AI expansion oddness

[teakbois]

Current game allows Pangeas, and i got one. There is a sizable chunk of it quarantined off by a small mountain range, however very easily accessible in coastal waters. In fact, one could just park a galley and use it to cross. The AI, however, doesnt want anything to do with it, as seen in this photo:

Spoiler :

Instead, the AI would rather do the usual crowding of a city it couldnt hope to ever get more than a 9 tile radius out of

Spoiler :

However, the untouched portion of the continent is a fantastic playground for my ranger, who the barbs seem to be leaving alone. All my cities have well stocked cages/carnivals.

 

[teakbois]

They finally did cross over, guess they just wanted to make sure everything accessible by land was used up before they took to the sea

 

[Koshling]

They finally did cross over, guess they just wanted to make sure everything accessible by land was used up before they took to the sea

Yeh, the AI gets especially confused about land that is in the same area (aka same landmass) but can only be reached by water. It's not a simple issue to resolve, and since it's not that common it has never really been a priority.

 

[teakbois]

Im not even convinced they used a Galley, Im going to go back to a save to see. I saw a warlord chief roaming the mountain range, they could have used a lead troops over mountains promotion. They have tons of ships, but mostly Triremes and Dromons, havent even seen a Galley

However, their attempt to claim the lands beyond the mountains didnt go to well...

Spoiler :

 

[sp00n]

Yeah I had that once too, it was animal galore for my hunter. It didn't help the AI that the land before the mountain tile (it was really just one tile with a mountain on it that connected both land masses) was occupied with my cities.

 

[Nakkimunakas]

I've noticed the AI seems to have hard time expanding to a new continent or large islands overseas. This seems to happen on landmasses where every coastal square is occupied by animal or barbarian so the AI might not know how to land in that situation? I've seen them land troops on a spot where I've settled and killed all the animals off.

 

[Koshling]

I've noticed the AI seems to have hard time expanding to a new continent or large islands overseas. This seems to happen on landmasses where every coastal square is occupied by animal or barbarian so the AI might not know how to land in that situation? I've seen them land troops on a spot where I've settled and killed all the animals off.

This is true - it doesn't current consider amphibious attacks when planning its landing. That is just standard BtS AI, but high animal density on uninhabited landmasses make it much more of a problem in C2C

 

[Thunderbrd]

Must be why they were destroying off barbs who remained trapped within a confined space for too long. Is it something you think you could resolve to get them to consider amphibious attacks like that? Otherwise I guess you could hold off all enemies by stationing a unit on every coast adjacent tile (now that I think about it... I've done that before! lol Standard Vanilla but yeah... forced 'em to attack the city directly if they wanted in.)

 

[Koshling]

Must be why they were destroying off barbs who remained trapped within a confined space for too long. Is it something you think you could resolve to get them to consider amphibious attacks like that? Otherwise I guess you could hold off all enemies by stationing a unit on every coast adjacent tile (now that I think about it... I've done that before! lol Standard Vanilla but yeah... forced 'em to attack the city directly if they wanted in.)

I'm sure it's resolvable, but I don't have immediate plans to prioritize it.

 

[hotrodlincoln]

For a while, the AI did use galleys to transport units. When the barbarians got a hold of galleys, the game got very interesting. I don't know why that was done away with.

 

[Koshling]

For a while, the AI did use galleys to transport units. When the barbarians got a hold of galleys, the game got very interesting. I don't know why that was done away with.

That's not the poi t being made here. The point is that the ai cant figure out where to land (so doesn't start) if all the coastal tiles are occupied.

 

[TheDisco]

Does anybody think the AI doesn't value "marginal" terrain for cities enough? Whenever I have a map with a large desert/barren/salt/rocky/etc region, the AI is very very very late in trying to settle it. I usually end up grabbing the entire territory, taking my time, because the AI thinks it is worthless. But with all the food producing buildings in particular, it is real easy to make those cities worthwhile and the resources that pop eventually make them very worthwhile.

Last game I played me and my main rival each reached 12 cities almost simulateously, then he (and every other civ) stopped expanding and I ended up getting to 27 cities and my "rival" is still sitting at 12 or so. Now the game is obviosuly a blowout but it should have been closer...

 

[Koshling]

Does anybody think the AI doesn't value "marginal" terrain for cities enough? Whenever I have a map with a large desert/barren/salt/rocky/etc region, the AI is very very very late in trying to settle it. I usually end up grabbing the entire territory, taking my time, because the AI thinks it is worthless. But with all the food producing buildings in particular, it is real easy to make those cities worthwhile and the resources that pop eventually make them very worthwhile.

Last game I played me and my main rival each reached 12 cities almost simulateously, then he (and every other civ) stopped expanding and I ended up getting to 27 cities and my "rival" is still sitting at 12 or so. Now the game is obviosuly a blowout but it should have been closer...

Are you sure it was marginal terrain that sopped it, as opposed to maintenance costs or REV stability? The AI log would be sufficient to answer that question I think.

 

[TheDisco]

Are you sure it was marginal terrain that sopped it, as opposed to maintenance costs or REV stability? The AI log would be sufficient to answer that question I think.

Well, marginal terrain is what I assumed it was. In the last game there were at least 15 or so civilizations that could have expanded into the marginal zone but essentially none of them did (or at least barely did). Some of these civs had only 3 or 4 cities so I don't think it would have been maintenance costs or stability stopping them. And this went on for a long, long time. Long enough for me to go from 12 cities to 27 without hurting my own economy or stability.

My other theory was maybe the AI was getting scared about war and kicked into a warmonger mode, as I did declare one brief war right about the time they stopped expanding. I'm not sure how the AI works though so not sure if that is possible. But if so I think it needs to balance prioritizing expansion a bit more alongside war preperation at least over the long haul.

As for AI log I will try to track it down (not sure where it is, I haven't poked around in the mod files in a long time)

I kept hoping the civs would start expanding in there as now the game is not worth playing as my lead is so great now. Once those marginal cities can get to population 6, which isn't very hard with all the food buildings, they really become worthwhile.

 

[ls612]

Does anybody think the AI doesn't value "marginal" terrain for cities enough? Whenever I have a map with a large desert/barren/salt/rocky/etc region, the AI is very very very late in trying to settle it. I usually end up grabbing the entire territory, taking my time, because the AI thinks it is worthless. But with all the food producing buildings in particular, it is real easy to make those cities worthwhile and the resources that pop eventually make them very worthwhile.

Last game I played me and my main rival each reached 12 cities almost simulateously, then he (and every other civ) stopped expanding and I ended up getting to 27 cities and my "rival" is still sitting at 12 or so. Now the game is obviosuly a blowout but it should have been closer...

I've actually found the opposite, taht the AI will totally overvalue cities in marginal terrain (Permafrost). Those cities won't get much past size 10 or so, but the AI I've found will continue building them so long as there is room, even when there is open space in better climates.

My mostly uneducated guess is that the AI sees the founding spots farther north (this game was in the southern hemisphere for me) as more contested and dangerous, and so doesn't found there. Instead he is going for hte non-dangerous (but also crappy) frozen cities and therefore loses in the long run to his northern neighbors.

 

[TheDisco]

I've actually found the opposite, taht the AI will totally overvalue cities in marginal terrain (Permafrost). Those cities won't get much past size 10 or so, but the AI I've found will continue building them so long as there is room, even when there is open space in better climates.

My mostly uneducated guess is that the AI sees the founding spots farther north (this game was in the southern hemisphere for me) as more contested and dangerous, and so doesn't found there. Instead he is going for hte non-dangerous (but also crappy) frozen cities and therefore loses in the long run to his northern neighbors.

Hmmm, interesting. Using that interpretation, the large block of marginal terrain in my recent game was right in the middle of the giant continent. So by your theory all the civs were afraid of expanding there as it looked too contested? Maybe that is what happened, although my main early rival wasn't afraid to plop down a city right next to my terriroty in a spot I wanted to expand to (good terrain on coast).

Not sure what happened, but I would consider the AI expansion strategy broken in my last game for whatever reason. I'm gonna start a new game and hopefully they put up more of a fight for territory in the early game.