handle it (the barbarians in particular) almost always crippled.
Creating Wildlands IMO always sounds better in theory than it actually works in practice, because there is just a fundamental problem with creating terrain that isn't really worth contesting.

Can't comment on FF too much. It's pretty clear to me where FFH, Wildmana and Orbis start getting into a "modern" feel for me (when I get bored and quit). *Edit: I also can say that none of these (and probably not FF I'm guessing) actually thought about the problem in the same way that I am here. The problem I see is that all of these give civs a very "modern" amount of food production. The result is a modern (or post-modern) looking world, no matter what other mechanisms you add to try to counteract this.

 

[PawelS]

2011-10-18

I really hope PawelS is not offended by the thread hijacking...

I'm not offended at all, feedback is always welcome, although this time the posts are really long and I think I'll spend more time reading them and answering to them than actually working on the mod tonight...

[Also, change some mapscripts to give more jungle; most standard scripts now hardly give any jungle.]

See the thread I linked earlier, the cause of the problem is allowing jungles only of Plains, while most of the tiles near the equator are Grass. This will be fixed in this mod, so even the standard mapscripts will give much more jungle. Clearing jungles will require a later tech indeed.

I strongly recommend adopting many of Thalassicus's economy balance changes - in particular changing MarCSs so that they give a flat amount of food per era (spread across however many cities you have) rather than flat food per city per era.

I think it won't be a big problem when the number of cities you can build is severely limited by happiness, so I think I'll just keep it as simple as possible: 1 food in all cities from friends, 2 from allies (no extra food in capital).

The limited number of cities will also, as I noted earlier, increase the amount of unsettled areas. Barbarians will play a more significant role in the mod, and they need some places to appear in middle and maybe even late game (of course it depends on the map size and number of civs). Even if the AI is not good at managing Barbarian invasions, they are too important thing in this mod not to have them.

Be very careful how you do this. The gold should be to have the choice of which improvement to build be an interesting one. If you give mines bonuses near mountains and farms bonuses near fresh water, then you make the choice of which improvement to build an uninteresting no-brainer; every river-adjacent tile gets a farm, every mountain adjacent tile gets a mine.

Well, I think the truly important choice should be where to place your cities. I don't mind some no-brainers when placing improvements.

About the sea tiles: I want the coast and ocean tiles to be different, coast will probably give 2 food and 1 gold and ocean 1 food and 1 gold.

Part of my problem is that in a big open world which is very sparsely occupied, you have a lot of downtime moving units around through open areas. To launch an invasion, you have to march many turns to just get to the enemy, and then many turns back again. It makes warfare kindof whack-a-mole if your army can take my city before my army can get there, and vice versa.
I think warfare is much more fun when it feels like your empire vs their empire, with your army vs their army, and this means that it needs to be reasonably easy for your army to get around your core provinces.

That's why to make defending your territory easier you should found your cities close to each other, even if there are more interesting lands farther. This doesn't contradict with large areas being unsettled. This also means that in most cases you will be able to found a new city when happiness allows instead of going to war. I don't think it's bad, you should have a choice to play peacefully or attack the enemy to grab the best city sites. But the AI can attack you even if it has lots of land to settle, I noticed it when playing my previous mod, which modified the vanilla game by restricting the number of cities through happiness like in this mod (of course you can also play a crowded game if you like by increasing the number of civs and decreasing the map size, but it won't be the default setting).

About the terrain yields: terrain without resources won't give good yields, but they will be much better from resource tiles (including mana nodes), natural wonders, and from Villages that you can build using a type of Great People. Also, terrains like Tundra, Snow, Desert, Jungle and Marsh will be really bad for most civs, because they won't be able to build any improvements there.

The AI performs best when city placement doesn't matter much.

This is a bad thing, because it strongly contradicts with my plans for this mod, I guess I'll just have to adjust the AI bonuses so it remains competitive.

 

[PawelS]

2011-10-19: More about happiness and number of cities

I think in the normal game the number of cities is way too insignificant factor, and population too significant factor when calculating unhappiness. In my non-fantasy mod, I set it to 10 per city and 0.5 per population, and it worked well (there were also some other changes to the happiness system, like getting +1 happiness from every policy, less from buildings, and a higher base value). In this mod I think it will be the same, and some civs will have these factors adjusted: Halflings to 12 per city and 0.3 per population (so they will get benefits above size 10) and Orcs to 8 per city and 0.6 per population (so they will get benefits below size 20).

There won't be any modifiers to unhappiness from the world size - bigger maps should have more civs, not bigger civs.

 

[Pazyryk]

With the system you describe, you might want to think about some strong intervention to limit AI settler building. This should be doable via Lua, and you can make it a function of happiness or whatever else you want. If you use AI bonuses instead, then you will not change the behavior or achieve your desired effect on city number. (I wish the developers understood this.)

Hmmm... what you describe would seem to limit cities to one type: big. The cost of building an outpost city in a non-food optimal place (for example to hold a military position or claim some resources) is prohibitive. Is this what you want?

 

[PawelS]

With the system you describe, you might want to think about some strong intervention to limit AI settler building. This should be doable via Lua, and you can make it a function of happiness or whatever else you want. If you use AI bonuses instead, then you will not change the behavior or achieve your desired effect on city number. (I wish the developers understood this.)

I made Settlers unbuildable, and created a LUA script that gives them for free when you have enough happiness. My tests with the non-fantasy mod showed that with buildable Settlers the AI sometimes didn't build enough or them, and sometimes built too many and had high unhappiness. In case of minor civilizations (formerly city states), they don't use Settlers at all even when they get them for free, but I managed to give them free cities using LUA, it checks player:AI_foundValue (x,y) on all map tiles and places a city where the found value is the highest using player:Found (x,y).

Hmmm... what you describe would seem to limit cities to one type: big. The cost of building an outpost city in a non-food optimal place (for example to hold a military position or claim some resources) is prohibitive. Is this what you want?

When you get a Settler you should use it to found the city in the best possible place, which usually means one with a potential to create a big city indeed, but things like claiming strategic resources can be important too, and you still get some unhappiness from population, so it's not always the size that matters the most

But your idea about "outposts" (small cities that are there only to claim resources or a defensive position) sounds interesting, I just don't know how to implement a system that prohibits you from having too many cities and at the same time allows a number of small ones that cause less unhappiness... Let me know if you have any ideas how to do it.

 

[Pazyryk]

But your idea about "outposts" (small cities that are there only to claim resources or a defensive position) sounds interesting, I just don't know how to implement a system that prohibits you from having too many cities and at the same time allows a number of small ones that cause less unhappiness... Let me know if you have any ideas how to do it.

Not an easy solution, no. My system is so different that it doesn't really even need a happiness limiter for growth (it's still there, just won't be an issue for most civs). In fact, I've even done away with the growth multiplier and exponent. I'm also re-working social policy so that it doesn't punish you for # of cities. I never like these mechanisms and you just don't need them IF food production is changed from always-in-excess (as it is in base and all balance mods, I think) to limiting. You naturally get some mega-cities (they grow with tech level because this affects improvement output, and they grow fast) and other cities grow or don't grow depending on the region. No punishment at all for the outpost cities. On the other hand, only limited benefit from these so there is no real motivation for ICSing.

Edit: I've been emphasizing terrible terrain yield in describing my system. Perhaps a better (or better salesmanship) way to describe it is to start with the idea of dropping the food multiplier and exponent (0 and 1, though I think I should be able to drop the latter to 0). Obviously, this won't work without major yield changes. But, in essence, I've given a monumentally huge benefit to large cities (or I like to think of it as removing a huge penalty). 1 pop growth cost no more in food for a city of 20 than 1. With this monumental boost to large cities, I can now drop a lot of punishment for small cities without getting into ICS territory. It's a tough balancing act, of course, and I can't say I've got it right yet.

 

[Ahriman]

in my non-fantasy mod, I set it to 10 per city and 0.5 per population

Oh, wow. That is a pretty huge difference. I really think that radically changing this kind of thing is just too much. So cities all have more unhappiness than vanilla until size 16??

It forces you into having only a tiny handful of cities. Happiness is no longer about population, but only about number of cities. It severely limits empire size no matter the tech level, because you can't possibly have enough fixed happiness sources to cover the extra cost of having many cities.

It would make conquest pretty much impossible, and would force you to raze nearly every city you capture. The AI doesn't raze, and so a conquering AI would also be totally screwed. It also means that AI happiness bonuses or not would have a *huge* impact on the game, rather than a modest one.

So there are no more tradeoffs - you always want to make a given city as big as possible, because small cities are totally useless. There is no longer a tradeoff between tall and wide empires, you can only go tall. Making settlers unbuildable and generated by script also sounds really not fun, you're taking an important player decision out of the hands of the player and making it arbitrary.

I think this kind of model would lead to very, very frustrating gameplay.

Honestly, I don't think I'd want to play a game like that.

I think it is fine to tinker with the vanilla mechanics - VEM for example has 4 unhappiness per city and +1 per pop - but radical changes like this seem unnecessary.
What is it about a fantasy mod that needs this kind of radical overhaul? I would start with something simpler and likely to have broader appeal.

 

[Pazyryk]

I think it is fine to tinker with the vanilla mechanics - VEM for example has 4 unhappiness per city and +1 per pop - but radical changes like this seem unnecessary.
What is it about a fantasy mod that needs this kind of radical overhaul? I would start with something simpler and likely to have broader appeal.

Need? None at all. Because there is no money to be made here, the only motivation is very personal. For some that might be broad appeal. For others it might be the enjoyment of experimentation. For others it may be producing something truly new and different. A modder might have more or less interest in pursuing one or all three of these areas.

 

[PawelS]

2011-10-20

@Pazyryk: If I understand it correctly, there are no punishments for having too many cities in your mod at all (other than using your production and growth potential to build Settlers), so you can settle/conquer the entire world in the early game if you like. It's nothing wrong, and there are many games where there are no such limits (MoM for example), but it's not my way of doing things

Oh, wow. That is a pretty huge difference. I really think that radically changing this kind of thing is just too much. So cities all have more unhappiness than vanilla until size 16??

It forces you into having only a tiny handful of cities. Happiness is no longer about population, but only about number of cities. It severely limits empire size no matter the tech level, because you can't possibly have enough fixed happiness sources to cover the extra cost of having many cities.

As I wrote earlier, there are other changes to the happiness system as well, like getting extra happiness from policies. My intention is that you should have only a few cities at the beginning, but you can get more as the game progresses. In the late game you can have about 15 cities, which is enough for my tastes.

It would make conquest pretty much impossible, and would force you to raze nearly every city you capture. The AI doesn't raze, and so a conquering AI would also be totally screwed. It also means that AI happiness bonuses or not would have a *huge* impact on the game, rather than a modest one.

I've seen the AI razing cities quite often. Conquest is possible if you accumulate enough happiness and don't want to use it to found a new city, and there is no extra unhappiness for captured cities, they work just like the ones you found (also there are no puppets). And yes, the AI happiness bonuses are reduced.

So there are no more tradeoffs - you always want to make a given city as big as possible, because small cities are totally useless. There is no longer a tradeoff between tall and wide empires, you can only go tall. Making settlers unbuildable and generated by script also sounds really not fun, you're taking an important player decision out of the hands of the player and making it arbitrary.

One of the things I don't like in the normal game is that often you're afraid to let your cities grow, because of the high unhappiness per population, that's why I reduced it by half. So it's true that generally you should make your cities as big as possible, but it's not the only factor when deciding where to settle, things like strategic resources are important too. Also many buildings with specialist slots will be buildable only near specific resources (or have other terrain-based requirements), so a city with lots of food but no interesting resources will have nothing to do with all that extra population.

It seems my approach to strategy gaming is different than yours, for me more decisions isn't always better. I prefer to leave only some important strategic decisions to the player - where to found a city, which tech to research etc., while some other things can be arbitrary. About the "wide empires" (ICS in other words), I eliminated this possibility because I want some parts of the world to remain unsettled.

Getting Settlers for free is mainly to help the AI, but also I don't like when in the early game you have to allocate most (or all) of the production and food to build Settlers, and stop any other development of your civ completely. So you research techs, but don't use them at all because of building Settlers. Of course it's the player's decision, but usually the best decision is to build Settlers, which makes the early game less interesting (at least it's like this in the normal game, with my happiness changes you can't build many cities in the early game, but still free Settlers are better for the AI).

I think this kind of model would lead to very, very frustrating gameplay.

Honestly, I don't think I'd want to play a game like that.

I think it is fine to tinker with the vanilla mechanics - VEM for example has 4 unhappiness per city and +1 per pop - but radical changes like this seem unnecessary.
What is it about a fantasy mod that needs this kind of radical overhaul? I would start with something simpler and likely to have broader appeal.

Need? None at all. Because there is no money to be made here, the only motivation is very personal. For some that might be broad appeal. For others it might be the enjoyment of experimentation. For others it may be producing something truly new and different. A modder might have more or less interest in pursuing one or all three of these areas.

Pazyryk is right, I'm not making a mod with the intention to appeal to a broad audience, I'm making it according to my personal preferences. Also there is always a possibility that someone will create a "modmod" that adjusts the game to the expectations of those who don't like my approach to things like happiness.

 

[Ahriman]

Pazyryk is right, I'm not making a mod with the intention to appeal to a broad audience, I'm making it according to my personal preferences. Also there is always a possibility that someone will create a "modmod" that adjusts the game to the expectations of those who don't like my approach to things like happiness.

Fair enough. I should stop whining. You have a lot of cool ideas here, and I should wait and try it out once you have something working and see how it plays in practice.
Best of luck!

 

[PawelS]

2011-10-21: Civ#13 - Gnomes

The Gnomes are great constructors and inventors. They have an unique improvement called Workshop, which can be built on flatlands, and gives extra production and science. Their unique building is Inventor's Workshop, which replaces Carpenter and gives extra science (I'm also thinking about a building that replaces Jeweler*, giving more gold and culture). They aren't strong in combat, but their siege machines are superior to those of other races, and they can construct many advanced war contraptions (I think about giving them Submarine and Helicopter). They have access to the industrial branch of the late game tech tree.

* Jeweler is a building that can be built when the city is near Silver, Gold, Gems or Pearls resources. It will probably give +1 happiness, and some gold and culture. I'm not sure about the name - 'jeweler' means a person that creates jewellery, but how should I call the place where that person works? (maybe "jeweler's workshop", but I already have enough workshops )

 

[PawelS]

2011-10-22: Early gold building

I needed a building that gives gold before Market, and the only idea I had was Bead Maker. What do you think?