I too prefer to replay MOO, but I had more fun in the beginning with MOO2 until the amount of micromanagement (not only on the late colonies, but the population transfer to maximize growth in the beginning is horribly boring, and mandatory in MP) made me realize MOO was more fit to be replayed (and I think the AI plays better in MOO).
Yep, but but there's a funny twist in it: the AI in MoO is actually tragic, yet it keeps giving the impression of being good. It's totally incomparable to the MoM's one (which probably means: to the MoO2's one as well), and it's prehistoric compared to Civ2/SMAC AI. Spear and tank kind of "incomparable". But it works somehow better.
There is some absurdity in this, but maybe something that modern developpers should reconsider. A good bunch of games had a very poor AI, yet the gaming experience is top notch and intellectually challenging. Prime examples: MoO, X-COM, and lately a new surprise member, Eador. The trick for me seems to have 2 major points and a third minor one:
1) These games are excellently set-up to cover for any AI defficiency. They use simplification of rules, they even premeditate the rules (it seems) to help the AI. Example: in MoO you cannot attack a fleet in the space, only at planets - so the AI easily know what's the E.T.A. and destination of your ships - very effective. In X-Com, the sheer lethality of alien weapons make the AI look good. And the 10 waypoint weapon is an AI cheat, because they obviously could not make the pathfinding routine to use only 3-5 points necessary. But Eador leads all the way here, the tactical AI is dumbly advancing all the time, yet there are many combat traits to make that trivial tactic actually worthwhile - like move-and-shoot, spend-move-and-attack, and overall advantage of attack compared to defense. Eador has very smart conception.
2) The erratic character of the AI. This is strange, poor and amazing in the same time. X-COM (especially TFTD) is known for unpredictable alien behaviour, mix it with fog of war and close fight and you get the formula. Surely, the code was broken and we know there pre-programmed paths. Yet, after all those years, I keep losing my Aquatoid off sight, and popping 2 or 3 turns later killing my soldier. MoO has some incredible situations as well, massive AI decisions are separated by some kind of butterfly effect, that can send the game in an "alternate universe". Just ask the people who play the game form the same save, how differently it can spin. (Eador is not erratic, though.) The AIs that are good but formulaic stand no chance against a clever human after 25 games. The AIs that are strangely inpredictable (but not dumb-inpredictable) can win after 100 games, though.
3) The minor point: the bugs. Huh? Si, the bugs sometimes add to the game experience. Like the chryssalid hatching its egg even when it fails to hit your soldier. Or the MoO strange ship distribution (creating the fleets of doom) which is actually a bug. Or the unpredictable choice of planet target in the same game. Or the invasion to a settled planet of a player with Non-aggression-pact. Many bugs make the AI very funny to play against.