AI's UNcompetitiveness

[maltz]

I recall the days when AIs declares war on the superpower/winning civilization (often it is the human player) just to preserve its own shot to victory. But after patches, AIs are quite uncompetitive. They are now just a tragic decorations of a human player's certain victory if the player has the faintest idea of how to win before the AIs eventually reach their own.

(1) Civilization A has no military, and has started the Utopia project. The AIs do nothing and let A win just like that.

(2) Civilization B finished the Apollo project and is popping out spaceship parts. The AIs do nothing and let B win just like that.

(3) Civilization C has been buying away all City States and will win a diplomatic victory if the vote happens now. The AIs do nothing and let C win just like that.

I know the game has to let the most human players win most of the time, or few people will buy the game, but it just feels silly when the AIs keep one eye shot to let another civilization (doesn't have to be human player) win 100% of the time. Is it very hard for AIs to realize they can block an impending victory easily?

(1) Declare war on a weak but a winning neighbor. Invite its other neighbors to declare war on it.
(2) Form an alliance of weaker civilizations against a super power approaching domination victory. (A lot of strategy games have this design.)

Anyone thinking similarly?

[mekane]

Originally Posted by maltz

I recall the days when AIs declares war on the superpower/winning civilization (often it is the human player) just to preserve its own shot to victory. But after patches, AIs are quite uncompetitive. They are now just a tragic decorations of a human player's certain victory if the player has the faintest idea of how to win before the AIs eventually reach their own.

(1) Civilization A has no military, and has started the Utopia project. The AIs do nothing and let A win just like that.

(2) Civilization B finished the Apollo project and is popping out spaceship parts. The AIs do nothing and let B win just like that.

(3) Civilization C has been buying away all City States and will win a diplomatic victory if the vote happens now. The AIs do nothing and let C win just like that.

I know the game has to let the most human players win most of the time, or few people will buy the game, but it just feels silly when the AIs keep one eye shot to let another civilization (doesn't have to be human player) win 100% of the time. Is it very hard for AIs to realize they can block an impending victory easily?

(1) Declare war on a weak but a winning neighbor. Invite its other neighbors to declare war on it.
(2) Form an alliance of weaker civilizations against a super power approaching domination victory. (A lot of strategy games have this design.)

Anyone thinking similarly?

yeah I agree this is a problem. The funny thing is that some other threads were created saying that the game isn't realistic because the computers ARE trying to win... i.e. Gandhi building nukes isn't realistic... You can't please everyone...

[I_qua_I]

I think the problem lies with there even being winning conditions at all.

[charon2112]

Actually, I'm all for the Civ's having their own personalities...like Ghandi going for a peaceful win. Part of the fun of Civ IV was seeing what Civ's you started near, because this would change the flavor of the game. It added to the re-playability. I think it's pretty lame to have a close ally suddenly declare war because it thinks you're getting too close to winning.

[blac]

If the computer was actually trying to win and tried to actually stop me, then I could play on a much longer difficultly and have it be a challenge. I would rather have that. There have been times when I was doing OCC and going for a cultural victory and 2 huge AI's where right beside me and they could have wiped me out. They just let me build the utopia project and off I went to victory. To me thats a problem because I can spot when a computer player is trying to do that so I stop them

[Martin Alvito]

Originally Posted by maltz

I know the game has to let the most human players win most of the time, or few people will buy the game, but it just feels silly when the AIs keep one eye shot to let another civilization (doesn't have to be human player) win 100% of the time.

I think that this problem can easily be solved. Have higher difficulty AIs ruthlessly attack any civ that nears a win condition with all the force at their command. That's all that is really required.

That would make peaceful Deity OCC victories extremely difficult to attain, but a smart human might be able to play possum and sneak a win through.

[Strategist83]

I recall the days when AIs declares war on the superpower/winning civilization (often it is the human player) just to preserve its own shot to victory.

Is that in a former Civ game? Civ5 AI never did this.

As for the topic itself, this problem is what I call the AI's identity disorder. The AI is allegedly 'playing to win', but as you give examples of, it really isn't, it just plays. IMO Firaxis needs to decide on whether the AI should be 'roleplaying' or 'playing to win'. These two are mutually exclusive, but we actually currently have an in-between - which is part of why the AI is so awful.