Village "mayhem"

[Rusty414]

Hey Fanatics

Ive been playing a couple of games through 3.5 - 3.10 and I enjoy most of the modifications. I have one issue though in most of my games. The village improvement has been altered and improved a lot and it seems to me that the AI don't build anything else than villages. This means that in most situations I will end up having cities with 15-20 citizens where the AI is still struggling around 10. Is this intended somehow? and if it is then what is the reasoning behind it?

I have usually been playing on the emperor difficulty in the normal BNW, but when I play with the mod it doesn't seem to be hard enough. Is there a specific difficulty where the AI behaves "best" with the mod?

Regards
Rusty

Edit: Both of my issues seems to be because of my starting conditions. I usually start at classical era and I usually gives around 1000 gold to each player/AI. When i do this the village improvement is available from the beginning and the AI only build this then. Furthermore the 1000 gold seems to give me a big head start because the AI either don't use them or spend on something not so useful. If i start on ancient era with no gold the game is a lot more difficult on emperor and I don't get as much "village mayhem"

 

[Ahriman]

I definitely agree that there are problems with improvement yields, particularly with the village base yield and the fact that it gets bonus yields with fresh water so early in the game.

On balance and difficulty, advanced start can drastically change the balance. If you start the game from the ancient era I suspect that you will not find it easier than vanilla.

 

[tlaurila]

I've only ever played with ancient start, but "village mayhem" doesn't seem that big an issue. Even if an AI city is founded in the later years, the AI seems to build farms ok.

City States do village mayhem, however, and with my own workers the advisor suggestion is always to build a village. So that does make it surprising that the AI builds as smartly as it does. With CS I'm guessing it has to do with them being on gold focus in governor. I do think the governor of the city effects automated workers.

 

[tlaurila]

Have to take back my earlier comment now that my game advances. AI village galore seems like the single biggest issue with the game now, close to game breaking. The AI has tons of gold but too little of everything else. They're breaking their own cities by not giving them food. Something needs to be done.

 

[Ahriman]

The most important thing is unifying the base yields, and the yield increases from techs, to cover all improvements and to have their tech increases at roughly the same eras. The other important thing would be to make sure that there are roughly the same potential for yield increases across social policies - eg we could put the food farm bonus back in Freedom.

 

[tlaurila]

Assuming the yields are balanced well now, then one is worth less overall than one . Then it's ok that Villages give more units of production than farms. That is perfectly ok for game design and balancing. But it's a huge problem if the AI doesn't understand it.

Ideally the AI would realize that food is more valuable when scarce, and less valuable if there is already plenty. I.e. the values aren't simply linear. Now apparently the AI just compares 3 gold to 2 food (or 2 gold to 1 food) and always goes for the gold because it's more, and as result starves in the long run. Even more utopistically the AI would couple improvement decisions to other circumstance, such as favoring food when there is plenty of happiness.

To patch this over quick, one option could be to weigh village to 0 for AI, so that no villages are built. This wouldn't be too bad, as I see it, if the AI gets in trouble for too little gold, then the AI could also get (more) reduced maintenance costs to compensate.

Another option would be to make villages niche use in the early game for the human, so that the AI thinks they're useless. But have villages grow to be a potential late game economy, given aprropriate policy picks. But again so that the AI understands this as well, assuming the AI took those policies (ideally AI should understand to pick village policies together or not at all). Even then, to not gimp the AI at any point, at no point should no food production at all be favored by the AI. So village needs some food. So something like:
Village base 1
Caravel hulls: on river +1 (!)
Economics without river +1 (!)
Wealth policy +1
Rationalism policy: +1

That should be a safeguard against Artificial Stupidity in all cases, while reserving a use case to base your economy on the village.

 

[mitsho]

Is it possible to make villages not buildable next to other villages? That takes away choice but let's us better fine tune it towards social policies and yields in general. Not completely convinced, but why not try it for a version or two

But I do think it hasnt' been managed in any version of civ5 over all its history to create a AI that varies in what improvements it builds. (At least I can't remember one...)

I would not like the 'short fix' of giving (more) bonuses to the AI (reduced maintenance et cetera). I think we should keep them as close to the human as possible.

Your other proposal might work for the AI (surge mid-game?), but is such a mix preferable for the human player? Not sure, but definitely an option

 

[Ahriman]

Assuming the yields are balanced well now, then one is worth less overall than one . Then it's ok that Villages give more units of production than farms.

But it's not ok to have villages giving 2-3 gold while farms give 1 gold. Just because they're maybe not 1:1 doesn't mean that it's ok to be 2:1 or 3:1.

I think 1 food is probably on average slightly more valuable than gold, but not by much. They're also different: gold income flows into trade route income, and there are many more and larger gold multipliers than there are food or production multipliers. So what we get is something where villages want to be concentrated in gold production cities; even if villages give 1 gold, rising to 2 with tech and 3 with policies and farms give 1 food, rising to 2 food with tech and 3 with policies, then I will still want to build villages around my gold production cities.
I think it's much better to have this approach than to have villages just be better from the start.

Is it possible to make villages not buildable next to other villages?

I don't think this would be desirable: I think it is good to have gold cities with lots of villages. I think it would be frustrating if building one village then blocked 6 other tiles.

 

[mystikx21]

Right, I think the easier fix is to reduce villages' early game yields. They will look more in line with farms and mines at that point.

I'd also rather they stayed mostly to gold production. Policy picks of culture or science on villages are more specific effects than what you get through a tech, and the reason is that they can be powerful effects (or useless ones that you skip). Adding food to villages would simply make farms less useful (for us) rather than increase the value of villages appropriately. This is not the direction of balance we need to go in to fix village spamming in other words is to make the village more useful.

 

[ExpiredReign]

It almost sounds like you want to revert them to vanilla Trading Posts.

 

[Ahriman]

Yes - except that there should be freshwater bonuses for improvements other than farms, so that you don't have no-brainer decisions of farms on freshwater and other improvements elsewhere. You should build the improvements that you need for how you want the local city to function, so we can have gold cities and industrial unit production factories and science cities and so forth.

 

[ExpiredReign]

So less generalization and more specificity.

I can't see why this change would be problematic, the rest of the game is structured this way. While you may get a small assistance to another yield, other than the one intended, from building a specific building, we certainly don't want city buildings giving boosts to all the yields at the same time, why would we want that for city improvements outside the city itself?

 

[Ahriman]

I would say rather more ability and need to customize. You want food, build a farm, you want gold, build a village, and then build your markets and banks near the village cities and your university and public school and garden near the farm city.

 

[tlaurila]

City specialization is what effectively playing the game seems to be about. And it's potentially a good and fun concept to balance the game around. However, no matter how fun and games anything might seem to the human player, creating a game that the AI cannot play makes the whole thing worthless. Balance versus the AI is far more important than balance of a science city versus a gold city.

Regardless of any and all other virtues of a game design choice, if it causes the AI to break its game, it destroys the fun of the game to the player. Game without opponents is pointless. This, to exaggarate the point a bit but not much, is what current Villages do.

So either the Villages need to be idiotproofed against the AI (Villages, when producing top unit yields, also provide enough food for a city to subsist even if entirely surrounded by Villages. Or Villages rarely produce top unit yields), or the AI needs to be changed/improved.

If the weights the AI values food/gold in improvements cannot be changed, then maybe the value of a unit of gold could be increased so that the AI values it right, and same amount of food vs gold is good? Particularly the N% (generally N=200 for buildings, N=300 for units) of hammers versus gold. Then again, this would of course increase the power of other gold production, such as trade routes.

 

[Ahriman]

I agree that AI performance is key. But the vanilla game AI does ok as far as improvements go, and reverting to 1 food farms/1 gold villages/1 hammer mines is returning to vanilla.

 

[Thalassicus]

I set the AI weight for gold to 200% a few months ago because the AI was building too many farms. It sounds like that's no longer a problem, so I'll reduce that weight back to the normal 100% value.


====================

The history of improvement changes was:

Vanilla (starting point)
1 farms + 1 with civil service (wet), fertilizer (dry)
2 villages + 1 with economics
2 mines + 1 with steam power

Mod changes
1. split village and mine tech bonuses into wet and dry
2. delayed 1 from mines to iron working, because hills were too important
3. moved the mine change back further to industrialization, because hills were still too important

Undoing these changes would would make hills too important again. If there's too much money in the early game, it would be easier to balance that by increasing early gold costs. If the concern is mines were reduced instead of hills directly, we could try an alternative by moving production from hills to mines. I'd undo changes #2 and #3, and reduce hills to 1.

 

[Ahriman]

The history of improvement changes was:

Vanilla (starting point)

Don't villages and mines only give 1 gold and production in vanilla (ignoring any tech or other boosts)? Or am I totally mis-remembering?

 

[ExpiredReign]

This is yield table as of vanilla BNW:

Improvement|Yield|Tech change|Yield|Total
Villages (Trading Posts)|1|Economics|+1|2
Mines|1|Chemistry|+1|2
Farms (Fresh water)|1|Civil Service|+1|2
Farms (no water)|1|Fertilizer|+1|2

 

[Ahriman]

Thanks ExpiredReign, that's what I thought.

I propose using the vanilla values, except that adding freshwater bonuses for villages and mines at ~medieval era, and adding freshwater and tech bonuses for plantations, pastures, quarries, and camps.

Undoing these changes would would make hills too important again.

I disagree. Why would it make hills too important? Do you think hills are too important in the vanilla game? I never experienced this. I think it is good that places with hills are better suited to production and places without hills are less suited, and I think it is good that there is meaningful variation in terrain: grassland areas are good for food, hill areas are good for production.

Gold is flexible (it can instantly produce buildings or units, can buy city state alliances, can buy tiles, etc), it has larger multipliers from buildings and other effects (eg market +25% vs smithy +10%), and gold purchase multipliers are lower than they are in vanilla. So production isn't "better" than gold (or food), and so hills aren't better than others.

 

[Thalassicus]

I'm referring to the original vanilla, not G&K or BNW. I've only made a few adjustments to improvements since vanilla.

I did not adopt Firaxis' changes because small income numbers are hard to balance, since we cannot use fractions. We can't set income to 1.2 if we feel 1 is too low. It's easier to balance costs instead of income. Costs have larger numbers, so we can do things like multiply 100 cost by 1.2. It appears you feel we build things too quickly with production and gold, so I will raise construction costs 20%.