Retarded AI behaviour

I'm thinking that the behavior is related to the game difficulty setting.  If you have the AI set to "Normal" in TotA, it's a rather disappointing adversary. I mentioned something like that in our GalCiv2 team forums.

Bumping the AI to "Intelligent" helps a lot and they start playing to win. That would be the "Tough" setting if you'd like to go back and try again.  It's still not as aggressive as I'm accustomed to, so I'll be bumping the difficulty up a couple of notches next game.  One of my teammates recommends starting at "Painful".  We'll probably all find a level that can balance the challenge to our individual tastes.  

Of course, the AI in GalCiv3 is a total rewrite, so it's a blank slate.  

This behaviour might disappear at higher intelligence levels but the point is, it's retarded behaviour at a normal level. That just doesn't make sense.

I really don't see what use these type of threads have. The game mechanics are different, and the AI is being rewritten. None of the old weird behaviors are going to apply.

It's like saying "remember that weird noise that a Ford Pinto used to make? They should fix that in the 2014 Fusion."

I don't have intimate knowledge of GalCiv2's AI and as such I don't know what caused the behaviour I described. It might be something simple that finds it's way back in the GalCiv3 AI.

So I do think this thread has a point. Also, a Ford Pinto vs a 2014 Fusion is not the same as GalCiv 2 vs 3.

Starbases were essentially an after thought in the GalCiv 2. They weren't designed well. They didn't get improvement when ships got more powerful. The AIs weren't great at using them.

Since this is going to be a new game, I hope things will turn out better. Considering that they are working with new code, its possible that they will work, or not be included at all.

That behavior is present in all difficulty levels the problem is less seen though most of the time being that someone else reclaims the resource prior to the original race causing the loop to end...

As the GalCiv 3 AI has nothing to do with the GalCiv 2 AI and the mechanics of Starbases are different anyway, it doesn't matter. That the problem existed in the old game has nothing to do with the new game whatsoever.

A new AI in a new game with new technology and mechanics is going to get new quirks. Those are the ones they need to worry about, not going back and looking to see if the old quirks reappeared.

Pretty close. They're a generation apart technologically. They share no parts in common.

I laughed for a good few minutes.

Thank you so much

But OP; start playing on suicidal. and the sheer number of enemies will simply inundate whatever stupid quirks the AI has.

For instance, I was playing against the Drengin on suicidal, only for them to find the telenath and get like +3000% race abilities across the board. They had this weird little bug where they wouldn't move around some of their ships, and they just sat there like stupid, red bricks that looked like fish. However, there were so many of them (Literally tens of thousands) that the immobile ships weren't even a consideration.

It was, coincidentally, one of my best games :3

Well the reason it is retarded on normal is because normally normal people are doing other things then video games. For example games are targeted for everybody no matter how stupid or intelligent they are. This means that if you choose normal, then according to gaming logic the majority cannot comprehend the game well enough to play at that level.

On the other hand you got the mentally challenged who cannot seem to win at anything or the the people who never ever touched a strategy game before, so you cannot expect everybody to know what is good and what is bad.

Besides in all my gaming experience, The hardest difficulty AI is almost always easier then playing against a actual human even if the human has normal intelligence.

I tend to find it's not the AI that I need to worry about in a game of GalCiv 2, but the pace of my own development.  If I dawdle too long the AI has a military and starts with the extortion, but when I have a military and I'm not blatantly parking transports next to their worlds, they tend to leave me alone.  Especially when I'm paying them all to go ten rounds with each other.

The AI poses a threat right up until I start winning.  I find that toppling one race and taking their planets is enough to tip the balance overwhelmingly in my favour, through sheer force of economy - I can afford to field more ships than the other guys, and I can afford to continuallly manufacture more or upgrade the ones that I have.  The alliances which spring later in the game up can prove to be a challenge, but essentially the initial victories serve to create a runaway victory for the human player.

It can easily go the other way - the AI can gang up and take me out in the earlier stages of the game.  Starting position, exploration and the initial colony grab can heavily influence how the game goes.

Despite the quirks and shortcomings of the AI it can still play a pretty mean game of GalCiv.  It's just that when the human starts to lose heavily, there's not much incentive to tough it out and see who wins.

The 'steamroller effect' is pretty hard to control. Mostly because strategic players hate 'rubber banding' types of mechanics. Unless you have games like CK2/EU4 where 'total conquest' is basically impossible. Or 'closed' systems where you can determine when the 'tipping' point is and do something, like Shogun2's Realm Divide.

Even at the point u take an empire I bet the other Ai's don't have stock exchanges. Probably have thousands of Bc, but hardly any ships until u declare war on it. If u still have room to improve the Ai. This would be better to do at harder levels. I still would like to see the Altarians send warships over there to protect their starbases until u get enough constructors over there to militarize the starbase. The sad thing is with a 20% penalty on weapons they would have to work a lot harder than the Drengin at this. I guess the real stradegy for the altarians would be better weapons or hulls; because of their research bonus.

The Altarians have a late-game building (the Dark Energy Lab) which gives them a significant bonus to Weapons.  This effectively cancels out the penalty.  But you're right, until they build it they're at a significant disadvantage.

That's why, when you're setting up a galaxy with Altarians, you should always have a race in the mix which shares their alignment so that they can use their Super Organiser ability.

They don't need to build the Dark Energy Lab (it only provides a 10% bonus anyhow). The tech unlocking it, Dark Energy Research, already provides a 30% bonus, which more than offsets the Altarians' -20% penalty.

Also, the Dark Energy lab is far from late-game. You only need to research three techs to get it: Xeno Mysticism (50rp), Precursor History (100rp), and Dark Energy Research (300rp).

I know, it's just that typically I tend to find that it's a while before they get around to it.

I guess I forgot about the bonus on the tech.

That was another exploitable AI flaw - they never considered alliances or super organizer when declaring war. If the Altarians were pretty weak and the Torians were very strong, you could pay other races to attack the Altarians and let the Torians declare war on the aggressors.

The AI in GalCiv II scales in that it's not fully active until you play at a certain difficulty level, and "normal" isn't it.  It's either tough or challenging.  It's weird, since you might THINK normal would equate to normal AI, but it's a few bumps higher to fully activate it.

The AI still may do some questionable things since, this may shock you, no AI is perfect or as good as a human, but when you play with the AI below it's full capacity it makes more mistakes - by design.

I think it's a pretty slick setup for AI because for one, it's nice that the different races have some different behaviors instead of fighting a bunch of AI clones, and for two, no matter how good YOU are as a player you're going to make a mistake here and there, so an AI that makes mistakes sometimes is playing more like a player, and that's cool to me.