So did GalCiv II have good AI or not?

GalCiv2 had a great AI. Consider what else was in the wild at the time and the fact that it was a 32 bit game.

Those of us who got caught up in the scoring game used exploits that took months and several teams of players to devise. 

Facing a Suicidal AI on a level field without giving ourselves buffs or using cheese would in fact be suicidal.  I doubt any one of us could beat it without cheese tactics, reloading, etc.

I know you worked hard on the AI there and you should take pride in it. 

The Ai is pretty good.  Normal is way too easy for anyone but beginners, challenging is too easy for good players but gives an ok game.  Above that the AI gets some strange perks like no FOW for them and random shot at moneys and techs / turn.  It gets pretty aggressive on godlike, not for the faint of heart.   SD is still doing work on the Al so it will keep getting better.

gal civ 2 ai was weird,　when using the max number of ai on the biggest galaxy 1-3　ai randomly would be active then the rest of the ai would be much less active like in some kind of hibernation (probably what people are calling braindead). Starting new games changed which race would become active each time too.

I liked the GalCiv II AI, I always think there are areas of improvement, however, I don't remember any specific flaws.

I loved it. That game got me interested in SD thus provided the curse that has plagued you with my continued presence on these here forums.

The AI was the single best bit about galciv2. By far.

I cant wait to see galciv3 AI reach a similar point

Yes The GC2 AI was better than what I have come across. If there is two words to sum up the GC2 AI it would be devilishly annoying. Ships running away flanking threw the FOW picking off weaker fleets and so forth. The AI seemed much more aware of the player and knew when to run and attack compared to most other games without cheating.

The only complaint I can really recall is when the AI would as mentioned above, hibernate or seem to go in depression and not be active during long games.

I know allot of people just want to non stop attack and conquer but the AI was robust enough to provide a nice atmosphere for those of us who like to play USA style and police the galaxy and see how it shapes over time.  For a immersive play style a good robust AI is pretty important. GC2 accomplished that. 

I guess you could say those of us who play long drawn out games are subconsciously admiring the AI do it's thing. I mean heck if the AI was boring I'd just kill em and move on.

 Look at it from this perspective; Just non stop conquering game after game is more brain dead when there are so many different ways to win and enjoy doing it. This is Galactic "civilizations" after all.

Brad,

I don't know the bonuses / maluses the AI got on GalCiv2, but my last game (i checked) was a Huge map, Normal scenario, as Altarians, against 9 (random) AI, difficulty level Crippling. At those game settings i consider the game winnable while delivering a sufficient challenge to be enjoyable/relaxing.

There are higher difficulty levels, but these appears to have been my goto settings for relaxing games. Looking in the high score table however a few are at Painful, most of them are at Crippling, some are at Masochistic and one at Obscene.

So if you can make the GalCiv3 AI at Normal perform at least like the GalCiv2 AI at Crippling then i will be a happy camper.

However my thoughts on the AI depend on which facet of the AI we are talking about. Trading and diplomacy were weak (easily exploitable and sometimes strange), but the strategic and tactical AI were well above the AIs of other games.

I like the diplomatic depth of (GalCiv2 contemporary) EU3 and the ancient "Balance of Power" game, not because their AI is great but because every decision you make has consequences some of them quite complex.

PS,

my GalCiv2 edition is the "Galactic Civilizations II Ultimate Edition" in case that matters for the AI levels.

I'm going to say "No", GalCiv2 did not have good AI. I'll explain:

(Disclaimers: I only played Twilight and not the previous expansions or vanilla game. I enjoyed the game for many many hours and it was well worth the money. Not complaining at all.)

The basic problem was that the AI was very complacent and apparently not interested in actually winning the game. Symptoms:

• Declares wars but makes little or no effort to invade or take territory.
• Sends solitary transports to unlikely targets with no realistic chance to take the planet (this shows there's no "plan", just targets of opportunity).
• Starts wars with unreachable opponents.
• Foaming infatuation with destroying starbases, even with no possible chance to win the battle.
• Fleets tended to target unimportant fleets on the other side of the galaxy and would send forces clear across the map (30+ turns away) to get to them.
• Wave after wave of crappy fighters would throw themselves at my superior fleets.
• Exceptionally easy to persuade AIs to go to war with each other. (War should be costly if not fatal)

Pretty much every game came down to this:

• Grab as many planets as possible
• Turtle until i can get Large hulls and weapons
• Build doom fleet and ravage the galaxy.

@Brad: if you can find a way to disrupt that formula, that makes the game much more interesting.

Suggestions:

• Create early game danger that prevents exploration and "colony rush".
• Provide a means and an incentive for early game warfare. (With Transports being a mid-game tech, everybody essentially gets a free ride for a long time)
• AIs should try to (and conspire to) destroy you, not just harass.
• On higher difficulty levels, the AI should recognize that i'm the human player and work together to beat me.
• Provide some problem that makes "just sitting there building my empire even bigger" difficult.
• When AIs attack, it should be sudden, concentrated, hard to see coming, and with the intention of crippling me or taking territory. Knocking out shipyards, doom fleets, and taking my homeworld should be the enemy's Opening Move / First Strike.

If TV tropes is to be believed, GC2's AI was bloody brilliant. As for me? It's been so long since I've played GC2, and I don't think I've ever won a game, so I can't comment on it, but from reading several AARs, it certainly looked like the AI was smarter than average, and that's half the battle.

To me GC2's AI was pretty good, but for me, I'm always comparing strategy game AIs to Paradox titles, which seem to be more complicated and seemed much more challenging strategy-wise.

I'm playing the ultimate edition with the community update and I consider myself to be above average on this genre, not elite. That being said, I'm having a blast and the Ai is certainly doing well. I'm satisfied until I get a better comp to play 3. I would give it a 9/10.

@frogboy people are stupid. Did you see the raving reviews about endless legends when the Ai is as dumb as a doorknob and can't strategize properly or manage units at all? I don't bat an eye at any 4x games other than this and legendary heroes, they're my favorites and you've made a happy gamer out of me.

People who complain about it don't get that it is heads and shoulders above all other 4x games out there. You need to compare it to what's out there not by your own standards.people will complain about it being exploitable, occasionally weird etc but theyre missing the point. It beats everything else out there and if they want to whine, ask them of they'd replace it with the endless legend Ai and they'll be kissing your programming fingers in no time.

What the hell are you on about. He asked to compare it to other 4x Ai on the market not your personal dreams of super Ai games.

The AI in GC2 was very good.  The only AI I consider better is Civ4BTS Better AI. That provided a first class playing experience for a wide range of skill levels.  It is the response to the range of skill levels that impresses me.  In GC2, that seemed to come from Brads avoidance of scripted behavior, which was revolutionary for its time and still not truly duplicated.  There are some justifiably high expectations of any AI that will claim to be a worthwhile successor.  Utilizing data mining is a wonderful technology to add to the AI. It may be just as revolutionary.

But I had the advantage of experiencing GC2 as a complete developed package, Ultimate Edition.  That saved me about two years of waiting for the AI to mature to the state I found it in.  This time, I am in at the start.  I fully expect this to take two years to mature to Brad's ambitions.  Given the nature of AI work, it is likely to last longer than that.  It is a field rife with unexpected and exponentially increasing complications.  If you really want to experence the GC3 AI, hop in a time machine and go forward a while.  Meanwhile, I am waiting patiently because time machines make me nervous.

What I noticed is that the AI in GC2 seemed aware of what I was doing and reacting to it.  As I learned to play differently, it played differently as well.  It worked to fulfill the role of engaging opponent.  You could beat it.  I got up to crippling, which impressed me.  But it could easily be adjusted to where it gave a challenge but didn't run over you.  That was a function of both the map settings and the flexible AI.  For me, that was a big appeal factor in GalCiv.

There were many AI artifact behaviors as mentioned above, but the overall effect was engaging, challenging, and could actually teach you how to play if you emulated a bit of its behavior.  I am hoping/expecting that the stubborn artifact behaviors will be continually stepped on. Some of what I hear being called "brain dead" comes from these behaviors.  The behavior of sending one transport at a time is one good example.  It was jarring to have this otherwise very organic behavior mechanism go into effectively a programming loop.  It was like the perfect robots in West World suddenly developing mechanical behaviors.  Otherwise, I dismiss "brain dead" as typical hyper gamer speak.  It is never a bad thing, it is always the worst catastrophe since Star Trek got cancelled, maybe worse!

I am hoping a lot of effort goes into expressing personality through AI behavior.  I note the priority list we give our custom factions.  This is the kind of thing that makes a lot of sense to me for personality, but without looking at stats, it is not very obvious.  I really want a wide range of flavor text and interaction, and visible signs that a race is xenophobic or cowardly or whatever.  I think the main goal is not to make the AI clever, but convincing as a fun opponent, or crowd of opponents in this case.

This, a thousand times this. To me, a fun, immersive AI is better than a blood-thirsty, perfectly optimal AI.

I play a lot of strategy games, mostly 4X, some not (MOO2, GalCiv II/III, Civ 2/3/4/5, AoW 2/3, Hearts of Iron III, et al.). On every forum I've visited, for every game I have played, I see the same comments: "This AI sucks, it's total crap." Sometimes with "and Game X did it better." Go to Game X's forum and you see "Game Y does it better."

It seems to me that this is often a combination of two factors: a vast overestimation of the capabilities of modern strategy game AI, and an inherent disagreement with the purpose of any particular game's AI - whether it should be only out to "win", or be an immersive AI with character; whether it should be predictable, random, optimal, etc. And when it comes to challenging a player, many expect it to be able to play at human-levels of competence.

Disregarding the sheer improbability of that occurring in a game (the first person to code human-level AI will be ushering in an incredible revolution of technical development), some seem to forget the most basic fact, that "human-level" competence is already a widely varying attribute. Some people struggle with the easiest difficulty levels in games; if you could build AI to perfectly imitate those individuals, there would still be outcry of "how poor the AI is."

Anyway, in regards to the OP, I found GalCiv II to have quite decent AI, certainly comparable to other games, and often superior to them. The fact that Brad and Stardock continue to work on and support their games for years just ensures that the AI will continue to improve - and it already has a pretty good start.

Civ4 is probably pretty close to GC2, but I can't really say because I'm 4X stupid even if I like them.  I just don't stay focused on managing the shit out of things, so I ended up losing most of my noble/challenging games in both.

I thought it was good but it was AI you know, and there is just so much you can do. The AI will never make the complex decisions that we do, no matter how much you fiddle with them and a lot of what you do has an unintended consequences.

I have always talked up galciv2 AI. I told all of my friends that it has the best AI in any 4x game I have played. I fully expect galciv3 to be as great. Sid Meier's Civilization has been my favorite gaming franchise since 1991, I would love for you guys to make me bump GalCiv out of number 2. I think that GalCiv3 just might do that, give me characters that make me want to burn their homeworld.

it was more challenging than any other civ-like game I had played

i don't know how much of that is because of cheats or because of intelligence.

but from what I recall, it needed pretty heavy bonuses to stay challenging. RTS games didn't need to cheat anywhere near as much

galciv2 had much more customization than regular civ games, so humans still had tons of interesting choices to make even when the AI cheated

in regular civ, the human doesn't have much fun when the AI cheats too much

The AI was good, but it had holes which were found out eventually, which is inevitable when AI runs off of set parameters and doesn't account for certain human behaviors/exploits.

GalCiv II AI is top-tier in the genre.

The next step in AI is AI that takes player into account and adjusts from game to game at times to keep things fresh.   I really wish you had implemented this in GalCiv III.

I'm going to agree with MottiKhan that a suicidal GC2 game without using any of the known exploits is impossible to beat on most settings, although this is due to the brutal bonuses the AI gets (+100%) and not because it's so smart. 

Still, it's better than any other 4X game I know, most probably because there's actually 4 main AIP that do things quite differently in certain regions. And I still have some hopes that you might get back to it one day and close some of the loopholes or weaknesses mentioned in the Community Update thread to make it perfect. It could even reflect or inspire your current work here, because the GC2 AI is superior in some regions than the GC3 AI (eg. he doesn't trade Colony Ships)

With some tweaking it might become impossible to beat it on obscene as well (he gets +50% increase to most racial stats here)

The GC3 AI already seems a bit more feisty and well-rounded than the GC2 AI was. It will need to be taught the tricks that players develop over time, but already i note that it fights back in ways the GC2 AIs never did. Good job on that so far.

The issue is the definition of "cheese."  One man's cheese is another man's bread and butter strategy.

My take is that EVERY strategy is cheesy, in which case the person who finds the best one wins.  There are no rules in love or war.  If you're playing PvP, are you going to hold back? Or are you going to pump out the cheesiest strategies ever, and crush your opponent like a fly and make them cry as they die? (it even rhymes.  AHAHAHAHAHAHA).  

I'm going to address some stuff you mentioned.

War was mostly the result of a really good and complex diplomatically interplay. When it appeared randomly it might have been the Drath with their SA. Even war without being able to reach an opponent can do certain things, War Profit, blocking trade, forcing some AIP to dive into specified research.

still the invasion will do more damage to you because an invasive force gets a bonus to attack and the AI was very keen to use invasion techniques. naturally you'll loose more pop than him resulting in a loss of economy & influence. especially considering that AI planets didn't build too many farms and were capping to the popcap fast (at that point it's always a good idea to load troops into a transports otherwise you'll loose pop growth potential)

Some of this is correct   although if an AI had sufficient MMR it was next to impossible to make someone declare war on him, and playing at a high diff level usually remedied that to your loss.

Well, there pirates in that seem to have the same function like monsters in FreeOrion, but yeah, perhaps one AI that is kinda totally destructive and doesn't care much about winning the map instead simply trying to put some harm on his neighbours.

That's a good point. As much as i hated that SA, those were probably the more interesting games. Games without the Drath were noticeably placid.