[Freeciv-Dev] Re: AI's advisors actually do stuff

["Baumans" <baumans@xxxxxxxxxxxxx>]

Of course that would be a good idea, but this is just very basic 
to help the ai along a little in the beginning. One way of doing 
part of what you suggested(minimally) would be to change the 
weights the ai assigns to each advisor. I know the mechanism 
is there, but is it used at all. Of course, it wouldn't change more
 minute things like attack vs. defence, but it might be best to let 
the circumstances dictate those, and to change the circumstances
(diplomacy, anyone?).
----- Original Message ----- 
From: "Jordan Crouse" <jordanc@xxxxxxxxxxx>
To: "Baumans" <baumans@xxxxxxxxxxxxx>
Cc: <freeciv-dev@xxxxxxxxxxx>
Sent: Monday, March 05, 2001 8:11 PM
Subject: [Freeciv-Dev] Re: AI's advisors actually do stuff


> Bravo - For those of us who care about the AI, this is a good step. 
> However, I wonder about the fixed route that the AI takes when
> determining which technologies to research, and which wonders to build. 
> I kind of like the way each nation tries to get a particular set of
> wonders built (i.e., Italians go for the Leonardo's wonder), so you have
> different AI players going for different things.  I would hate it if
> every single AI was trying to get the same technologies and wonders
> (especially if they all thought like me).  I think that it would be
> better if each AI took a different strategy or attitude as to where they
> dedicate their time.  This could also be useful when we start
> considering AI diplomacy, and how each AI player treats the other.
> 
> Just for a small example, consider three attitudes:  Neutral,
> Capitalist, and Warrior:
> 
> The Neutral player would build up a strong military, but only
> defensively.  His main technology goals would be defensive military
> units, and anything that encourages trade.  His Wonder goals would be
> defensive (Great Wall, etc...), and diplomatic (Marco Polo's Embassy,
> etc...).  When this player meets a new AI or human player, his attitude
> is always one of peace (unless he is attacked).  
> 
> The Capitalist player would build up a moderate military, some attack,
> some defensive.  His main technology goals would be a typically modern
> army, anything that encourages trade, and income.  His wonder goals
> would be economic and growth oriented (pyramids, trading post).  When
> this player meets a new AI or human player, he will encourage
> neutrality, but he will attack when he believes his best interests are
> at stake.
> 
> The Warrior player would build up a strong offensive and defensive
> military.  His main technology goals will be anything military or
> transportation oriented (to move his troops easier).  His wonders will
> be military (warrior academy) and transportation (lighthouse, etc.)...
> 
> This is just an example, of course, we would define more AI
> personalities and make them easily to modify much like nations and
> rulesets.
> 
> It would be nice if these attitudes were random at startup time, so that
> we wouldn't be stuck pigeonholing each nation with a given attitude
> forever.  Maybe for those who prefer a more realistic game, this could
> be an option.  
> 
> I know that this this may bend or even break the ICS rules that everyone
> is so passionate about, so ignore me if this is a bad idea.  But when I
> play the game, I would like to know that I have a 33% chance that I will
> meet a AI player who will want to encourage trade, and a 33% chance of
> meeting a warrior.  That gives me more options and considerations other
> than immediately choosing war.
> 
> Thoughts, flames and death threats are welcome... :)
> 
> Jordan
>