Quicksilver did not make up the 4x name. Its been around at least since the original MoO. Also, games like Silent Storm and Fallout are not TBSs in the same sense as GC2, Civ, or MoO. Completely different AIs for the same reason that you can have brilliant chess AIs but cant really have a brilliant Civ4 AI. Comparing GC2 to something like Silent Storm is pointless.

If I compaire it though to Civilization and Alpha Centauri and those games i'd say the AI is simular to SMAC if not slightly worse but FAR better than any Civ game.

Hu ?

Civ3 and Civ4 AI dust the floor with SMAC AI...
So I fail to see how you can have an AI better than Civ's but worse than SMAC's ^^

[[There are no TBS games where the AI is all that smart.]]

Obviously you haven't played ALL TBS games then because there are sevearl TBS games with AI's that can whip the socks off of you.

"War of the Lance" for instance is one of the better games of the 80's with an AI that is downright ruthless and hard to beat.

"Spartan" v1.013 has one of the best AI's I've ever played against and another that will whip the socks off of most players on just the "normal" setting (it was so hard they had to DUMB IT DOWN in patch 1.017 because of all the whinners saying they couldn't beat it). Move on up to "impossible or very hard" and you will see smart and ruthless and grueling play vs the human player.

Panzer General has an excellent AI in campaign mode another that will beat the soxs off of you.

Civilization IV has an excellent AI and will beat the soxs off of you.

Combat Mission with certain settings will beat the soxs off of you.

TBS AI can be smart if players will give them their smarts (that have been programmed in) and up the difficulty levels. Normal is for unexperienced players. Very hard and Impossible are for the brains and those with IQ's over 100. And the AI show's it's smarts when you play the extreme difficulties. No, it's not cheating, it's just smarter, people just tend to call things cheats when the AI beats them over and over and over again. lol But, of course humans never cheat the AI right? Humans never exploit a mechanic of the game to beat the AI right? And when a human does this he/she feels "smart" correct? So, I see no cheating going on by the AI when it exploits something vs the human like advantages in what it sees and techs it might get at the start of the game or speed of teching up, etc. etc. They humans cheat, the AI should have the same advantages and cheat too. That's SMART!!!

Civilization IV has an excellent AI and will beat the soxs off of you.

And the AI show's it's smarts when you play the extreme difficulties. No, it's not cheating, it's just smarter,

Uhm, no. The Civ IV cheats quite a lot, even on easier settings. Try it out.

Obviously you haven't played ALL TBS games then because there are sevearl TBS games with AI's that can whip the socks off of you.

There was never a game where the AI did not cheat and beat me. That doesn't mean that I'm very smart or very good at TBS. Quite frankly I suck at them. But its easy possible to use AI weaknesses to win.

I can beat the Civ IV AI on the hardest difficulty, and it's not the AIs smartness that is a problem (the AI wastes its units on my units on hills and in forests) but the fact that the AI spawns dozens of units out of thin air.

If I had to rate the quality of different game AIs I've played I would rate them as follows:

On a scale from 1 to 10 with 10 being perfection and 1 being a joke.

Galactic Civilizations II: 8
Civilization IV: 7
Civilization III: 5
Space Empires IV: 2
HOMM IV: 2
HOMM III: 5
Warlords IV: 5
MOO 1: 3
MOO 2: 5
MOO 3: 1
Rise of Nations (RTS) 4
Age of Empires 3 (RTS) 4

Anyone saying that the Civ IV AI is "great" though needs to define what is great. It's easy for a game like Civ IV to have an AI that beats you if it cheats. The AI in Civ IV cheats very early on in the difficulty levels and cheats severely as you go further up.

I still feel that Civ4 has a better AI than GC2. If they can get the AI to build good planets in this game it might become the best, but right now it just doesn't use it's resources well enough.

The game that provided the best challenge was Kohan 2. While the AIs that came with the game weren't very good, the community built some that were using the latest strategies and provided an outstanding game - almost as good as playing multi.

I don't think it is totally possible to compare Different AI's from games directly. Why? Well even in teh 4X Sc-fi genere, there is enough variation between games that they would be handled totally differently from game to game. Take Ship design in GC2. It is very simple (attack vs size vs type) when compared to MOO2 (attack vs size vs type vs special vs imporvements). Thus (hopefully) GC2 AI has an easier time building useful ships than MOO2.

Its a double bladed sword really. The more options, techs, and cool things we put in our games, the more stuff the AI has to deal with. Think of some other Games where there are very contoverial options for the player that are hotly debated over the forums (Like which infantry to use in RTR or Phasor vs. Deathray in MOO2). Now imagine an AI trying to figure it out

While the AIs that came with the game weren't very good, the community built some that were using the latest strategies and provided an outstanding game - almost as good as playing multi.

How is that possible? Did the company have to release the source code? I agree that is the ultimate solution to having a smart AI. Imagine hundreds of people competing to make the smartest AI that could beat the others or the human players. No single group of paid programmers and testers could compete. Just like Stardock could never have designed all the awesome ships that people have uploaded by themselves.