City founding AI.

[Epicnessman9000]

Bad city spot:

Really bad city spot:

Really, REALLY bad city spot:

EDIT: But what bothers me most is not the sheer poorness of that settle. Its this, which only came to my attention recently:

Look at that PERFECT city location right there. Within your territory, on your island, easily defended and supplied, and almost completely overcrowded with resources. You get access to three (later four) strategic resources, which get massive buffs with rationalism (which Polynesia has invested in, the culturemongers them), and even a chance at the Petra if you were quick enough. Buffalo and wheat provide some nice yields, and get your city off the ground faster. Theres a good balance of hills, fertile lands, and even lakes for more profitable farms. Theres ALSO a luxury resource on top of all that, which as you can see Polynesia already owns a few of, pushing it closer to a monopoly.

But no. We can't build any Moai there. Trash tier.

We must settle in the middle of someone else's territory, completely destroying our already strained relationship. Make sure its stupidly far and vulnerable, that it has no resources or production whatsoever, and that it is placed in a way such as to invite maximum pisseyness from our militarily far superior neighbor. We must also make sure to lose the one redeeming aspect of the settle, claiming that tasty natural wonder and using our 50% bonus yields UA.

I assumed that unclaimed land was all desert or tundra or completely lacking anything of worth (which the south was, except for a respectable amount of forests, which have some potential) and thus Polynesia skipped it and went to settle islands farther away. When I first discovered it, I probably had a glancing thought that Polynesia would settle there soon. The stupidity of the AI continues exceed my expectations in ever greater quantities.

As a side note, every single city Polynesia owns is coastal. Perhaps the weight put on coastal cities is too high?



EDIT AGAIN: I think I have also found out why Polynesia isn't settling that spot:

Seems like any city that isn't a coastal city is voodoo for Polynesians, but is okay for even other very naval focused Civs like Carthage. Also explains why every single city they founded was coastal. Looks like that other forward settle was some unfinished business. I can see Dido telling Kamehameha all about her sovereignty and how she settles wherever she pleases and if you don't like that you can go bugger yourself. I know Kamehameha's a nice guy, but this was a bit far, especially in the CLASSICAL era. Too bad the bastards spent so much time grudge fighting each other the vikings rolled them both.

I may have been wrong about the AI. Its developed to the point where they spite each other like real players do, even being capable of irrational and emotional thinking. Even if that thinking is mostly used to evaluate long settles that tick the other guy off the most.

 

[infidel88]

It's not even funny. It's like AI deliberately tries to lose his settler and start a war. It happens constantly, AI can go 9-10 tiles to found their 2nd or 3rd city. Which is silly, as it's hard to defend and, if relevant, completely ahistorical. It's like in real life Chinese civilization would go to India or Indonesia to "found their cities" because... reasons.

One of the problems behind this are completely irrelevant barbarians. I usually play on Emperor with raging barbs and can't remind when was the last time barbarians were actually posing any threat to my city. Or burning my improvements. Or whatever.

And as a fallout, we get something mentioned in other topic - by the time you get pioneers (not to mention colonists) there are almost no places to settle, which makes late colonisation non-existant.

 

[Horvath Broncos]

It's not even funny. It's like AI deliberately tries to lose his settler and start a war. It happens constantly, AI can go 9-10 tiles to found their 2nd or 3rd city. Which is silly, as it's hard to defend and, if relevant, completely ahistorical. It's like in real life Chinese civilization would go to India or Indonesia to "found their cities" because... reasons.

One of the problems behind this are completely irrelevant barbarians. I usually play on Emperor with raging barbs and can't remind when was the last time barbarians were actually posing any threat to my city. Or burning my improvements. Or whatever.

And as a fallout, we get something mentioned in other topic - by the time you get pioneers (not to mention colonists) there are almost no places to settle, which makes late colonisation non-existant.

I agree that AI has some quirks when it comes to settling, but settling city 9-10 tiles away isn't that much. I do that pretty often to prevent someone from settling there and then maybe settle in the middle later.

 

[matlajs]

I do that pretty often to prevent someone from settling there and then maybe settle in the middle later.

agree.

 

[Gazebo]

A friendly reminder that individual experiences do not a bug make. I can guarantee that I've seen worse city placement by humans that have sent me test games to debug in the past.

The AI is substantially better at settling now than in vanilla. And that's the goal. Not every city perfect. But better.

G

 

[Ixenzo]

Slightly tangential: is it theoretically possible to implement a setup option that would in some way limit maximum distance a new city can be founded at? I.e. you "can't settle more than %number% tiles away from any of your existing cities on land or across coasts" hard limit or soft cap "it will take more resources to maintain the city the farther it is (increased building maintenance cost?)". Settling islands over ocean would require at least pioneers or something like that.

I'm just pissed at AI settling their cities half across the map just to be a dick to me. They won't be able to defend it in time anyway and such cities should be a liability AI needs to consider each time. No, I don't forward settle myself.

 

[notque]

Slightly tangential: is it theoretically possible to implement a setup option that would in some way limit maximum distance a new city can be founded at?

Yes, and I've used it in my AI mods before. It works really well in all cases except the need to settle an island, or where terrain makes it so that it is just past the max where it makes sense.

The other option that I used was to lower the values for tile already in the area of another civ, or taken already.

All reasonable options, but have their own negatives where once you set them, you now complain about those issues.

I did like the max distance settling setups better than the non-ones. AIs would tend to stay more of a connected territory that way, and it ends up best for them.

 

[grmagne]

One of the problems behind this are completely irrelevant barbarians. I usually play on Emperor with raging barbs and can't remind when was the last time barbarians were actually posing any threat to my city. Or burning my improvements. Or whatever.

I like this suggestion of making the early-game barbarians stronger so that civs are forced to either build cities close together, or earn the right to build farther away by fighting/clearing a path to the distant location (which is how I remember Civ 4). Right now, especially on the higher levels, it feels like a race in the early game to research pottery immediately and settle the good locations before the AI takes them all. A strong barbarian buffer zone between civs would do the trick.

The problem is that it’s very difficult to balance it all. I’ve seen screenshots from Barbarians Evolved where entire continents are overrun by barbarians. And then there’s the risk that the Authority policies will become too powerful if there are too many barbs to kill. But there must be some happy medium where they’re just tough enough to slow down early expansion without destroying the game balance.

And a lot of players don't even want barbarians in their games. It'll have to be a VP-specific mod rather than something added to the base game.

 

[infidel88]

Spanish Royal Court
Isabella: We need a new city!
Advisor: My Queen, all locations are taken, maybe we should conquer one instead?
I: Nonsense! We need cities! Bring me my map!
A: There you go my Queen, as you can see...
I: There!
A: Where?
I: There, that wonderful beach!
A: Umm.. My Queen, it's on the other side of the world, it would take us hundreds of years just to get there.
I: Nonsense! Imagine what a wonderful beach far away can do for us. Tenerifa rings a bell?
A: My Queen, but there is nothing there. Literally nothing!
I: Nonsense! I can see some fishies on the map. And delicious bananas!
A: But my Queen, just supplying this city should cost us way more than we can make from it. Ever.
I: Nonsense! We will build a lighthouse, the magical teleport inside will provide wormhole to our capital.
A: Fair enough. Alas, my Queen, but this is also in the middle of our sworn enemies lands. They will be mad.
I: Nonsense! Why would they. I settle wherever I please. I am Isabella! And blah blah blah!

Spoiler :

 

[Stalker0]

It might make sense to increase the weight of cities within initial trade route range for placement.

I do settle cities that I can't give internal trade routes to in the early game....but they need to be very nice spots. A weaker land spot that I can feed food or hammers too to get off the ground often does much better overall

 

[Gazebo]

It might make sense to increase the weight of cities within initial trade route range for placement.

I do settle cities that I can't give internal trade routes to in the early game....but they need to be very nice spots. A weaker land spot that I can feed food or hammers too to get off the ground often does much better overall

That's already how it works.

G

 

[Iamblichos]

I don't see why that's a bad city spot, it's a strategic location squeezing out potential threats and gaining military sea access to Carthage and Utique. It'll be a powder keg yes, but think of the history of Kypros here.

 

[Stalker0]

That's already how it works.



G

Sure, but the weight could be further increased if we thought it prudent

 

[Stalker0]

I don't see why that's a bad city spot, it's a strategic location squeezing out potential threats and gaining military sea access to Carthage and Utique. It'll be a powder keg yes, but think of the history of Kypros here.

If that city stretched its borders just a bit in the northwest it would actually cut out the sea lane access for 3 major cities...no joke if the city can be held.

The issue of course is when an AI is doing this. If it has its core cities, then options like this become more feasible, but as a second or third city option it's not ideal.

Does ai city placement logic change depending on the number of cities it already has?

 

[Funak]

I've seen really questionable city-placements as well recently, India settling his second city 4 tiles from my capital (and 18 tiles from his capital) and there were pretty much 10 different way better locations in between those two. I mean I'm purposely letting him get away with it because taking the city would cripple him, but the AI surviving on my mercy doesn't seem like a very good general strategy.

 

[Enginseer]

I actually like settling on one-tile lands. No one else?

 

[matlajs]

with at least 2 fish and 2 attols, why not.

 

[balparmak]

Seen similar behavior, eg: Darius is bordered by Gandhi and Theodora, but not by me, our relations are pretty friendly. He then settles faar away from his capital (and quite far from his closest city), at the meeting point of the two above and me, in a way that mostly harms my development. It was also a really bad spot with no outstanding resources. So in the end Theodora invites me to declare war against him, I'm forced to take his city lest the others take it, Darius remains crippled for a long while, and eventually gets vassalized. Alongside any chance he had, he also ruins my plan to play him to balance a runaway Theodora. This was the clearest example I can give right now, but in general my impression is that the AI values forward settling a bit too much, and generally does much better when it settles without considering to block others. If such a variable exists in AI evaluation that is, I may be dreaming all that.

I've seen really questionable city-placements as well recently, India settling his second city 4 tiles from my capital (and 18 tiles from his capital) and there were pretty much 10 different way better locations in between those two. I mean I'm purposely letting him get away with it because taking the city would cripple him, but the AI surviving on my mercy doesn't seem like a very good general strategy.

 

[Epicnessman9000]

You're right, if they could hold the city it might be of good tactical positioning. If.

The city is a dozen dozen tiles away by sea from their large home island.

It has no production except for some stone. Which can be easily pillaged. And just happens to be on an tiny island that offers great vantage points for siege weapons. Theres even more spots for siege weapons once artillery comes into play.

Circled by 4 (5 counting Gades) port cities.

Belonging to a major naval power.



But what bothers me most is not the sheer poorness of that settle. Its this, which only came to my attention recently:

Look at that PERFECT city location right there. Within your territory, on your island, easily defended and supplied, and almost completely overcrowded with resources. You get access to three (later four) strategic resources, which get massive buffs with rationalism (which Polynesia has invested in, the culturemongers them), and even a chance at the Petra if you were quick enough. Buffalo and wheat provide some nice yields, and get your city off the ground faster. Theres a good balance of hills, fertile lands, and even lakes for more profitable farms. Theres ALSO a luxury resource on top of all that, which as you can see Polynesia already owns a few of, pushing it closer to a monopoly.

But no. We can't build any Moai there. Trash tier.

We must settle in the middle of someone else's territory, completely destroying our already strained relationship. Make sure its stupidly far and vulnerable, that it has no resources or production whatsoever, and that it is placed in a way such as to invite maximum pisseyness from our militarily far superior neighbor. We must also make sure to lose the one redeeming aspect of the settle, claiming that tasty natural wonder and using our 50% bonus yields UA.

I assumed that unclaimed land was all desert or tundra or completely lacking anything of worth (which the south was, except for a respectable amount of forests, which have some potential) and thus Polynesia skipped it and went to settle islands farther away. When I first discovered it, I probably had a glancing thought that Polynesia would settle there soon. The stupidity of the AI continues exceed my expectations in ever greater quantities.

As a side note, every single city Polynesia owns is coastal. Perhaps the weight put on coastal cities is too high?

 

[Navanod]

Ramesses II once sent a settler half way across the map, through hostile Greek lands all the way to settle a city in the 7 tiles gap between my 4 cities

The city can never expand and royally ticked me off. I suspect it is because Greece had totally blocked all expansion sites on their end of the map...