I'd enjoy making a real good AI that can handle battles on ground, air and sea. The final version should even be enhanced with some variables defining how the AI 'plays'. It is careful ? It is aggressive ? Is it specialized in ground battles, but is weak when it comes to air threats ? Will defend itself or waste all resources for 1 big attack force ? That way the AI could act like different persons which should be quite fun to play against.
Anyway, i thought about it and made these basic rules for my 1st version of the AI:
- a Base minds its own business (troops of different bases don't work together in this version)
The AI wants to gather info into its variables. It's like seeking an answer to certain questions.
Order of important questions:
1.) Where is the enemy ?
2.) What is the enemy ? Which weapons ? How many units have been spotted?
3.) What is my best reaction force to this threat, regarding my resources ?
4.) Where is the best place to attack / to defend for my groups ?
5.) Judging my info of known battle outcomes, what went wrong and what was my greatest victory? What do i need to do to get 'better' and win the war ?
Obviously there are many variables needed that will simulate a 'thinking & judging' routine. The point is that the AI mustn't cheat in any way. It has to use 'real' info and judge everything with these infos like a real human would. It would certainly be easy to code an AI that never fails and knows everything from the start, but i don't want to do that. If the AI has wrong or outdated info because the players chose the stealthy commando-gamestyle then the players should win the war this time. That's ok, it should be fun to play - maybe too much realism isn't always the answer.
Any comments, ideas ? If you feel i'm getting off topic within this thread just say so, i'll start a new one.