How to combat AI's ICS? And ... what's up with Great Engineers?

[Chibiabos]

I keep getting dominated in everything because the AI spams cities like mad. I try to focus on science and wonder building, putting my cities in ideal spots, but the AI just crams its in and by the medieval age is way ahead of me in everything. I can't complete a single wonder anymore, even when I try to rush with a Grate Engineer ... which is another consternation ...

Great Engineers seem utterly worthless by this point in the game ... since regular workers can build Manufactories, all GEs can do is initiate wonder or hurry production, and by medieval-era Wonders, that's not even a quarter of the production. Am I missing something with Great Engineers?

Yet the AI seemed to instantly build a wonder from the Industrial era ... the very turn it entered the Industrial era. How the heck did it do that!?

 

[markusbeutel]

I keep getting dominated in everything because the AI spams cities like mad. I try to focus on science and wonder building, putting my cities in ideal spots, but the AI just crams its in and by the medieval age is way ahead of me in everything. I can't complete a single wonder anymore, even when I try to rush with a Grate Engineer ... which is another consternation ...

Great Engineers seem utterly worthless by this point in the game ... since regular workers can build Manufactories, all GEs can do is initiate wonder or hurry production, and by medieval-era Wonders, that's not even a quarter of the production. Am I missing something with Great Engineers?

Yet the AI seemed to instantly build a wonder from the Industrial era ... the very turn it entered the Industrial era. How the heck did it do that!?

First of all - what difficulty are you playing on? Also, the only way to really counter the AI is to more or less match them in the amount of cities they build. With chopping and policies, you should be able to get out a city every 15-20 turns. Worker manufacturies aren't straight up great people manufactories either, they've been scaled according to the mod. And GE, even if they don't hurry later wonders, still give you a solid boost towards them, which is still stronger than either a mod or vanilla manufactory.

It can be difficult, and I'd suggest lower difficulties if you're struggling, as the AI expands slower, both in how many cities they build, and in how fast they grow.

 

[Chibiabos]

Prince, which is the same difficulty I play Thal's mod and vanilla.

So the only way to beat the AI in this mod is to practice ICS like it does?

 

[markusbeutel]

Prince, which is the same difficulty I play Thal's mod and vanilla.

So the only way to beat the AI in this mod is to practice ICS like it does?

Thal's mod is more or less tweaked vanilla, while nights is a completely different animal, and similar styles of play won't necessarily work. The AI is terrible at expanding in vanilla CIV V, and that's not the case in nights. While you'll be able to compete if you have maybe at least 60% the amount of cities an AI has, 3 cities probably won't cut it (and to stay competitive while going tall, you'll really need to play the CS and RA game).

Nights is modeled more on managing a grand empire, not on competing scientifically with 3 cities. Bigger is essentially better, and even in vanilla this is the case, its just that the AI is too pathetic to take advantage of its bonuses. Rather than having half the world unsettled in 500 AD, I took the mod in the opposite direction - which essentially is what the standard was in Civ up until CIV V, where now population = science, which means that ICS will always be the optimal strategy.

 

[Huck Finn]

Is it possible to cut out (or halve) science= population from Civ 5, and just leave science up to those generated by buildings and tile improvements?

Obviously you'd have to make the base rates of the buildings increase science output quite a bit, and include a few more science type buildings, but if its possible it might be a way to move away from the ICS strategy being the only viable one.

That said, even in the other civ games if you were playing with only a handful of cities it took quite a bit of skill to pull off, its just even harder in Civ 5. At least in this one you will be getting tons of social policies with so few cities...

 

[joltis]

Thal's mod is more or less tweaked vanilla, while nights is a completely different animal, and similar styles of play won't necessarily work. The AI is terrible at expanding in vanilla CIV V, and that's not the case in nights. While you'll be able to compete if you have maybe at least 60% the amount of cities an AI has, 3 cities probably won't cut it (and to stay competitive while going tall, you'll really need to play the CS and RA game).

Sorry to pop in the discussion but how did you mod AI expansion. I try to look files but didn't find them. I have been doing same thing (see my AI mod). My approach can be summarized next lines:

At defines:
<!--early REX-->
<Update>
<Where Name="AI_STRATEGY_EARLY_EXPANSION_NUM_CITIES_LIMIT"/>
<Set Value="5"/>
</Update>
<!--early REX-->
<Update>
<Where Name="AI_STRATEGY_EARLY_EXPANSION_NUM_UNOWNED_TILES_REQUIRED"/>
<Set Value="7"/>
</Update>
<Update>
<Where Name="FLAVOR_EXPANDGROW_COEFFICIENT"/>
<Set Value="4"/>
</Update>
<Update>
<Where Name="AI_CITY_SPECIALIZATION_EARLIEST_TURN"/>
<Set Value="6"/>
</Update>

At AIEconomicStrategies:

<Update>
<Where Type="ECONOMICAISTRATEGY_EARLY_EXPANSION"/>
<Set CheckTriggerTurnCount="1"/>
</Update>
<Update>
<Where Type="ECONOMICAISTRATEGY_EARLY_EXPANSION"/>
<Set WeightThreshold="-5"/> Side note not sure about this
</Update>

At AIEconomicStrategy_Flavors:
<Update>
<Where AIEconomicStrategyType="ECONOMICAISTRATEGY_EARLY_EXPANSION" FlavorType="FLAVOR_EXPANSION"/>
<Set Flavor="30"/>
</Update>

At AICityStrategy_Flavors:
<Update>
<Where AICityStrategyType="AICITYSTRATEGY_CAPITAL_NEED_SETTLER" FlavorType="FLAVOR_EXPANSION"/>
<Set Flavor="60"/>
</Update>

 

[markusbeutel]

Is it possible to cut out (or halve) science= population from Civ 5, and just leave science up to those generated by buildings and tile improvements?

Obviously you'd have to make the base rates of the buildings increase science output quite a bit, and include a few more science type buildings, but if its possible it might be a way to move away from the ICS strategy being the only viable one.

That said, even in the other civ games if you were playing with only a handful of cities it took quite a bit of skill to pull off, its just even harder in Civ 5. At least in this one you will be getting tons of social policies with so few cities...

The mod actually did this when it first started out - Citizens only provided 0.5 science each. The problem with this is that you need way more science buildings, and AI runaways were far more common-stance, because if an Civ missed constructing a few of these buildings, they'd be behind the eight ball right from the start.

Sorry to pop in the discussion but how did you mod AI expansion. I try to look files but didn't find them. I have been doing same thing (see my AI mod). My approach can be summarized next lines:

At defines:
<!--early REX-->
<Update>
<Where Name="AI_STRATEGY_EARLY_EXPANSION_NUM_CITIES_LIMIT"/>
<Set Value="5"/>
</Update>
<!--early REX-->
<Update>
<Where Name="AI_STRATEGY_EARLY_EXPANSION_NUM_UNOWNED_TILES_REQUIRED"/>
<Set Value="7"/>
</Update>
<Update>
<Where Name="FLAVOR_EXPANDGROW_COEFFICIENT"/>
<Set Value="4"/>
</Update>
<Update>
<Where Name="AI_CITY_SPECIALIZATION_EARLIEST_TURN"/>
<Set Value="6"/>
</Update>

At AIEconomicStrategies:

<Update>
<Where Type="ECONOMICAISTRATEGY_EARLY_EXPANSION"/>
<Set CheckTriggerTurnCount="1"/>
</Update>
<Update>
<Where Type="ECONOMICAISTRATEGY_EARLY_EXPANSION"/>
<Set WeightThreshold="-5"/> Side note not sure about this
</Update>

At AIEconomicStrategy_Flavors:
<Update>
<Where AIEconomicStrategyType="ECONOMICAISTRATEGY_EARLY_EXPANSION" FlavorType="FLAVOR_EXPANSION"/>
<Set Flavor="30"/>
</Update>

At AICityStrategy_Flavors:
<Update>
<Where AICityStrategyType="AICITYSTRATEGY_CAPITAL_NEED_SETTLER" FlavorType="FLAVOR_EXPANSION"/>
<Set Flavor="60"/>
</Update>

I tweak much of the same code. Some important ones you're missing include:

TargetNumCities in the CIV5Worlds.xml (increase these values by at least double).
AI_STRATEGY_MINIMUM_SETTLE_FERTILITY in GlobalAIDefines.xml (lower this value).
AI_STRATEGY_AREA_IS_FULL_PERCENT (raise this to 100%).
FLAVOR_EXPANSION for Settler -> raise this amount.

There's also the fact that in the mod, you start off with much higher happiness, and this is also closing off other options the AI would normally be focusing on instead of settling, (but I doubt you'd want to add happiness to Civs at the start of a game).

 

[joltis]

I tweak much of the same code. Some important ones you're missing include:

TargetNumCities in the CIV5Worlds.xml (increase these values by at least double).
AI_STRATEGY_MINIMUM_SETTLE_FERTILITY in GlobalAIDefines.xml (lower this value).
AI_STRATEGY_AREA_IS_FULL_PERCENT (raise this to 100%).
FLAVOR_EXPANSION for Settler -> raise this amount.

There's also the fact that in the mod, you start off with much higher happiness, and this is also closing off other options the AI would normally be focusing on instead of settling, (but I doubt you'd want to add happiness to Civs at the start of a game).

Thanks for info. The TargetNumCities is really new to me. Others I have been messing around. I think the values are quite good for me now maybe coming back to expansion in future. But I will move next to diplomacy and try to make thing a bit more better in that area.

I am fine with happiness which I see as a speed modifier in game (Not possible to all time make settlers; without happiness hit this game would be infinity settler spamming). AI can handle happiness issues quite well.


Joltis

 

[Mu'min]

Is it possible to cut out (or halve) science= population from Civ 5, and just leave science up to those generated by buildings and tile improvements?

Obviously you'd have to make the base rates of the buildings increase science output quite a bit, and include a few more science type buildings, but if its possible it might be a way to move away from the ICS strategy being the only viable one.

That said, even in the other civ games if you were playing with only a handful of cities it took quite a bit of skill to pull off, its just even harder in Civ 5. At least in this one you will be getting tons of social policies with so few cities...

The trouble is the pretty much linear relationship between pop and science. 12 pop 1 cities will produce the same science as 1 pop 12 city. When in reality 12 pop 1 cities would produce virtually no science, while a size 12 city would produce more science.

So, how about using the formula: science = pop of city/total number of cities in civ + bonuses + % mods.

That way ICS civs would be slowed down science wise, while still being able to out produce and out gold everyone else. And tall civs would still be able to keep up.

I hope that makes sense

 

[chaddougless]

The trouble is the pretty much linear relationship between pop and science. 12 pop 1 cities will produce the same science as 1 pop 12 city. When in reality 12 pop 1 cities would produce virtually no science, while a size 12 city would produce more science.

So, how about using the formula: science = pop of city/total number of cities in civ + bonuses + % mods.

That way ICS civs would be slowed down science wise, while still being able to out produce and out gold everyone else. And tall civs would still be able to keep up.

I hope that makes sense

I would like to state that I have yet to play this mod, but have been reading up so that I will have some idea what I am doing this weekend when I begin a new game. I am also relatively new to Civ5 in general. However, I can understand what you are saying about wanting to essentially disincentivize rapid ICS spread as a way to out tech a tall empire and I do believe that your point is valid. However, an easier way would be just to fix an integer number on it to scale it around based on population, so maybe start out at 1/4 science per pop and scale up to near 1 as pop grows (before bonuses). That would essentially look like Science = PopCity/(PopCity + X) where X is a constant. For this, let's say that X is 3. Your example would then yield 12 * (1/4) for the wide empire or 3 Science. The tall empire would yield 12 * 12/13 of a science or about 11 Science. Obviously this number can scale differently depending on the constant, but I do think it makes your example scale with pop in a city a little bit better. Unless of course you start to factor in transmission of scientific ideas across different locations as a way to speed up scientific ideas based on different platforms for observation.

 

[Mu'min]

Scrap what I said early. What we need is a logistic curve, or s-curve, as defined by the equation: P(t) = 1 / 1 + e(power of)-t. Where P=science and t = pop

the output of this equation would then need to be multiplied by the pop of the city, or maybe a figure derived from the pop and number of buildings in the city. That way a more developed city will produce more science than an undeveloped one.

 

[vonbach]

I'm running into the same issue. I'm just getting spammed endlessly with enemy cities
and units. I can never gain a real edge. Great engies are also pretty pointless.
Oh yes the not able to build wonders without a certain government types get
obnoxious real fast. Still its an interesting mod.

 

[markusbeutel]

Scrap what I said early. What we need is a logistic curve, or s-curve, as defined by the equation: P(t) = 1 / 1 + e(power of)-t. Where P=science and t = pop

the output of this equation would then need to be multiplied by the pop of the city, or maybe a figure derived from the pop and number of buildings in the city. That way a more developed city will produce more science than an undeveloped one.

That could work - but it's something that'd require .dll access.

I'm running into the same issue. I'm just getting spammed endlessly with enemy cities
and units. I can never gain a real edge. Great engies are also pretty pointless.
Oh yes the not able to build wonders without a certain government types get
obnoxious real fast. Still its an interesting mod.

The AI more or less expands at the rate they did in previous Civ games. CIV V is an exception, in that the AI just fails to expand, in many cases on even the highest difficulty settings. Mechanics like increasing social policy costs were put into place to discourage expansion, (and traits favoring tall empires) - were even put in place to promote this play-style. There's one problem with all of this: it doesn't work.

A large, sprawling empire will always dominate small/tall empires - especially when population = science. A slider from the previous games would actually do a better job of promoting tall empires... Rather than have the AI expand at a snail's pace, they expand at a competitive rate in Nights, and with less happiness bonuses than in vanilla CIV V. Coming from CIV V, their rate of expansion might seem too fast - but with practice, it's easily possible to keep pace, (although on Deity it will be very difficult). On Emperor/Immortal, to keep pace with the AI, you'll probably have to get Settlers out every 15 turns or so - which is entirely possible with proper policy/tech choices, and effective use of chopping and specialists.

 

[vonbach]

Having to expand like crazy just to match the computer is just silly.
I like wide empires myself but theres a limit. Another issue is the techs
the computer has no issues at all with science while you do. Theres so
many techs that you get mired in a tar pit of techs while the AI just spams
settlers and units. Or wonders. It just leads to a giant stalemate at best.
Getting swamped at worst.

 

[markusbeutel]

Having to expand like crazy just to match the computer is just silly.
I like wide empires myself but theres a limit. Another issue is the techs
the computer has no issues at all with science while you do. Theres so
many techs that you get mired in a tar pit of techs while the AI just spams
settlers and units. Or wonders. It just leads to a giant stalemate at best.
Getting swamped at worst.

Having to expand like crazy at the higher difficulty levels isn't crazy, it's how CIV V should've been designed in the first place based solely on the fact that it's always optimal to have as many cities as possible. As for techs and science, again - it's not difficult to keep pace with the AI with proper use of policies, tech paths, and expansion. If you're having trouble keeping pace scientifically, adopt Rationalism early on, invest in science specialists, build mines, sign some RA's - and focus on recruiting barbarians to provide your early military power if possible.

As a newcomer to the mod, you probably haven't played all that many games yet - and with all due respect, you can't expect to just jump in and dominate - this isn't vanilla CIV V.

 

[BTAxis]

I kind of feel the same way though, Markus. I've tried several games with the mod now, and I find I can never really keep up with the AI - I can keep up in some aspects if I focus on that, but other areas inevitably suffer.

I've repeatedly read comments like "it's possible to do X with proper techs, city placement and policies", but the flipside of that, and this is important, is that if you DON'T play with absolute efficiency you'll get swamped due to the crassly cheating AI. There's no room for mistakes. And that's not a good thing.

I've played vanilla, and yes it is too easy to beat the AI. But in NiGHTS, it is too hard. I can't enjoy the game this way.

 

[markusbeutel]

I kind of feel the same way though, Markus. I've tried several games with the mod now, and I find I can never really keep up with the AI - I can keep up in some aspects if I focus on that, but other areas inevitably suffer.

I've repeatedly read comments like "it's possible to do X with proper techs, city placement and policies", but the flipside of that, and this is important, is that if you DON'T play with absolute efficiency you'll get swamped due to the crassly cheating AI. There's no room for mistakes. And that's not a good thing.

I've played vanilla, and yes it is too easy to beat the AI. But in NiGHTS, it is too hard. I can't enjoy the game this way.

What difficulty are you playing on? Between chopping, free policy workers, (and settlers), barb recruits, and specialists, you can change things up a lot and still stay ahead of the AI.

Lowering the difficulty would be my best advice, as there are players beating the mod on higher difficulty levels. Granted i've played the mod a lot, but I can easily vary up my playstyle and have 7-8 cities up after 100 turns while fielding a competitive army and maintaining a top science score on emperor and immortal.

The AI also gets less tech and happiness bonuses than in vanilla CIV V.

 

[vonbach]

Once again your answers aren't exactly helpful.

 

[BTAxis]

Warlord, most of the time, which is difficulty 3.

In one game where I shared a small continent with China, playing as Iroquois, I thought I was doing reasonably well, getting 4-5 cities up against Wu's 2, and I was ahead in score too, but then she declared war on me (because I boxed her in with 2 cities). Although I'd gathered up some barbarian units by then, I was still vastly outnumbered and outclassed, and I lost my nearest city. And never mind the other AIs, most of them were far ahead of me points wise at that point.

I reloaded an autosave to see if I could save my city, in which I succeeded, but I was unable to turn the war around. Wu's cities had grown much faster than mine and she had built walls around them (strength close to 30). I basically couldn't touch them at all, not even with multiple siege weapons. Since the continent was pretty much full by then I was basically at a stalemate.

Now, it should be clear I'm not the best player around. I tend to pick techs that are easy to research and policies that will benefit me the most at that particular moment, because I can't anticipate what I will need later very well. Still, it felt very much like my enemy got advantages that tended toward the unfair. I'm hoping that, in time, the AI can compete without relying on such drastic bonuses.

 

[vonbach]

This is the issue I found it just turns into a stalemate at best. As far as I've seen this mod just isn't balanced.