[NFP] Map and Gameplay Settings to Encourage AI Pursuit of Victory

dw0914

Chieftain
Joined
Jan 31, 2014
Messages
2
Hi all - been a while since I posted here, but a newborn in the family has me back into the occasional Civ game because it's a lot more compatible with interrupted gameplay :)

I know it's a given that the AI is bad at actively pursuing victories from a strategy standpoint. With that said, I have been experimenting with map settings to find a sweet spot where I can feel like the AI is at least trying to make a go of it, particularly if they have a strong position. For context, I have all of the DLCs but am otherwise running unmodded, and the games I've played have mostly been on Ancient start with Emperor or Immortal difficulty (Deity can be fun, but I don't like feeling pressured to early warmonger every game). Indeed, I mostly try to play a peaceful game, perhaps with some liberation, to see how the AI responds. Here's what I'm seeing for each victory type:

  • Culture: AI can kind of pursue this from a building / Great Person / theming standpoint, but never really kicks it into overdrive. Actually achieving AI victory this way would take 100s more turns than the typical game. Culture Victories are also often about targeting the other high Culture civs militarily, which they don't seem to do.
  • Diplomatic: I haven't had my ear to the ground for CivFanatics' opinions on Diplomatic Victory, but the AI's decisions are so predictable that I don't think it even makes sense to talk about "pursuit" of Diplomatic Victory (probably my least favorite victory type anyway).
  • Domination: AI can amass units, but of course some map types are better than others when it comes to actually waging war. But the AI doesn't really seem to "pursue" it i.e. press onward for capital after capital, even if it's strong enough to do so. Huge runaway Civs feel less common in VI than in V.
  • Religious: AI could do this so much better. Many games, I neglect religion, and there's a frontrunner AI generating boatloads of faith and converting maybe half the world, but then they just stop, or it would take 1000s of turns for the AI to actually accomplish this.
  • Science: AI actually kind of pursues this, pretty much by default, when it's in a strong position. They don't really defend Spaceports or anything intelligent, but if you let an AI build a good economy, they'll eventually lurch over the finish line via Science.
My games are usually Large map, default Civs (10) and default (or near default) city-states (~15). I tend to leave a lot of settings on Random (except for Standard Resources, Standard Start) and go with Fractal or Continents and Islands. I know the AI can sometimes do better with smaller maps / fewer Civs, but games with < 8 Civs just feel empty. I have been playing on Standard speed, though in Civ V I used to play on Epic. For some reason, Epic felt too slow in Civ VI, but perhaps changing speed is an option.

What I've found so far is that the AI fares best with less water and fewer hills / terrain obstacles. They also do reasonably well when they can knock off an early city-state or two, and perhaps when they can gain an early ally rather than just going to war (AI vs. AI wars often result in a loss of both armies with very little changing hands in terms of cities). Beyond that, though, I'm at a loss for what else I can change, specifically for settings that might encourage them to pursue Culture or Religion, aside from disabling other Victory types which I don't want to do. Right now, by mid-game, it mostly boils down to me having to keep the strongest AI Civ in check so they don't default into a Science or Diplomatic victory. Same situation, different game.

Any suggestions? I'm also not completely closed off to mod suggestions, as I know there are a few that modify the AI, but I'd prefer something that keeps the AI nimble in terms of victory types rather than just a mod that improves AI military tactics (which I know is the most commonly sought type of AI mod). Thanks!
 
I ran about 30-40 AI-only games (with Gathering Storm rule set and the latest patch), with various settings. The only mod I used was Better Spectator Mod v1.1.1.
I have generally the same view as you.
About 80-90 % of the games had Science Victories, the remainder were all Diplomatic Victories.
One thing that jumped out at me was that every AI almost always used almost all of its aluminium on military units which meant that it could not build Lagrange Laser Station and thus did not travel faster than 1 light year per turn. For example, I ran one game at marathon speed: the winning AI started traveling at around turn 850, and then traveled at 1 light year per turn for 150 turns.
I tried running games with sparse / abundant resources (my idea was to provide more aluminium), but this made no noticeable difference. With abundant resources the AI simply built more military units.
I tested with different AI leaders but they all showed this same pattern of play.
Another thing I noticed was that the average game length was about 20 turns longer (at Standard speed) with an Archipelago Map than with a Pangaea Map.
 
About culture victory - some civs actually CAN.
I am looking at you, wonderspamming and greatwall spamming China.
(Also religious civs can basically spam LaVenta heads every tile, so maybe maybe...)
Still do not expect it before turn 300.
I haven't tried it though on specific mapscripts, but I suppose if you run China against low culture civs on a mapscipt with little coast (lakes preferably) we can expect pretty reasonable date of tourism victory
 
The one thing I've noticed is that on maps that don't wrap (such as inland sea) the AI is absolutely awful and doesn't perform normally.
 
I ran about 30-40 AI-only games (with Gathering Storm rule set and the latest patch), with various settings. The only mod I used was Better Spectator Mod v1.1.1.
I have generally the same view as you.
About 80-90 % of the games had Science Victories, the remainder were all Diplomatic Victories.
One thing that jumped out at me was that every AI almost always used almost all of its aluminium on military units which meant that it could not build Lagrange Laser Station and thus did not travel faster than 1 light year per turn. For example, I ran one game at marathon speed: the winning AI started traveling at around turn 850, and then traveled at 1 light year per turn for 150 turns.
I tried running games with sparse / abundant resources (my idea was to provide more aluminium), but this made no noticeable difference. With abundant resources the AI simply built more military units.
I tested with different AI leaders but they all showed this same pattern of play.
Another thing I noticed was that the average game length was about 20 turns longer (at Standard speed) with an Archipelago Map than with a Pangaea Map.

So Helicopters are the problem, and a lack of power?
 
So Helicopters are the problem, and a lack of power?
I am not sure it is just about helicopters, I would say aluminum in general. For example, on archipelago maps I saw islands with aluminum that were never settled.
I forgot to look out for how the AI was using power in general, but I did look out for Terrestrial Laser Stations; I never saw the AI build one.
 
I forgot to say that I also saw that the AI generally builds more military units than it has strategic resources for, especially towards the end of the game.
 
Here is a screenshot to illustrate what I mean.
We are viewing from Matthias' (green colour) perspective.
We see that it is turn 315 (standard speed): Matthias' is traveling at 1 light year per turn; is researching Future Tech and Future Civic; is getting over 600 per turn of gold, culture and faith; is overusing aluminium.
 
Here is a screenshot to illustrate what I mean.
We are viewing from Matthias' (green colour) perspective.
We see that it is turn 315 (standard speed): Matthias' is traveling at 1 light year per turn; is researching Future Tech and Future Civic; is getting over 600 per turn of gold, culture and faith; is overusing aluminium.
I see a lot of Gurus too, which is something I also noticed. Really wish the strong religious civs would keep cranking out Missionaries and Apostles to convert other cities rather than parking a bunch of Gurus in their home territory. The AI almost never tries to convert my cities even when I have a weak religion. What motivates them to try?
 
I forgot to say that I also saw that the AI generally builds more military units than it has strategic resources for, especially towards the end of the game.

It would help if the AI built mines and oil wells to take advantage of all the resources they possess.
 
So the AI does not understand resource upkeep. Lovely.

Exactly. Just had a game where Mali had one source of uranium and in addition to a GDR already built, they were building two more GDR's AND a Nuclear Power Plant. Then I watched their GDR's get decimated one after the other due to the lack of resource upkeep.
 
I actually lost a game to an AI culture victory because they were spamming rock bands and one of them turned into a runaway train, and i wasn't paying attention. However, now I block rock bands in all my games (and the amenities are never a problem) so that's off the table now. I agree with your views.

The diplomatic votes are also maddeningly predictable. I wish there was some skill or horse trading involved at least.
 
This is why I just disable Diplomatic Victory. It's just easier and more fun to not have to think about it.

Yeah I'm a total purist when it comes to these things, I don't like turning anything off, but I should try turning it off too.
 
I almost always have Diplo disabled, I agree it needs to be reworked. I think it's the most boring victory type, even more boring than a score victory. :lol:
 
even more boring than a score victory. :lol:

Just played for my first score victory, standard speed. While it definitely was a slog, I did find it to be pretty challenging/interesting to hold back the progress of the entire world, especially towards the end as multiple AI civs became space capable. Managed to do it without resorting to war either! That said, I probably won't play for score again.
 
Might have nothing to do with it not being wrapped though. Inland sea typically has less rivers, so perhaps the AI suffers with city placement.

It happens on the GOT LAKES? map script when set to not wrap too
 
Back
Top Bottom