So, I've recently picked up the GC2: The Ultimate Edition. I had never played any of the GalCiv games before, but decided to give it a try based on the many positive reviews. One thing that the reviews cited in particular was a superior, challenging AI that GC2 supposedly has. This was a big selling point for me, since most 4x games are ruined by brain-dead computer opponents that don't take proper advantage of a game's many nuances and fail to provide a good challenge for a human player.

Well, after playing a few games on the Tough AI setting (which supposedly maximizes the computer's strategic ability), I have to say I'm unimpressed and rather puzzled by the praise the AI received from the reviewers.

The AI is bad at developing its planets. It will often fail to take advantage of the bonus tiles other than the Research or Production ones. It will not rush-build factories to speed up a planet's initial development. In general, it will have undeveloped planets sitting idly for many, many months slowly building advanced projects which the planet's industry is incapable of completing within reasonable time. It will not cultivate large, taxable populations and is nearly oblivious to the population Approval.

The AI is terrible at diplomacy. It will trade its technology, warships, and even planets away for very small amounts of money or low-level, mostly useless tech and without regard to the loss of its strategic advantage resulting from such trades. It is far too willing to be bribed into making war (or making peace) with the most dire consequnces for itself.

The AI fails at geopolitics. It does not mind its borders and those that violate them. You can mass warships close to a computer player's key assets for a future crippling assault, and AI will only respond with a mild diplomatic penalty towards you. As long as you manage to maintain good diplomatic realtions with AI players in the first place, they will allow you to get away with extremely dangerous behavior.

Speaking of which, the AI sucks at the Influnece game. I am able to stage hostile takeovers of entire civilizations with Influence Starbases built inside thier territory and, again, suffer no reprisal aside from a minor diplomatic rebuke. Nor will the AI attempt to counteract my influence with its own starbases or planetary structures, and I have rarely seen the AI build influence starbases of its own at all.

The AI is oblivious to another civilization (human or AI) actively pursuing and approaching one of the victory conditions. I've been able to pursue a Technology victory undeterred. I've achieved Diplomatic victory without other civilizations making war on me, trying to ally with me and my allies, or attempting to break my alliances via pitting my allies against each other. I've witnessed an AI civilization go for an early Ascension victory and come within 100 days of it without other civilizations ganging up on the culprit until I paid them to declare war.

Finally, the AI is poor at war-making. Its entire war strategy seems to be limited to pumping out lots of ships and engaging enemy fleets until total military superiority is achieved, at which point it will send some transports to take over the undefended planets.  The computer's attacks seem to be very broad and undirected in scope: it doesn't purposedly seek to achieve local superiority to capture/destroy key enemy assets quickly, it does not preemptively position fleets in preparation for war, when on the defense it does not stage effective counter-attacks. Overall, when confronted with an opponent of similar technological/industrial level the AI wages a protracted, ineffective, economy-draining war of attrition.

The only thing the AI is truly good at is maximizing its industrial capacity and achieving high military production output. Its planets tend to be factory-heavy and it seems to like to adjust its economy settings towards industral productivity. Thus it can surprise a novice human player with a very large amount of warships. But lots of ships hardly equals good AI. Once the human player figures out how to min-max the game's (atrociously bad!) economic model he can match and surpass AI's ship production and make its advantage go away.

Although it's not an AI issue, I feel that the horrible economic model desreves a special mention. I can find no merit in this unintuitive monstrosity governed by multiple sliders and setting across different interface panels that forces the player to maximize one type of production (usually industrial) across his entire empire to the complete detriment of another or else suffer gross inefficiency across the entire planetary infrastructure. The income model is likewise extremely flawed. By making wealth generation be driven primarily by an exponential growth population function the game assures that the civilization which pursues Economy/Growth/Approval bonuses will run away with game's economy including prodution, while the one that does not will remain chronically poor and underpowered across the board.

Despite all I've said here, I'm enjoying GC2. I love the vast variety technologies, civilizations, and empire building options. The extensive spaceship design workshop is great in particular! I like the crisp graphics too, although I just wish the game would give me more reasons to zoom in on the action rather than play from the zoomed out strategic view.

But as to the claims of great AI... I just don't see how these are justifed. Anyway, just my 2 cents.