Complete.Org: Mailing Lists: Archives: freeciv-dev: March 1999:
[Freeciv-Dev] AI Diplomacy theorization.
Home

[Freeciv-Dev] AI Diplomacy theorization.

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: freeciv-dev@xxxxxxxxxxxx
Subject: [Freeciv-Dev] AI Diplomacy theorization.
From: Tony Stuckey <stuckey@xxxxxxxxxxxxxxxxx>
Date: Wed, 24 Mar 1999 23:12:14 -0600

        Since it's come up a couple of times, and things seem to be heading
to a point where it would be useful to have, we should talk about the AI
and how it should react to human players.

        One current problem with the AI is that it does not take world size
into account.  An 80x40 world with 14 players puts people basically in
sight of each other to start.  A 200x100 world with 4 players means that
you could quite possibly get into advanced years and technology before
ever meeting another player.
        Clearly different strategies are useful in each of these worlds.
If the world is large and there are few players, tech leading to
fast-moving explorers and/or seaworthy vessels would be good investments.

        Once another civilization is met, though, there has to be a
completely different AI calculation, that is not so objective.

        The Civilization 1 diplomacy options were basically useless,
especially for any purpose other than going to war with a computer player.
There was very little option for negotiation, and the AI attitudes were
laughable at best, and annoying and insulting at worst.  Getting a "we will
agree not to crush your worthless civilization if you give us Robotics"
from a civilization 1/5 my size and 20 advances lower on the tech tree was
a common experience.  Refusing the AI demands was often an immediate
declaration of war.
        Civilization 2 added useful propositions like "please remove your
army from my land", and provided somewhat more steps to the interaction,
which could often reach acceptable complexity levels.  However, once again
the AI would often act megalomaniacally.  It was common to see an
established city with a border army in a fortress, a computer settler found
a new city close to that border fortress, and the next turn, a demand that
you as the human player "remove your army from my land", even though the
fortress was both older and within your own city radius.

        International politics were no better.  When approaching launch of
a spacecraft, it was common to see round after round of computer players
frantically trading technology to each other so that they would each be in
the space race as well.  The final rounds of every game tended to be deeply
vicious defensive actions against multiple allied computer players.

        Now certainly part of this is called for.  It's generally a bad
idea to simply let someone else win.  But it's my opinion that the
interaction could be made much more satisfying.

        By giving the AI many stated reasons for its actions, ranging from
friendly relations to outright blackmail, we should be able to provide an
entertaining diplomatic experience.
        By keeping the civilizations somewhat more close together
throughout the game, we should be able to increase the necessity for sharp
wits and decrease the "everybody beat on the leader" mentality.

        It's pretty much a fact of the Civilization world that it's very
hard to win a war against a technologically superior opponent.  But it's
also the case that an early resource lead can become an insurmountable
advantage later in the game.

        Should civilizations which have more than their "fair share" of
land turn turtle, and try to keep a science lead?  Or should they continue
their current strategy of trying to win by military conquest?
        Should civilizations which are clearly losing attempt drastic
actions like provoking global warming?  Should they then hold that over a
player's head as well?  "Give us Edinburgh, or face global warming?"
        Should a computer civilization offer a stripped, rioting city?
After stockpiling offensive units nearby to take it back and plunder a
technology?  Or simply to destabilize a Democracy?  How much history should
they keep for such "hot potato" items, to make sure that they don't get
screwed on a return within a couple of turns?
        How contrite should they be when actions fail?  "Okay, I sent 6
spies against you, and didn't get much out of it.  Let's let bygones be
bygones, okay?"
        Should lesser civilizations enter into mutual improvement pacts?
Planning to research different things to trade to each other and speed the
rate of research return?

        What level of AI cooperation wouldn't be fun any more?
Self-promotion is fine and all, but part of intelligence is realizing when
things aren't working, and changing strategies.
-- 
Anthony J. Stuckey                              stuckey@xxxxxxxxxxxxxxxxx
"When I was young, the sky was filled with stars.
 I watched them burn out one by one."  -Warren Zevon

[Prev in Thread] Current Thread [Next in Thread]
  • [Freeciv-Dev] AI Diplomacy theorization., Tony Stuckey <=