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!

No, you've just whine like a little girl when some note that your suggestions suck, you don't know what you are talking about, etc. And when a developer notes how bad your suggestions are, you go about condeming the game based on some utterly childish and unreasonable expectations aren't validated.

Then, when you realize you are loosing the argument, you try to shift the focus away from those embarrassing comments and recase yourself as the good guy who was simply trying to speak truth to power. Don't worry, you're not the only fool in this thread; note the couple of equally ignorant people who chime in to help you shift the focus of the debate.

Successfully updated karma reason!

Idiot. Now you've just stooped to my level.

You've lost all credibility now...according to your own criteria. Allow my to stoop a little lower for a sec because I just love watching you unravel.

C.A.R.L (Craves Anal Ravaging Love). Ya, that's you all right from the sounds of it.

Successfully updated karma reason!

it's easy to shut down a dialogue when one can appeal to misunderstanding in order to keep a power dynamic.

recovering power by sacrificing grace doesn't contribute to a mutual understanding.

disavowing agency of the effect one has when communicating or stating ideas is dishonest. or naive.

i have to admit it is hilarious to read the hysterics and melodrama. precursors performance during his noble monologue where he says, "...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." had me glued to the screen.

His performance gave me goosebumps. It was like I knew EXACTLY how he felt.

And the saber rattling with "Jasamcarl" is great. I wonder what will happen next...

Will Precursor's noble quest to improve the AI fall on deaf ears--not only to the devs, but to the court of public opinion?

Can Jasamcarl prove victorious as the sole voice of reason?

Has KraziKarl's appeal for mutual understanding been seen as a veiled attack against precursors credibility?

could vimes2004 be setting us all up for a thread hijack?

we will see...

Successfully updated karma reason!

onowak. I just have on thing to say to that.........i like you.

Successfully updated karma reason!

#54 by Jasamcarl
Friday, May 12, 2006 7:55 PM

Trolling
Insightful

onowak. I just have on thing to say to that.........i like you.

See? Now that's a great way to end this miserable thread.

But I'm sure some ass will come along and mess it up.

Er...Full Retraction there.

Successfully updated karma reason!

Franco fx

Join Date 04/2003

+59

Amazing thread. I don't doubt the points that P10 has stressed (0 attack 0 defense, AI will not attack, etc. ) but I can say that I have not seen this in the 30-40 games I have played.

I have seen the AI build zero attack and 2-3 defense ships in some of the versions that have been released but I haven't seen that recently. In my games I try to avoid being attacked by the AI and I have not found that to be very easy when playing a full array of AI. All it takes to get the AI agressive in my games is to act counter to their ethical allignment, let my military strength go south and not research diplomatic techs.

I have not found it that difficult to win any type of victory up to tough difficulty. The one win that I have at painful was well, painful and too demanding to be enjoyable to me. For those that want a more challenging game at the no AI cheat level what is wrong with letting the AI get a bit ahead and do the things that piss them off.

In other words, I do not believe that in game after game you are confronted with a stunted AI that refuses to attack unless you have played the game in a way that make them play that way. Things like maintaining the strongest military, exploiting every possible mining resource and pressing with influence and diplomacy.

Try this: After you have gotten your share of planets and made whatever ethical decisions , let your military go fallow, set your spending to zero and click the turn button about 40-50 turns, then come and tell us how the AI behaved.

Successfully updated karma reason!

The problem is we don't want to handicap ourselves only to have an AI that acts correctly. We want the AI to challenge us without needing a 200% economic boost, or 50 turns to build it's military.
I was playing a game 5 minutes ago: Gigantic, 9 AIs on tough. I was doing my stuff in peace when the drenging, the terran and the torian attack me all at once. I had the time to research missiles all the way to harpoon, get warp drive, and build a couple of ships before ever seeing an enemy ship (I also have eyes of the universe, so I can see them coming from afar). And what kind of ships does the enemy throw at me? Fighters with 6hp, 1 attack (mass drivers) and 2 moves/week. And unescorted tranports...
All this proves the AI gets weaker as you increase the map size. In small or medium games I never saw this happen, as the AI usually has fleets of several high attack ships by the time I get a decent army, so it's not a cakewalk like in a gigantic game.

Successfully updated karma reason!

Actually, pre. You're still an idiot.

Which of my criteria was I breaking. I was not the one who made this thread irrationally personal.

And besides, unlike your snide insinuations, you've provided all the proof I need to call you out.

Successfully updated karma reason!

Actually Jasam, I'm not. I'd give it another shot at trying to convince you but I don't have the necessary training for handling people with 'special' needs as yourself. Nor does it interest me to try.

So J-ass-man (I typed it that way because you're probably dyslexic also), I'll just leave you with your anger. Anger at knowing that someone capable has challenged your laughable dominance and comfort level on these boards, and knowing that yet again...

YOU"VE BEEN OWNED BITCH!

Successfully updated karma reason!

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.

The AI actually does it, if it is at war. You seem to want the AI to do it, whenever they have a chance to do it, without taking into account whether they want to go to war.

[Quote]
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.

I think you are assuming that influence boundaries = boundaries. In theory it isn't if you read the manual, BUT if you look at some of UN resolutions, some of them do implictly assume this

My problem mainly is that the AI doesn't seem to value grabbing resources. The way i play, early on, I don't grab resources either within 'my area' since colony rush is more important, but once i see a scout in the area and i know the opponent knows about the resource, i quickly send a contructor to grab it.

[Quote]
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.

I find it pretty amusing that the way the AI plays early on is pretty defensive in that it tries to stop troop rushes early in the game, but hardly anyone plays like that I think.

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!!!

You are assuming the AI has tons of spying on you. Eyes of the universe don't do the Ai much good, but i suppose it can piss off the human player... In most cases, the best colonies with the most stuff are the most heavily defended anyway.

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.

Too general. Even a human player would say huh??

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.

Ah tactical AI. if you take this, then i take that, and you take this then i win by taking this hence you won't take this...
I think in a posting the developer explained why the AI was weak in this area, because it was computationally expensive. It's not like Chess Ai you know, where all the computer needs to do is to render the board pieces and keep track of the rules.

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.

Nothing wrong with diplomatic translators, and they love to built Xinathium hull plating quite a bit (compared to others) i notice. They benefit very little from eyes of the universe, because using all that info is computionally expenseive. I think Galactic privateer sucks though.

Agree about tir-quan training and less so about gravity acceletors. I just wish they put more focus on soldering techs, but all this is moddable.

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.

AI defense DOES react to the weapon class. But I think again, limited computional resources mean they redesign a lot less then optimal.

Personally, I find nothing wrong with going deeper down just one line of research.....

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

Fixed in 1.2. No more first attack bonus.

Successfully updated karma reason!

It's amazing what you can find with Google. For anyone reading this discussion and giving that prick Jasamcarl some benefit of the doubt, let me fill you in a little. Besides being a massive D&D dork, Petitioner to the Cartoon Network, Crusading critic of Halo 2, and Super Hero Role Player, he's an all around offense to the Internet...

Look how he provoked this guy:

RE: Welcome to HNN, Jas (#20197)
by NYGuy on October 14, 2003 at 8:54 PM
Jasamcarl

What the hell are you talking about? I asked you to provide evidence that the Bush administration showed China and Southeast Asia the way in terms of their domestic security policy. Instead you launch into the ridiculous, hyper-generalized rant about the new ideas that Bush&Co. have. Nevermind that Asian policy is grounded in a specifically Asian context.

NYGuy

Ignorance is not a defense. Certainly your meaningless comment that “Nevermind that Asian policy is grounded in a specifically Asian context” betrays your inability to engage in meaningful debate, and only shows your ability to say nothing but in your mind you are thinking you said something brilliant. To repeat your comment, “What the hell are you talking about?”

With 10 little words you think you have expressed an idea, even though these 10 words lacks substance. But with your arrogance you can feel superiority. On the other hand you claim no one understands your political philosophy. Perhaps it is your inability to achieve the type of simple communication for which Ernie Pyle was so famous.

Evidently living in a cloistered atmosphere you don’t come in contact with the real world and with world business leaders who interact with global problems each day and must make serious decisions each day. GW, Rumsfeld and Cheney have had that responsibility, and even if you can not understand their public sector leadership positions in creating jobs, growing revenues for their companies, creating shareholder value, interacting with world governments and with the dynamic trends in the world economic scene, you say in your smug way, “I know more then they do even if I know, and say nothing.” You just prove yourself to be a self-centered liberal, not an unpatriotic American.

Ignorance is bliss and creates special privileges for people such as you. Since you plead ignorance in anything I say, you will always remain in that state because you lack the experience needed to comprehend what I am saying. Ignorance does not make you right, particularly since you are afraid to tell us what your objections are beyond ignorance.

Oh there's sooo much more....pages and pages of this guy being hated. Seems Jasamcarl can't make a point without being a condescending prick.

Jasamcarl, you are dismissed brother. Shut the hell up or I'm bringing out the GOOD stuff. And take this guy's advise will ya? How about unplugging once and while. Go outside, get some fresh air. Talk to a girl maybe. You might like it. Get some perspective.

Successfully updated karma reason!

Precursor, you are looking like an idiot, with your latest posting. Best case scenario both of you Jasamcarl and precurosor are sad losers... I

regards
Third best loser.

Successfully updated karma reason!

@Richrf

Unfortunately, the subject of my last post was the first person to make things deeply PERSONAL around here. A brief amount of investigation work revealed his heinous MO. I prefer not to plunge to such depths but it's necessary. My genetics prevent me from ignoring such ignorance.

He's a cancer.

Successfully updated karma reason!

I simply cannot believe this thread has not yet been locked. Precursor drops a "Frogman," spells out CARL (I can't help it - that made me laugh), calls someone an "offense to the Internet" (which is a difficult feat, as I think we all know) - and yet it's still going!

Precursor...are you a witch?

And to think that the whole thing achieved closure a few posts ago when someone tried to be civil and polite and - what do you know? - that actually worked.

P.S.: Although I agree that the AI shouldn't treat the human player differently, I think most players would agree that it could be smarter in the way it waged war - which was Precursor's original point, whether or not you agreed with the his suggestions of how that could be done.

Successfully updated karma reason!

I'm didn't think you could be such a took, precursor, but you've opened my eyes. I'm 'hated' but some pathetic fanboys, such as yourself.

And you have stil lost this argument, otherwise you would be able to debate the content of this thread. Alas, you are retarded.

Successfully updated karma reason!

DW, no one was questioning that the AI has weaknesses (though this is hardly an orginial point), but merely all the extra crud that was attached to it (lame suggestion, general aura of ignorance).

Successfully updated karma reason!

Were there suggestions in the original post which cannot and will not be implemented? Most definitely. Could the poster have realized this if he'd read more on these forums? Yes. But so what? Not everyone has the time or inclination to read extensively on the boards or developer's journals (I think they're fascinating, but I'm a geek). So there will be posts, from time to time, which contain "lame suggestions" and "ignorance." Ignore them or try to be helpful - or gear up for hot flame war action.

And not all of the suggestions were lame, anyway. For instance, 2e brings up a useful point for debate: does the AI spend too much time chasing freighters around? And 2c raises a point that, while it has been raised by others (and addressed by the developers), still needs to be stressed: the AI does not do a good job of invading planets.

But if the "lame" and "ignorant" parts of the OP are too much for you, it's usually better to just let it go than to step in with a "your suggestions suck, and you suck" comment. Don't get me wrong, this thread was a riot, but it could very easily have been avoided.

Successfully updated karma reason!

I dont understand why people are blaming precurosr10 for what he's saying.
I mean, in this game, war is a major component. By the way, talking about enlisting for irak in this thread is not apropriate in any ways.

I also agree with that fact that the A.I. need tweak about what they are doing in a war. I dont think they should focus on human, but they should simply focus on doing thing corectly.

I mean, I'm playing at challenging difficulty, where is the challenge when the stongest military fleet of the better A.I. in your game is a fleet of 15 small ship with a total of 50 missiles points for attack ??
2 huge hulled ship I designed beat them easy. So with 2 ships I beat all their entire fleet. That's lame.
They are been war between different races in almost all the games. It look like the A.I. is not able to build good fleets.
I mean I never see large or huge ship from the A.I., even if they researched it before me.

That game is great, but the A.I. suck at doing war, even in tactical plan.

Successfully updated karma reason!

I'm didn't think you could be such a took

A "took" is that some Lord of the Rings reference?

Learn how to spell moron. I love watching you de-evolve into the maggot you truly are.

Successfully updated karma reason!

PRECURSOR10: ON THE RECORD

The tone of my original post contained some frustration based on having to abandon 4 consecutive games and many days of work on a particular strategy. Admitted.

I guess I sadly misunderstood what was involved in making these suggestions happen because I've played so many good games out there (and, yes...some really bad ones too).

Thank you to those of you who saw that I love the potential of this game and want to see it succeed on a higher level while maintaining it's accesibility to new GC2 gamer. The only reason I bounce around on this forum (all but two of my posts to date are here in this thread) is to monitor the development of this game and, of course, kick some CARLs around.

For those of you legitimately interested in this topic or the suggestions I made, I am sorry but the thread has disintegrated into an exercise in destroying CARL-credibiity...which I'll admit, is barely more difficult that the AI on a HUGE GALAXY setting (I kid, I kid).

Sorry to the innocent bystanders who have to watch the ancillary stuff. But it's necessary. Idiots like Carl have to be dealt with. It's an ego thing. I'm sure that those of you with testicles understand.

Successfully updated karma reason!

Precursor, is this your way to try to retreat from a fight? You just more or less conceded to acting like an ass, but are lashing out against one of the ones who called you on it.

You haven't destroyed my cred at all. In fact, you come off as sort of defensive. You haven't even bothered to debate the point.

Successfully updated karma reason!

DW, his 'ignorance' came in the form of the personal insinuations he made about the designers' incompetence, which continued after his ideas were shot down. He himself admitted to his 'frustration', i.e. his moronic jibes.

And, if you actually read the entire thread, this would be obvious. He made this personal first. I and one other tried to explain what was wrong with his attitude and to reiterate the difficulty with making the A.I. do what it was going to do. He lashed out. Can you read?

Successfully updated karma reason!

Precursor, is this your way to try to retreat from a fight?

Never buddy. NEH-EH-VER! There's this great little forum feature called 'Watch'. I get a convenient e-mail everytime you show up here. So I just log on, check out your pseudo-intellectual jibberish, laugh a bit and then egg your simpleton ass on a bit more.

You can insinuate that I'm losing the 'debate' but how can that be when the AI issues I've raised EXIST? Fact brother. I'm negotiating from a position of strength. You've chosen a side that CAN'T win. THAT WAS A SMART CHOICE IDIOT!

So if you wanna give me a failing grade for my delivery. Fine. I'll admit that I could have chosen SOME words better but I'm hardly regreting the post.

Jam-tart. At first I day-dreamed about smashing your nose cartalidge into your brain cavity but now I just pity you. Now that I've seen what your lowly exisitence online consists of, all violent urges have left me.

The only thing that you've proven is that Frogboy wants nothing to do with you defending him. Put that into your pipe and smoke it buddy.

Successfully updated karma reason!

FACT: You must be a moron, or else you might have been able to read the disclaimers i put in most of my posts that the issues you raised were valid, but your delivery and suggestions were crap. Given that you've just conceded AGAIN that you were full of shit, I believe I have won the argument. The fact that you even admitted to having violent inclinations also more or less makes the moron case for me; oh, and the internet stalking, which besides being pathetic on the substance (you've already conceded your poor performance in this thread) is also a bit scary.

Poor boy.....

Successfully updated karma reason!

No internet stalking. Just a simple Google search. I had to hit about 5 whole pages before the picture came clear into focus. You were THAT obnoxious. You stand out THAT much.

It's easy to de-construct (i.e. say that suggestions are crap). Try constructing something (i.e. making suggestions).

But that's not in your nature, is it parasite?

So tell us the reasons why Wonder Woman's golden lasso attack should yiield 4 points rather than 3.75. I'm dying to hear you talk about something you're qualified to discuss.

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