AI on gigantic eternity (and other settings)

MAvL

Chieftain
Joined
Jul 9, 2016
Messages
93
TL;DR: can the AI handle gigantic enternity? Which other settings can help and hinder it? (asked from a strong monarch player).

Hi!
First thing first, this mod is awesome.

I'd like an epic and challenging game. I am thus wondering what settings I could choose to sit my taste for epic but do not cripple the AI. Especially how wise is it to use "start as a minor civ" when AI seems mostly inept with early warfare (it can foster decent attacks, but will unecessarilly suicide units on defense, so much that its percveption of battle odds seems rigged/bugged).

Below are my ideal settings and the issues I've found or speculate about them :

  1. Difficulty level. I first tried monarch, my favorite difficulty in the standard game, hard enough to be challenging but fair (non cheating) enough that the game still feels symetric and that the variety of viable strategies is still wide.
    Free archery while understandable, is bugging me in standard BTS. It seems monarch AND AI also gets it for free, which sounds much more problematic as it is now a much more advanced tech relative to the start of the game (or maybe I don't remember correctly and it's just slinger?). Thats why I restarted and played a game in Prince (I'm 2200BC). Combined with "start as a minor civ", and them using their extra unit aggresively, I might be screwed right from the start.
    On my current Prince playthrough I'm comfortable first with a lot of bad plays from not knowing the mod that I reckon in hindsight.
  2. Eternity speed and gigantic map. It sounds so epic and I feel like game speed and map size should be correlated for good experience, but I am worried the AI might handle them both very poorly. It will callously take bad fight right from the beginning, like attacking my woodsman warrior for no reason and leave his city barely defended as a consequence. Also experience and choosing the right promotion looks a lot more crucial and I would expect AI to be bad at it, but havn't seen it in action (I've seen nonsensical amphibious promotions, but they might be free seafaring promotions). I've captured a city with myth of the Hero so at least AI is half competent in hunting.
  3. Map: Perfectworld 2 (Terra). I like Earth like map. Regular BTS AI is completely inept at colonising the New World. In my only ever Terra game (which disgusted me), I got the whole new continent minus 2 cities all for myself (ok, on Prince where I was maybe already ready for monarch). Stability issues and some barbarians options should make it much harder to seize the whole world all by myself, but still, I'd like to hear whether old world nations and new world raising barbarian civs are worthy opponents, or if I'm better off with the option "start anywhere" (kind of like vanillia continent map).
    Related notes : (how) should I tweak the number of civs with start anywhere? Can barb civs raise into fully fledge civs no matter what, or do they have to wait for a civ to die to take their "slot"?
  4. Start as a minor civ. It sounds fun on the surface. On Monarch and above there is the issue of super close AI killing your starting warriors with ease. On Prince and below, there is the issue of AI suiciding its fresh warriors against your woodman II Warrior fortified on a hill forest. I'm barely exagerating, and it's really bugging me how unconservative the AI is. It's not like normal speed emperor+ where aggressive AI can just take bad fights for the odd chance of screwing you and recover anyway.
    Without start as a minor civ, I could just make war to everyone in the hope that they reproduce their really poor play, but at least AI would know what to expect and could also concentrate their effort on me if I choose to go that way.
  5. Ruthless AI. I haven't tested enough to know for sure what difference it makes, but the descriptions I've founhd here and there are really enticing, making the AI less of a simulated leader and more of a competitor.
  6. No rubberbanding. Tech diffusion and revolution should be enough.
  7. tech trading, no tech brokering.
  8. Surround and destroy. The AI seems completely helpless and clueless about that one, which is sad, but much understandable. I rolfstomped 3 cities very early (around the same time as my 1st and 2nd settlers, given that I build 3-4 warriors before) with a single javelineers and a couple warriors by just exploiting the mechanics. So i'll disable it.
  9. Ranged attacks. I didn't intend to use it anyway but out of curiosity, does AI handle it properly?
  10. Balanced resources. I don't remember if that's an option and if it impacts the food resource around the capital. I most likely won't used because I crave imbalance, but the AI seems to always settle in place. While sometimes suboptimal in regular BTS, it is always safe. OTOH, in AND, the starts are a lot less shiny, but there also are a lot of capitals with really crappy start if you settle in place. Especially considering that it takes a while to push your borders around your whole BFC.
  11. Advanced espionnage. No clue about it, sounds fun, does the AI handle it? can I toogle it in game? (so I can start a game and deactivate it before it matters if you anwser me AI is bad with it)
  12. high to low. I don't mind leaving an empire at its height, but getting used to a civ build crappily by someone else is very tedious, but I guess it might be a solution. It kind of both adds and takes away from the epic (takes away from the progression and adds to the world immersion). Problem I foresee is that you might get high scores by simply over expanding and never have to face the consequences. But at least it will act like a late start, mitigating the effects of "start as a minor", and like an start with odds, feeling better or at least different than just raising difficulty for AI to cheat.
    Question, is there any other possibility to switch civs? Like in RFC where you can change once, or like in Cursader King 2 where you can restart with anyone from any save? Just to get to see most of the world more freely, and to make interesting play.
 
Last edited:
4. I always play with Start as Minors (and slow big maps) but I also reduce the number of starting civs by 33-50%.

5. AFAIK Ruthless AI doesn't work so well in AND2.

9. My experience is that the AI is using Ranged attacks.

6. and 12. Revolutions and High to Low is never meant to work together. You will likely to get a tiny one-city civ in a very bad position as a 2nd civ.
 
Have you ever finished even one game at gigantic map? I am not sure it is possible at all - in modern time PC is moving too slow and then your starting to get CTD too often.
 
I had not looked at the starting techs before and I agree they need to be adjusted a bit. Archery is a dead-end tech in BTS so giving it to the AI gives them an immediate edge in units but doesn't springboard anything in terms of other techs. In AND this is not true and so the tech should be something different.

So I think this is what the bonus techs should be.
Monarch: Hunting [replaces Archery].
Emperor: Hunting, Weaving [replaces Hunting].
Immortal: Hunting, Weaving, Agriculture.
Deity: Hunting, Weaving, Agriculture, Stone Tools [replaces The Wheel].

The flipside is the techs the player gets at lower than Noble difficulty.
Warlord: Stone Tools [replaces The Wheel].
Chieftain: Stone Tools, Agriculture.
Settler: Stone Tools, Agriculture, Hunting [replaces Mining].
 
Vokarya, how AI will build Archers without Archery?
And without Archers it will be too vulnerable to early rush.
 
Vokarya, how AI will build Archers without Archery?
And without Archers it will be too vulnerable to early rush.

If the AI starts with Hunting, then all their defensive units are Slingers. Slingers are 2 Strength with +50% city defense, so only a little less than Archers. It's also harder to rush in AND because Axemen come later and Chariots have a city attack penalty. I think we should try it and see.
 
Slingers are 2 Strength with +50% city defense, so only a little less than Archers

Archers are 3 strength with the same +50% city defence ( and first strike, and 25% hill defence ), so it is more than 50% stronger.

In general:
Archers can defend cities from Axemans or Chariots.
Slingers are totally unable to do it.
I understand that AI will anyway get archery soon. But with archery at deity it got 4 archers at turn 0, and it is good defence from any rush. 4 slingers - not so much and he need later to find some gold to upgrade them.
 
I'm not worried about Axemen. Axemen don't show up until Metal Casting, and that is a Classical Era tech. The AI should get Archery long before anyone can research MC.

Javelineers can defend a city from Chariots. Chariots attack at 3.75 (base 5, -25% from city attack) and Javelineers defend against them at 4.5 (base 3, +50% vs. mounted), and that is before any city defense bonuses.

If necessary, we can increase the defense bonus of Slingers. As long as they don't defend better than Archers, it will not be a drawback to upgrade from them.
 
Javelineer rush and Spearman rush in AND against slinger might be a bit stronger than respectively Axeman and Swordsman rush against archer in regular BTS and similarly easy to pull off.

I've played another game with different settings, mostly more standard ones, most notably without "start as a minor" and with "limited movement", and I think AI having archers or archery would not be a problem under such settings ("would" because I played prince, so I haven't really tested). At least as far as using them aggressively, it's true that might them more impervious to rush than they ought to be.

I really like the idea of keeping rush possibilities open, but if AI is really too subpar at dealing with them, how about giving them archers without archery? Like one or two (or more) military police archers for the capital and one or two per newly founded city (depending on difficulty level).
 
I really like the idea of keeping rush possibilities open, but if AI is really too subpar at dealing with them, how about giving them archers without archery? Like one or two (or more) military police archers for the capital and one or two per newly founded city (depending on difficulty level).

The XML does not work that way. All you can do is dictate the number of units in each broad category: defense, exploration, worker. The game then calculates the best unit available in each category and creates that number of units. It's also how the game handles starts beyond the Ancient Era. (This is why the game generates Shaolin Monks as defensive units on Medieval starts. They are the most powerful defensive unit available with Classical technology.)
 
Top Bottom