The best and the worst AI's

[Lone Wolf]

The title says it all. Which AI tends to be the best and which one performs badly?

Good AI's -

Balseraphs. Apparently, Festivals beeline isn't a that bad thing. They are one of the few AI's that consistently aviod economical crashes.

Lanun, esp. Falamar. He tends to be the top in score. Possibly combination of +1 food from coast and the increased caps of Cha/Exp. Good relations with woman leaders may also be a factor . The Lanun, too, have few economical crashes.

Flauros. His beeline is rather good, vampires are strong, and Fin/Org + his favourite Aristocracy makes for quite an economy. Alexis usually performs worse.

Varn Gosam isn't usually that good, but in ErebusContinent, a script with flavour start and all the floodplains surrounding his position, he can become quite a tech monster.

Svartalfars experience economical crash much more often then other civs mentoined here (Hunting is rather an expensive tech), but their usual Elven economy is nothing to be sneered at.

The Elven economy also manifests inself in the Ljosafar.

Hippus, if they do rush someone (and they do it quite often).

Bad AI's -

Illians. I've seen a strong Illian empire only once. Even when I increased the bonus "raw" deer has from 1 to 2 food (since I play usually on ErebusContinent that has flavour start) it helped the Doviello, but not the Illians, with their manner of building Samhain in their only city .

Doviello can be OK/mediocre, if they suceed in rushing someone early game. If they don't, they are weak.

Bannor beeline, like the Doviello one, doesn't do them a lot of good. Their AI often economically crashes. They started to perform much better, when I changed the Bannor and Doviello beelines to Mining, and placed very high weights on Education. Other then that, they can be more or less OK if they have some Wine next to their starting position (since improved wine gives and and is on the way to BW).

Kuriotates. Aviod economical crashes, have a high score in the beginning, then are eliminated by someone more militiary strong. Once I've seen a strong Kardith, who succesfully defeated the attack of Tebryn's PZ's, and went down only because the foursome of Mahala, Flauros, Sheelba and Tebryn declared on him.

Speaking of the Sheaim, while their planar gates stuff can help them get quite effective in the mid- and late- game, and their PZ's are scary, they often economically crash and even if they are performing OK, AV is usually founded by some more "economical evil" like Balseraphs. Os-Gabella usually performs worse - it seems that her hatred of manhood isn't helping her diplomacy.

The AI experience bonus puts the Luchuirp at a militiary disandvantage, althrough their AI isn't really that bad aside from the usual ineffective AI way of using their hero, who, however, is very valuable to them.

Amurites are stupid in their refusal to trade techs, althrough they aren't that prone to economical crashes.

Hyborem. Poor guy, sitting in some faraway corner of the map, never amounting to anything. Makes you feel "Sympathy for the Devil".

 

[cIV_khanh93]

maybe this should be taken as a list of improvements neded by a certain Kael?

 

[kenken244]

Personally, I think that simply removing all the beelines would make all the AIs much better. Right now I believe the sheiam beeline sorcery, which has almost no economic benifits whatsoever.

 

[JonathanStrange]

My experience is mainly large, pangaea maps, deity level, playing the Ljosalfars, raging barbs. BUT I have played LOTS of games, admittedly I start get bored around turn 400-500.

My experience:

Ljosalfars: I'm usually them, and usually hardpressed in the beginning, but the AI Ljosalfars seem to do a bit better, though perhaps their weak start hurts them too. Still the AI Ljosalfars have raised large nations when they survive. Treants are overrated.

Balseraphs: Usually a contender, usually very strong. Crazy but true.

Hippus: Often successful in the early game, if too successful tend to run into problems (money? over-extended empire?). Bad neigh-bors.

Clan of Embers: Variable, but rarely hopeless, their aggressiveness helps them beaucoup. Do tend to like building and hanging on to Ritualists (that is, the Ritualists stay at home).

Svartalfars: AI Elves again seem as competent as the Ljosalfars, usually moderately successful. Rarely knocked out early but I have seen that too. Look better on paper or played by human.

Elohim: Mostly pitiful, usually destroyed in the middle game. I can't recall seeing any large Elohim nation. I can recall more than once their casting their worldspell "Sanctuary" needlessly.

Kuriotates: Often weak or destroyed early if near an aggressive civ. If they get an isolated start, they do moderately well. Never have seen them close to winning. Neither they nor I know how to play 'em.

Luichirps: Few Luichirp nations survive long in pangaea maps, in my experience. I'm not sure why. I don't play 'em. Once I did meet Beeri Bawl at the Olive Garden. Dude's like 3 feet tall, but tough looking.

Malakim: There's a Malakim advantages thread which is interesting but the only sign of AI Malakim nations I've seen are a few broken city walls and ruined temples. Hell, Varn Gosam only appears to announce his nations' extinction. The last several pangaea maps I've played, the Malakim disappear fairly quickly.

Doviello: I've seen huge Doviello nations; I think if the AI can knock out a rival civ and get their conquest rolling early they can be fiercesome. Make many troops but usually all beastmasters or something like that. No early kills, no strong Doviello presence.

 

[Lone Wolf]

Right now I believe the sheiam beeline sorcery

In the code, they beeline Sorcery only if they have KotE.

Luichirps: Few Luichip nations survive long in pangaea maps, in my experience. I'm not sure why.

Golems, by their nature, don't receive AI free experience. That puts the Luchuirp at a militiary disadvantage.

 

[Kyroshill]

I have seen no greater factor in determining future success than the ErebusContinent map as it applies to Varn Gosam (and possibly the Malakim regardless of leader)....

They always get mad floodplains.... since EC tends to have many rivers and also uses flavor starts (read: Malakim in deserts)

 

[DarkxL0rd]

i find falamar runs away with the lead early, often, and consistently.

Ljos are always a dominant factor

hippus usually do well

everyone else varies in my games,(i usually play svarts)

 

[Grakor456]

I don't think I've ever seen Kuriotates do well. From my observations, the AI doesn't seem to understand how the cities work. Instead they overlap them and then send settlements sprawling everywhere as if they were playing any old civ.

In the code, they beeline Sorcery only if they have KotE.

But isn't KotE the Sheaim beeline anyway?

 

[Nihilus]

in my own version of the game i have removed the Grigori adventurer mechanic and instead made it that every Grigori unit starts with Valor and Spirit Guide.
this was mostly meant for flavor/balancing/gameplay for a human player, but i was amazed at the impact it had on the AI. since this change Cassiel is consistently among the strongest/best AI players. makes sense in retrospective, instead of a few highly important hero units (that the AI always fails to keep alive) every unit becomes a little hero so to speak. more importantly when the AI uselessly suicides units half of the XP that would normally be lost is distributed to some other unit, thus speeding the overall XP gain of the Grigori.
luckily the Cassiel AI is not very aggressive or he would be a real killer with that change ^_^

 

[Lone Wolf]

Cassiel is usually OK/mediocre in my games.

But isn't KotE the Sheaim beeline anyway?

In Python it's not coded, althrough maybe they do beeline KotE due to weights.

 

[Emptiness]

since this change Cassiel is consistently among the strongest/best AI players.

Yeah, because giving all units Valor and Spirit Guide for free is a massive boost. Give that to anyone and they'll be much stronger.

 

[xalien]

I pretty much agree with AI rewievs so far, they're very consistent with my own experiences. Just to add that ljos seem to be more successful when led by Arendel than 2 other leaders, maybe because her traits are more "economical"? Svart are performing less well, their preference of CoE instead of FoL hurts them a bit.

 

[cabbagemeister]

Hippus (both leaders) always seem to be the major force in my recent games. Luckily I haven't started next to them in a while, but that just means that by the time I meet them they've always annexed at least two other nations and usually have a vassal.

 

[Domingo]

It's been discussed many times the idea of having the "beelines" become less weighted to make them more of "inclinations" (to allow the AI more flexibility). As far as I know nothing has been done in this direction recently though.

 

[Rystic]

The Lanun AI, while competent, needs to be taught about Pirates Coves.

Illians almost never build Temple of the Hand. We should see more of that.

Hyborem needs to start more wars. If he could conquer some cities, he'd be fine. As a general rule for he and Basium, after about 5 cities, they really take off the ground.

 

[JonathanStrange]

I wouldn't be surprised if the 1/2 of 1% of FFH2 players who post here come up with dramatically different lists of best and worst AIs. But it's still fun to do.

 

[WarKirby]

Good:
Balseraphs
Ljosalfar
Lanun
Hippus
Clan of Embers

Bad:
Kuriotates
Luchuirp
Doviello
Bannor
Hyborem

These are just based on general success rates. how often I'm seeing specific civs achieving world dominance, or dying early.

 

[cabbagemeister]

The Lanun AI, while competent, needs to be taught about Pirates Coves.

I agree. In my most recent game I see Hannah with miles and miles of coast and nary a Cove to be found. I think if Lanun DID use coves, they would be a dominant force in every game (and we'd see some cove-nerfs to tone them down).

Hyborem needs to start more wars. If he could conquer some cities, he'd be fine. As a general rule for he and Basium, after about 5 cities, they really take off the ground.

I agree with this as well. Both also should be better about playing the diplomacy game, in terms of trying to start wars with bribes. Specifically, they should be willing to offer a LOT of techs to coax people into wars, because they benefit economically so much from those wars (Hybo especially). And RP-wise, wouldn't a devious Hyborem who promises knowledge and power in exchange for blood and violence be exactly correct?

Also, I saw a thread a while ago about how Hybo doesn't appear to be settling his manes enough, which just kills him economically because he not only works fewer tiles, but he pays maintenence for low-strength extra units.

 

[Black Attila]

Illians. I've seen a strong Illian empire only once. Even when I increased the bonus "raw" deer has from 1 to 2 food (since I play usually on ErebusContinent that has flavour start) it helped the Doviello, but not the Illians, with their manner of building Samhain in their only city .

I agree that the Illian AI is one of the worst and the Samhain ritual is the biggest cause of it. Instead of a ritual that does nothing for them, it should be an Illian only building that gives them some kind of research bonus or some units or something. Then you can make an event that automatically triggers the next turn placing Mokka and his frostlings down on the world as the ritual would do.

Illians almost never build Temple of the Hand. We should see more of that.

raise the Ai weight of the building to make it more attractive to build.

 

[Lord Haldane]

I don't think I've ever seen Kuriotates do well. From my observations, the AI doesn't seem to understand how the cities work. Instead they overlap them and then send settlements sprawling everywhere as if they were playing any old civ.

Their city founding AI is horribly magnified on huge maps, where I will often find the Kuriotates have created no true cities by the time other civs average four to five! Who wants to guess what happens to Cardith's tech rate...?