Sirian Wrote:Montezuma will declare war and try to take cities, but does he ever threaten to reach domination or conquest? What use are his efforts? In the end, he does the same thing all the other AIs do: he techs along.
Although I stand by this criticism, I do want to elaborate on just how many things the Civ4 AI does right. I don't suppose I've ever compiled a list for public consumption. The list that follows is far from exhaustive, but shows off some of the most important achievements.
*
The AIs do not get tied down in endless useless warring that never goes anywhere.
This is probably the biggest strategic achievement for the Civ4 AI in its plus column. The Civ3 AI got tied down in endless wars, and they were useful only by accident, in that most if not all of the AIs would get chained in to the war via temporary military alliances, and dogpiles would form against those who ran out of money (including credit).
If the AI were to be changed to remove the "inevitable warfare" chaining effect of the cheap and mandatory (mandatory to the AI) alliances, then the thing that made all that warfare "go somewhere" would be gone, and all that would be left would be small, isolated wars and mostly phoney wars that would simply act as a drag on those who get involved.
Since Civ4 changed the diplomatics and removed the cheap alliances and reduced the inevitability of the AIs entering wars, if the Civ3 AI were to be copied in to that environment, its performance would be very sad.
The Civ4 AI may only be competent at pursuing the space race, but at least it IS competent there, given an appropriate handicap to match the skills of the player. This is chiefly because the AIs can now say "no" to an invitation to join a war, chiefly through "redding out" the option to attack their friends.
*
The AIs are not completely stymied by strong city defenses.
The AIs will pillage you bare if you huddle in your cities, and they will bring up siege engines that will take down city defenses if all you do is huddle behind the walls. This prevents the defender's bonuses from breaking the game, even though the pillaging ends up being universally applied (when it should not be). Making you come out of your cities to fight is a big deal. The chief problem with this behavior is that it's the only trick in the bag. If the AI had other capabilities and other strategies, and switched up from time to time, the player would not know for certain what to expect all the time.
*
The AIs are competent at managing their own cities.
They choose city sites pretty well. Probably the weakest remaining aspect of their selections is managing coastlines on water-dominated maps. They too often settle one tile inland. Other than that, they are going strong. They do tend to be cottage heavy, but not as bad now as on first release. Compare to the Civ3 AI who would be running food surpluses of eight and ten extra food at cities stuck at size 12 in pre-Sanitation days. Humans can still outmanage them, but they got a LOT closer than AIs were previously getting, as shown by how much smaller the production bonuses are (and how much more effective they are) at the various difficulty levels.
*
The AIs no longer cheat at spotting unrevealed resources.
This alone was worth buying Civ4 to experience.
*
The AIs are aware of city maintenance constraints.
They do not mindlessly expand as far as they can as fast as they can, as they did in Civ3. Though the AI gets huge breaks on all things commerce, including maintenance bills, war weariness, and unit upgrade costs, they also carry a lot of extra units around. They are set to pause in their expansion until they can afford to incur new costs, which is how a wide open map type like Highlands, with a higher number of available land plots per civ, will see some areas remain unsettled in to the industrial age. On higher difficulty, though, where the AI gets more discounts on its costs and more bonuses on its infrastructure construction and growth, expansion happens faster, so that it may at times feel as if there is no pause. This has more do with getting the game balance right on the settings for higher difficulty than it does with what the AI is programmed to do.
Some people who read my criticisms of the AI definitely got the wrong impression. Most of the Civ3 AI's problems got solved in Civ4, but the Civ4 game mechanics open up new problems, as do some of the solutions that fixed old problems. There are also many areas where the AI does something it should be doing SOME of the time, but is doing it ALL of the time, turning a potential strength in to a wholly predictable behavior, which then becomes a liability. (Predicted equals dead, in any fair fight.)
The first step in improving upon anything is a frank assessment of its strengths and weaknesses. The Civ4 AI is large and sophisticated (for a game AI) so it has plenty of both.
For any player, there are plusses and minuses with a game. The longer you play, the more that some of the plusses fade away (as you've come to have "been there and done that") while some of the minuses begin to loom (as you become more aware of them). Each player will reach a point where the plusses shrink and the minuses grow to where the net balance is negative. Sometimes "time away" doing other things will allow a return, sometimes not. I've certainly never met any game I could play exclusively for years on end.
That I'm much farther along my road with Civ4 than just about anyone else allows me to put together scenarios that (mostly, or sometimes entirely) work as intended. Honing my sense of how to shape the game to draw something new and fun out of it is at this point the larger game for me, so I don't expect ever to catch up with players like Kylearan or Blake, or what now seems like half of RB (I seem to have been passed by a lot of folks!) who have done well at figuring out the "current build" of the game and can do things like choosing all the right techs to research to maximize trading potentials, or exactly how to manipulate the pathing of invaders, etc. I'm nowhere near running dry on new ideas for scenarios, though, or ways to add wrinkles to tried and true concepts. Plenty of gaming to go, and what looks like a solid base of players to keep the reading interesting.
- Sirian
. 
. I take the AIs in the game as obstciles, not opponents. 
considering the fact Sirian highlighted, that software companies are reluctant to AI development.
