Does the AI stick to the same affinity entire game?

kaspergm

Deity
Joined
Aug 19, 2012
Messages
5,565
Or was my current game just a fluke?

It was only after I finished (my not very succesfull 1-city challenge) that I notice that the #1 civ went food on every era:

Harrapans - Celts - English - Haudenosaunee - Mexicans - Brazilians

Meanwhile, the #2 civ went military every era:

Hittites - Huns - Aztecs - Poles - Zulu - Zulu (again)

Have anybody else seen similar patterns?
 
Yes, AI personas tend to choose cultures that match their strengths and biases, so they often choose the same paths.

This can be useful knowledge that eg Mycenaeans and Hittites are most likely to become Huns.

If you want to see more variation try adding say 3 militaristic AI and 2 food-focused ones so they don't always get their first pick.
 
Yes, I noticed that, too. And I don't understand why. It's usually not optimal for them. It's not always the case though, even when all choices are available, sometimes they deviate. To be honest, I have no clue on what basis the AI picks its cultures at all.
 
Yeah, definitely a bad AI design decision there.
I disagree, some AIs should tend to be consistent as long as only some are consistent. (And I have seen them wander…although that might be because their ideal choice is gone)
 
It seems they pick culture based on both persona and a deterministic calculation of what cultures are more desirable. So you get AI who

1) specialize in food without picking up any of the cultures of other affinities that would allow them to actually apply that food to production/science,

2) deprive the other AI who would benefit from a chance to get a food culture

3) deprive the player of choice and variety if the same cultures are picked by AI each time.

4) the popular cultures also have really lackluster EU later in the game, at least in the hand of a leading AI (Mexicans, Zulu etc, which behave better when catching up, not leading).

Especially in late game; when you only usually interact meaningfully with 1-2 industrial and contemporary AI, it is quite disappointing that it’s most often transcendent zulu.

If the AI did a quick calculation of what affinity they needed most, you’d get variety each game and your get AI actually benefiting from the cultures they pick.

It seems the goal was to create flavor in the personas, but it has the opposite effect for me.
 
Thanx for all the answers. I don't think I even understand how the AI personas work. :blush:
 
So I'm not the only one who has to face the runaway transcendent zulu. I assume it's because there is no militarist civ in contemporary.
 
A real shame, usually taken by AI with tech lead, but Zulu would be awesome if you skipped science, since no tech requirement for Impi. Either you or another AI would suddenly be a major military contender. But what the leader actually needs to do is research homeland and line infantry. Too much fighting draftee-Impi armies with muskets that don’t even trigger the Impi bonus.
 
A real shame, usually taken by AI with tech lead, but Zulu would be awesome if you skipped science, since no tech requirement for Impi. Either you or another AI would suddenly be a major military contender. But what the leader actually needs to do is research homeland and line infantry. Too much fighting draftee-Impi armies with muskets that don’t even trigger the Impi bonus.

Not entirely my experience. In my last game, the largest Faction on the other continent (3 continent map with New World) was not the Tech Leader (I was, by a good margin), but had expanded rapidly early in the game as Harappans - Huns, took Zulus in Industrial (can't remember their Medieval/Early Modern picks), transcended and mopped up both the other Factions on their continent, giving them a boost of Militarist and Expansionist Fame which durn near caught up with me. Unfortunately for them, I had the tall man with the wide-brimmed hat and he nodded his smokey head over a couple of their cities to give me time for a Go To Mars win.

But I find that the AI makes some pretty good decisions when it comes to transcending throughout the game. The fact that they can't match a human for long-term strategy is not really relevant - neither can a lot of humans!
 
I’ve seen this but it’s not terribly common, except that whoever picks Carthage always picks Byzantium. Another weird thing is I’ve never played as or against the venetians, are they just that bad or does the AI not like them for no reason?
 
With the new beta, I got Beowulf the Militarist taking Nubians, which was a surprise, as the Militarists tend to favor the swords... and yes, there was one Mil culture available when they took it (don't remember which one)... so there is some variability there, don't know if it was already there or it came with the beta.
 
Actually, I had one time where the mongols turned into edo japanese and they went aestathic way and started to run away, it was definitely a challenge to deal with that one instead of the usual agrarian one or militaristic one...
 
It was only after I finished (my not very succesfull 1-city challenge) that I notice that the #1 civ went food on every era:

Harrapans - Celts - English - Haudenosaunee - Mexicans - Brazilians

Meanwhile, the #2 civ went military every era:

Hittites - Huns - Aztecs - Poles - Zulu - Zulu (again)
I've seen both of these a lot, but then in my last game an AI personality that did not have the Transcend quirk went ??? > Greece > Greece > Greece > France. :crazyeye:
 
I've seen both of these a lot, but then in my last game an AI personality that did not have the Transcend quirk went ??? > Greece > Greece > Greece > France. :crazyeye:

Even the AI knows you can never have too much Greece :lol:

Just don't tell Xerxes. He agrees in a different way... :shifty:
 
Even the AI knows you can never have too much Greece :lol:

Just don't tell Xerxes. He agrees in a different way... :shifty:

Hoplites with their 'adjacency bonus' are just about as good as PIkemen, so remaining Greek doesn't give you much of a Military Penalty until just before gunpowder, while their Amphitheaters give you both Science and Influence, both of which I find myself scrambling for in the first 13 to 1/2 of the game.

And going France in Industrial looks like an attempt to rush for Science in the end game: and it never hurts to have all the late game Infrastructure and Units available while your opponents are still marching with matchlocks and plate armor . . .
 
counterpoint: amphis are sort of weak considering they're research Qs that don't exploit science tiles

True, but they are one of the earliest Science Qtrs available, do 'exploit' adjacent Districts of all kinds for Science, and add Science per Era. Not as good as some later EQs, perhaps, but I find the critical Science advance rates occur in the Medieval, when it easy to start falling behind in Science progression compared to the game pace, and the Greeks can get ahead of that nicely.
 
Top Bottom