SPECIFIC Suggestions for Improved AI Aggressiveness

Precursor10

Join Date 04/2006

SPECIFIC Suggestions for Improved AI Aggressiveness

Okay, I was spazzing out a little earlier (https://forums.galciv2.com/?ForumID=346&AID=116644). Let's try to be a bit more constructive this time around.

It ticks me off when I invest a ton of time in building an empire, only to find that the AI is a virtual cakewalk to conquer when I finally decide to set-out and release my elite fleets. 4 games in a rows I have abandoned because of this and, yes, on the higher and highest difficulty settings. I have literally conquered a 23-planet, 130-ship Drengin empire...losing only 3 (1-2-3) of my own ships (not FLEETS...but SHIPS). Obviously, this shouldn't happen.

This may come off as seeming a bit preachy, with a know-it-all tone, but please keep in mind that I do respect what you've done here. I don't know squat about computer programming but I do know a good game when I play one. I just want to see the game to challenge folks (me specifically) more and realize its true potential as an addictive masterpiece. But, respectfully, the AI has considerable room for improvement guys and you need to address it.

Attn: Developers - SOME SUGGESTIONS

1. The AI's primary function at the higher game levels has to be to DEFEAT of the human player at all costs without cheating. It should not focus on battling other AI civs, nor ignore that element completely. But the first 'realization' the AI needs to make is that the human player is the greatest threat. A 200% economy for the AI shouldn't be the primary means of balancing things out in the computer's favor. It doesn't work anyways. Trust me. The AI economies can be run with efficiency that a human player could never achieve over 1000 turns. It knows every civ stat, every resource location and every planet's profile. That should be enough in terms of 'tools' to challenge a human player.

2. For AI at war with the human player, the PRIORITIES need to be dramatically altered as follows..from highest-to-lowest:

A. Attack undefended or weak resource starbases. This can cause civ-wide chaos for the human player...especially those economy resources. It will also force the human player to stretch his or her available defense force a little thinner...to more uncomfortable levels. How many of you have set-up resource mining in the heart of your enemy's area of influence? This is stupid and should not occur so easily. Doing so should cause the AI affected to dramatically increase its aggressiveness towards you, depending on the specific resource involved.

B. Attack the human player's economy starbases, especially during the the early-mid game. Another big-impact type of play. Worthy of some AI kamikazee? I think so.

C. Undefended human planets need to be RUSHED. The AI attack should always involve a troop transport following up within 5 turns of the orbital attack. Otherwise, the starship attack/battle for supremacy of the planet's orbit has no STRATEGIC purpose and the AI's resources should not be wasted accordingly.

D. After 'undefended' planets, planets hosting the human player's Galactic Achievements/Capitals/Trade Goods should be ranked next, provided the defense is exploitable. Use a simple ratio to determine. Take my EYES OF THE UNIVERSE or MANUFACTURING CAPITAL or GRAVITY ACCELERATORS and I will curse the monitor! But I'll also be savoring the thought of some revenge!!!

E. If the human player is relatively well-defended but has concentrated trade with one civ...and THAT civ is weak...THOSE planets should be conquered instead. Chasing a human player's mini-freighters around is stupid because a new one will simply respawn where the route originates. It's idiotic to watch the AI do that and an insult to your hard work and a really cool gaming concept.

F. Beyond those STRATEGIC targets, the AI should seek out TACTICAL victories during war where it can. That means locating, engaging and destroying weaker human fleets. AI fleet logistics should focus on this kind of tactical exploitation.

3. Otherwise AI fleets should flee superior-sized and powered human fleets. AI fleets should retreat and link up with other ships or fleets to man-up! If the AI can't possibly construct a fleet that can challenge a human fleet (due to an inferior logistics point cap for example), then AI fleets should become as big as possible and congregate together. The human fleet might be able to defeat the first AI fleet, but would the human even attack if the AI had 2 other fleets waiting and in range to counter on the very next turn? Maybe. Maybe not.

4. If the human player has Tech Victory enabled, research-advantaged civs should pursue it with a dedicated focus (if any are involved in the game that is). That puts the human player on a virual deadline and raises the in-game tension. The human player knows he's on-the-clock to go wipe those guys out before the final TECH VICTORY tech has been discovered. Again, defeat the human player at all costs (without cheating).

5. Likewise, if INFLUENCE VICTORY has been enabled, I expect to see at least one of the non-war-mongering civs doing their best to build and upgrade I-Starbases. AND aligning themselves with other militarily strong AI Civs via an Alliance.

Man...I can go on an on...

6. AI needs to rush-build GRAVITY ACCELERATORS, TIR-QUAN TRAINING, EYES OF THE UNIVERSE, GALACTIC PRIVATEER, XINATHIUM HULL PLATING...and deny the human player from gaining those if it can. Watching an AI civ build DIPLOMATIC TRANSLATORS or worse, a GALACTIC GUIDE BOOK just makes me roll my eyes in disgust.

7. AI needs to diversify their fleets with ALL 3 weapon classes, instead of going deeper down just one line of research. It's too easy for the human player to counter. AI defense research strategy should react to the weapon class of the human player.

8. AI needs to take more initiative to land those ATTACK bonus. Can't stress that enough really.

I know it sounds like there's little I like about the game as-is but please don't misunderstand. I'm motivated to type all of this out because I actually care about the game and I want to see it kick all @ss. Massive potential. The core model is a winner.

Fix it before a competitor does.

Peace out.

Locked Post

Advanced Search

Watch this post

Do not email me updates for this post Email me updates for this post

Successfully updated karma reason!

Alright, it's awesome that you spent a lot of time and took effort to not be offensive, but you have to realize a few things:
1) AI is EXTREMELY difficult to make. When you say "I want it to ID undefended enemy resources and knock them out in an intelligent manner", you are asking for an extremely difficult thing. Humans can do it easily, but that's because we eiher had several million years to develop our own AI, or we were crafted by a perfect being, depending.
2) Some of your suggestions simply go against the grain of the theory of the AI in the game. Here's two key things:
a) The AI is supposed to treat you like another AI. As far as the code controlling the opponenets (the AI) is concerned, every other civ is equal - it literally cannot tell which player is human. And that's the way the game was designed, and they are going to keep it that way. I might as well ask for a mixer mode, so I can make music CDs from home. Its simply NOT what the game was designed to do, and they aren't going to change it.

The AI is not out to "win". The AI, just as it was designed to not tell the difference between you and other AIs, is not out to "win the game". I don't think any president ever sat in the White House, hoping to "win" by conquering the world. Likewise, the other civs in GalCiv2 are supposed to simply do there thing. Torians spread their hippy love, Drengin feed themselves, and the Yor enjoy torturing folks. If they happen to "win" because of that, more power to them.

And really, when it comes down to it, if you really can crush the AI even on the hardest difficulty level, there's not much an AI, or most people, can do much against you. I would suggest fiddling with map settings - some settings are much harder than others.

As for the reviewers who said that this game does have one of the best AIs, they say that because it's true. Compared to games of equal depth, the AI in GalCiv2 makes the least stupid mistakes. I know that sounds pessimistic, but if you find a game with better AI that doesn't cheat until beyond "tough", email me.

Successfully updated karma reason!

And really, when it comes down to it, if you really can crush the AI even on the hardest difficulty level, there's not much an AI, or most people, can do much against you. I would suggest fiddling with map settings - some settings are much harder than others.

Everyone seems to be consistently missing the point. No disrespect intended. I wish I was this brilliant at computer strategy, but I'm not.

ON HUGE MAPS...THE AI DOESN'T ATTACK...EVEN IF I HAVE ZERO MILITARY...REGARDLESS OF THE DIFFICULTY SETTING.

EVEN ON CRIPPLING OR HIGHER, EVENTUALLY I CATCH-UP IN TECH AND CASH...SO I ATTACK THEM...THEY DO NOTHING. THIS IS BORING.

Sorry about the all-caps but, once again, it's an AI aggression issue primarily. The computer doesn't know what to do with a human attacker or opponent.

I am hopeful that ver. 1.2 somehow addresses this. If not, I'll check out ver. 2.7 or 4.3 or whenever I start seeing in the forums that the situation has changed. In the meantime, I'll be playing me some God of War of Battlefield 2 cause I feel like killing me some sh!t. Kratos rules.

Successfully updated karma reason!

I'm not pretending to be an AI expert here, I'm just developing a piece software with somewhat similar real-time decision making and these are my educated(?) guesses of what may be going on...

It's difficult to develop an AI that's good in all situations, i.e. that can generalize its skills from the cases it was developed and tested on to the random galaxies of all the zillion galciv2 players. But it's the only way to prevent massive loopholes and cheese to even some degree and provide some illusion of intelligence and strategy skills. If you just hard code some "expert" rules (do A if B and C etc.), you end up with very predictable, machine-like AI that works only in very specific circumstances and is easy to trap with some specific tactics.

Even a war-optimized AI will have some loopholes for sure and you'll win it every time when you discover its weakness, i.e. the simplistic rules it's based on. There is no free lunch, no silver bullet. Trying to develop a more general model is the only way to go even if it's futile with the resources available.

Frogboy, your posts indicate that you have been focusing on generality, not special cases and it shows. The AI feels more varied and fun to me than maybe any other TBS AI I have tried!

Maybe it could beat me easier if was optimized with only that in mind but I wouldn't like to be forced to "build the the same kind of perfect fleet ASAP in every single game" all the time. If I want challenge, I try to beat the game with weird ability settings, find new ways to win&lose etc. I like to create a kind of RPG-like story of each game, not necessarily repeatedly hone my skills of rushing the AI in some specific circumstances. That's what fun to me. YMMV.

Of course, there are bugs and weird things happening. Take the "AI builds way too much embassies"-bug for example. My wild guess: it would make no sense to fix the bug by just limiting the number of embassies the AI can build in a single planet. You would never learn what the AI was doing wrong in the first place. It's better to find out the root cause and fix the AI's general way of thought instead of duct-taping the problem away. It's still good to complain "please fix the annoying embassy bug, I hate it" so that the developer knows what is going wrong from the players' point of view. I'd like to see the bugs fixed, too. Just please don't forget to say please! Stardock devs are remarkably open about the product they are developing and making themselver vulnerable to all sorts of flaming and abuse. I don't want their dev journals and quick replies to these forums to stop. (BTW, the memory management and AI related journals have been great.)

Some of Precursor10's suggestions sound good to me; from "use case", not "force this rule to AI" point of view. There are areas where the AI plays too dumb. I guess the devs know about them because they seem to read these forums actively and can't possibly be so stupid to not spot the glitches in the game itself.

It just takes time and is really not trivial.

Oh, and "AI:s should gang up against the human player" is a very very bad idea, IMHO.

Successfully updated karma reason!

1. The AI's primary function at the higher game levels has to be to DEFEAT of the human player at all costs without cheating.

No.
The "everyone vs the player" syndrome is precisely one of the most annoying trick that AI use in countless games, and it's one of the most irritating feature.
I like the AI precisely when they DO NOT differenciate between other AI and human.

I'm happy to hear from a dev that it's "not gonna happen"

Successfully updated karma reason!

I posted some simple, easy to implement suggestions that might make the AI much more interesting. Have a bit more unpredictability in the AI. They can be attached to certain races to make it more predictable for the human, and still enrich the human experience.

Examples: have some AIs go for a military rush strategy right out of the gate. Be bold and don't worry if the enemy seems to have double their military. Superior ships will take out superior numbers, especially early, especially if one player plans to do this right from the start. Depending on map size have them build a very limited number of colony ships (0 for tiny, scale up to maybe 8 for gigantic), beeline for military techs and planetary invasion and go for it. Have them build constructor fuel starbases to find someone to fight if there is no one nearby. Again, maybe limit this to one or two races and don't always have them do it, maybe 66% of games they are in. Unpredictability keeps everyone off balance. Occasional early wars amongst the AI or human vs. AI will spice up the game.

Have one or two AIs build fast sleek armed ships, early. Again, this can be part of the personality. Also include at least one engine on ships for this race from day one, and then maybe a 70% random chance they increase the number of engines by 1, a 25% chance they increase by 2 engines, with each major new engine level (Impulse, Warp, Hyperwarp). Again, not always, but often, a player can expect to see fast armed military ships from one or two of the races.

Have some AIs tend to stack transports with their military ships. One transport for each battle fleet. Again, not always, but often, roll a die to see if the AI wants to do this for a year or two. One reason the AI does so poorly at war is that their transports are poorly defended.

These are three specific suggestions that will make the AI much better, much more interesting, without breaking the bank as far as testing and development, without making them gang up on human players.

Successfully updated karma reason!

Here are a couple more, have one or more AIs make a priority of resources over habitable planets. Again, perhaps have them build a set number of colony ships depending on map size and then go all out on constructors to find all available resources, without first seeing them. The minor races do this on occasion.

Again, don't always have a certain race or two do this, but maybe in 66% of their games. Figure out which races would naturally prefer each of these preferences and then add that to their AI. The specifics of coding each of these may already be in the huge code ready to be turned on.

Successfully updated karma reason!

I've come to realize over the last couple of days that GalCiv2 IS what it IS. It has a lot of great stuff and a ton of potential - no doubt folks.

But the AI is incomplete.

-building ships with 0, 0, 0 weapons and 0, 0, 0 defense ratings?
-building capitals and other non-producing structures on bonus tiles?
-8 embassies on a planet?
-failing to attack a defenseless civilization under any circumstances?
-focusing on chasing freighters around the map during times of war?
-troop tansports running around unescorted with 1 HP hulls?

These are facts, not opinions. This game is INCOMPLETE. Sorry Frogman, I am not trying to offend you. The fact that you roam and participate in these forums is one of the damn-coolest things I've seen as a gamer. I really can't give you enough credit for that. Your teams' dedication to the community is also not subject to debate.

But you shipped this thing in a box, with a glossy cover and collect cash for it. The fact is, this is still a beta version game. And I know, despite all of your hard work (not doubting your talent either for a minute) that when you put your head on your pillow at night, you know this.

And mark my words...WHEN you get it right, I'll be your number one fan man.

Peace out.

Successfully updated karma reason!

If you dont understand how even basic algorithms work, then you should be very careful about making inflamatory posts about said subject. Criticisms are fine, but shooting your mouth off when you dont know what you are talking about is a really good way of appearing to be an idiot.

In any case, one valuable thing to do is to compare the GC2 AI to the AI in other 4x games. When you do this, you will find that your statements about the AI being in the beta stage are completely false and make you look rather silly.

Specifically addressing your last points:

I have never seen the computer building large numbers of 0,0,0 ships except as scouts (which is a viable strategy). Also, keep in mind that building those ships as non scouts can be a viable strategy because its faster to build the hulls and upgrade when you get higher techs than to build the high tech ships when you get them. FOr those reasons, Im going to go ahead and declare this as not an issue. I think you are confusing scouts with fighting ships.

Unescorted transports were a big issue with older versions of the game, but nowadays it simply isnt. The computer usually escorts them. Some unescorted trans arent bad: its worth the risk on occaision, and human players lose them sometimes too. This could use some improving, but to call the game AI poor because of this is silly.

I cant address the rest of your points because I cant get any posts to load due to the horribleness of this forum.

However, the bottom line is:

1) Dont make strong, idiotic statements when you have no clue. To criticise is perfectly valid. To say that a game is still in "beta" for a bunch of issues that it does better than any existing similar games and cannot be fixed with any real world applicible solution makes you an idiot.

2) Many, if not all, of your ideas are either attempting to address problems which only exist in your head, or would require a solution that is not possible while still keeping the game playable.

Successfully updated karma reason!

Precursor, I think you need to think before you reply to the above, because it looks to be on the money.

Successfully updated karma reason!

@ Krazicarl + Jasamcarl (geez those names look similar. I'm shocked you share the same opinion).

It's funny how, when I post a criticism, I get slandered with terms like "idiot". Krazi, anytime you wanna compare degrees, just send me an e-mail buddy. I'm sure you have access to it. In the meantime, don't patronize me.

I also love how you tools consistently ignore the positive things I say in each post.

Look, I know these forums are flooded with Stardock guys and there's a loyal following here. Congratulations.
But how about pulling your collective heads out of your collective asses and scan the forums. I'm not the only one who's raising these issues. It's a horde out there fellas!

But you believe what you want to believe. Call it a difference of opinion. All I know is, I made some SUGGESTIONS and they got shot down. Cool. But if that's not enough,...if you want me to repent and admit my blasphemy well then I will simply say "blow me". The AI has serious issues. That's a fact.

I'll stop posting about it (which until now, I've limited to almost entirely this thread). But if you guys want to keep this discussion running, I'll keep coming back. I won't leave the topic with some fool, calling me an "idiot".

Successfully updated karma reason!

Uh. it's not that you have issues. It's not that you find the A.I. easy, the game not challenging or fun...

It's that you constantly insinuate that either the developers don't know what they are doing or that the game is an abject failure because it doesn't live up to your literally impossible standards for AIs in complex strategy games.

You could have posted your negative views of the game and that would have been fine. But you have constantly gone beyond that with ignorant condescention to those who enjoy the game at its current level of difficulty or who appreciate how hard it is to develop a very good A.I.

I am not a fanboy. I tend to hate fanboys; i know this games limitations and will probably cease playing in part because of them. But I have enough empathy to realize the constraints the developers are working under and enough maturity to not make everything into a personal persecution complex.

You share one notable featuree with the fanboy; notably a childish sense of entitlement.

Successfully updated karma reason!

I really despair (well a little anyway) at how a thread can soon become personal rather than focussing on the points made by the OP in his first post which was both well constructed and thought out.
If the OP made a misjudgement at all it was a lack (no offence to the OP) of understanding of the depth of feeling and passion shared by some people when they feel their game is being unfairly criticised, even if that criticism is not unfair in its content but it is not sugar coated to the degree needed when raising the issues that has been raised.

I had never played GC I or any other 4X game before and so there was no history' or allegiance to anything when I bought GCII - I hadn't found these forums until I had been playing the game for a little while. When I had played the game and understood, to a reasonable level what was going on, I was amazed at just how many irritating bugs and seemingly irrational choices the AI was making. I assumed it was because I started playing at the 'Normal' level of gameplay. Whilst I have now played a handful of games and now need to think beyond the 'Tough' level I am still concerned over the bugs, although many of them have been resolved through the patches, but I am still concerned that the AI makes some choices that make me cringe.

For me this game is very much a 'rough diamond' - it is very addictive and the responsiveness and attitude of the devs is nothing short of superb in the dedication that they show towards the people who play it. However it does, like any other game have its short comings. People like the OP should, in my opinion, be welcomed by any forum as long as the points that are made are objectively critical of the game with the underlying intention of trying to make it better. IMO that is what the OP tried to achieve.

The OP needs to consider that whilst the initial post was, imo, to the point and well constructed it also lacked a little 'sugar coating' for some people to accept without goading you to make the thread a little personal, all that does is detract from your objective. No matter how right you feel that you are being right is often not being popular. Just ignore the snipes and gripes and keep focussed on what you set out to do. Your first post made a big impact and a lot of sense however your follow-up posts needed to be thought over a little more.

Please people have a think before you post, this is a great friendly forum, lets not get personal

Successfully updated karma reason!

@ vimes2004

Nicely said and so much sensed.... and needed too !

Successfully updated karma reason!

I'll ignore that last sentence because it's not true. Please re-read my posts. Otherwise, I don't have a problem with anything you wrote here. Although I think that I have been seriously misinterpeted in some cases.

And don't doubt any of the claims I've made. Just ask and I'll provide screen shots. The 0/0/0, 0/0/0 ships were NOT scouts. First mission of the Dread Lords campaign. That is ALL the Drengin built. Sorry but it's true.

Successfully updated karma reason!

Acknowledged vimes.

Now that's an intelligent and balanced response. I appreciate what pieces of that you've extended my way and I can definitely respect the rest.

Thanks.

Successfully updated karma reason!

hooray for ambassador vimes!

Successfully updated karma reason!

Wait to see how 1.2 changes things. When the attacker no longer gets the massive advantage of that initial 'killer blow'. You might find things a little more difficult.

Successfully updated karma reason!

I too am looking forward to the ideas about the1.2 patch will bring to the game.
I am almost overwhelmed at the tenacity and drive which Frogboy and the team apply to their work in continuing to develop this epical masterpiece.
I would also like to believe that the OP should find the the future patches, as well as easier modding support, a positive step in the direction that meets with his / her approval.

I do hope that the OP continues to frequent the forums and contributes posts like the first one in this thread, with a little more sugar coating, as I am sure that the devs take on board constructive criticism and ideas as a catalyst to spur them on to make this game even better than it already is.

When they do have a little time to spare then please......

'Fix the Forums' - oops that is side tracking the thread a little, please indulge me

Successfully updated karma reason!

Uh, vimes. Have you read any of the replys. The OP didn't know what he is talking about. You don't know what you are talking about. It wasn't that he wasn't sugarcoating it, it is that he was plainly ignorant in his insinuations that some problems could be easily fixed and that the A.I. was unchallenging for anyone.

No one was questioning that the A.I. has faults. What they were questioning were his dumb suggestions and general ignorance.

Successfully updated karma reason!

Actually precursor. I never doubted that the observations of A.I. weakness were valid, but those were never the issue. Stop trying to set up strawmen.

You do have a dubious sense of entitlement. Otherwise you wouldn't have raised you own expectations of how well the A.I. should be doing and then label the game as 'in beta', despite the fact that it already performs better than every similar game on the market, atleast at initial release.

Successfully updated karma reason!

I dont really have a problem with the original post. It has many opinions in it which I believe are not particularily well reasoned, but people are welcome to their own opinion. I agree that people should be encouraged to post such opinions. My problem with precurser is statements like:
"The fact is, this is still a beta version game. And I know, despite all of your hard work (not doubting your talent either for a minute) that when you put your head on your pillow at night, you know this."
"That's too bad. Bought this game because Gamespot.com was raving about the AI qualities. Morons.
So in summary, your great STRATEGIC game only works on mini-map mode. No epics here. Gotcha. I'll spread the word."

These are statements that are not only inflammatory, but show a distant lack of thought from the poster. In particular, he makes strong claims while acknowledging that he knows nothing about programming or algorithms. Such statements should be strongly discouraged as they are neither useful nor constructive. He also claims to be a big fan of similar games, but is unable to find a similar game which has a superior AI despite my repeated encouragements to actually support his arguments using evidence.

The original poster also shows a distinct lack of maturity by saying things like:
"Krazi, anytime you wanna compare degrees, just send me an e-mail buddy. I'm sure you have access to it. In the meantime, don't patronize me."
and
"Personally, I play GalCiv2 because my 6-figure salary ..."

The original poster likes to claim that he has simply offered suggestions and they are getting unreasonably shot down. This is not how things happened if you read above. When it was pointed out how many of his suggestions either did not fit with the developers vision of the game or were not possible (including a post by the AI developer himself!), he proceded to call groups of people morons and attacked the developers of the game by suggesting that they should not be able to sleep at night. The poster's ideas were not "shot down." They were unreasonable, as was explained to him.

Finally, I did not call your actions idiotic because they were criticisms. I called them idiotic because you made them in an inflammatory mannor while admitting that you had no idea of how to program or construct algorithms and without comparing the game to other similar games (once again, I will point out that the "good" in "good AI" is a relative term). Such actions are undeniably immature and idiotic. Its like criticising the fact that physicists havent developed a warp drive based soley on watching Star Trek a lot. Your comments differ from my slightly inflammatory comments in that I make sure I support my argument adaquately when I say what I say.

Of course, I am not suggesting that you have to be a programmer to make criticisms/suggestions, but you have to control yourself unless you really know what you are talking about. The original poster has failed to do this in nearly every post besides his initial one. To return to the Star Trek analogy, it is reasonable to ask:

"Hey physicist, why havent guys made a warp drive yet? That would be really useful."

Its not reasonable to say:

"Hey physicist, why havent guys made a warp drive yet? That would be really useful."
"Its hard to do for a variety of physical reasons, including Einsteins General Relativity"
"And they said you guys were smart. Morons. I'll make sure to spread the word about how dumb you are."

Successfully updated karma reason!

I'm 100% sure for most people the AI of GCivII is aggressive enough as it is, yet not as aggressive as some of the posts here.

Successfully updated karma reason!

Krazikarl:

In not one instance above, did you quote me in proper context. Typical. You pseudo intellectuals are so busy analising and scrutinizing one little string of words that you can't see the big picture.

FACT: In this thread, I have not once launched the first personal attack. I have only responded in like kind. That reflects on you, not me.

FACT: You have still failed to note even one of the many positive comments I have made about the game.

karl, you are Krazi. Just cause you type a bunch of sh!t, 10 paragraphs long doesn't make it true. We essentially had closure here until you had to open up your big mouth again. Congratulations instigator. Now I'm pissed off again. Personally, I thrive on conflict so if you want to make this thread 10 pages long,...then it will be 10 pages long.

"Precursor10 versus the K/Carls". Someone phone MGM, we got a quality script here!

Successfully updated karma reason!

I agree with the OP here..the AI needs to be more agressive. Sure, some of his suggestions aren't possible or would make the game unplayable or cause performance issues, but I bet he's not the only one who abandons games early just because conquering all the galaxy is a cakewalk.
When I first started playing, I got thrashed by the AI even on lower difficulty levels than Normal...but once I started playing better I win all of my games with relative ease. I usually see Ais fighting eachother, and some races dominate the battlefield and destroy a couple of races before I even start building a nice military. The AI needs some work, but it's definately good for a good majority of the players.
Mya biggest problem with the AI is the way they manage their planets. They play exactly like I played when I first got the game: I built once of each building in each planet, only specializing the planets that had a capital on them. It's common that I already have 10 fully terraformed planets with every tile built on, and each planet is specialized to do one task (production, money, research), but the AI only has 2 or 3 planets and still hasn't built improvements on half their tiles.
Part of this is because they don't build enough factories. Another part is because they research techs to improve their buildings (lets say...bank construction or research academies) before they have built all the improvements they can. That means they spend all their production upgrading their xeno labs to invention matrixes instead of building more xeno labs until they don't have any tiles left.
If only the AI was better at managing their planets, they would advance faster in the tech tree (thus getting better weapons), build more ships, and have more money to bribe other players than in the current build. That would be a big step up in making the AI more challenging.

Successfully updated karma reason!

The total war series are a good example of an AI ganging up on you, the human player. It renders any diplomacy etc in the game pointless, and means that you are basically having to expoect to fight everybody at once. I don't think that would work in the Gal Civ world, especially not if you choose to go for a Tech or Diplomatic Victory.

And aren't the 0,0,0 stat ships there mainly to stop a transport rush to conquer the planet?

Successfully updated karma reason!

*SPECIFIC* Suggestions for Improved AI Aggressiveness

*SPECIFIC* Suggestions for Improved AI Aggressiveness

SPECIFIC Suggestions for Improved AI Aggressiveness

SPECIFIC Suggestions for Improved AI Aggressiveness