This is the mail archive of the
xconq7@sources.redhat.com
mailing list for the Xconq project.
Re: ...and the AI's decision-process around resignation
On Sun, 2002-06-09 at 11:21, Jim Kingdon wrote:
> I believe that point value only affects resignation, not building or
> other play. So maybe cut the point values which are more than 10 and
> increase the ones of the units which can build cities. The choices in
> the standard game aren't so great (if I had time to futz with it, I'd
> try cutting cities to 5 like towns or maybe 7; I might also put
> battleships and carriers down to 0 like the other ships but I don't
> think that one would matter as much).
As I continue to examine the AI's behavior, I'm no longer sure that
point values are affecting the AI's game play (short of resignation). I
think that the AI instead needs to colonize more quickly.
In this game, exploration usually is not an issue like in most Civ-style
games because the world is already known at the start of the game. So
the AI should immediately be able to find a nearby spot to build its
first city.
>
> What to build is controlled more by things like u_offensive_worth (see
> offensive_worth in ai.c) and the others.
Looking at that code, I see that there is a rather complex process used
to determine the value of units. However, I can't figure out where it
makes decisions about resignation. I can see the functions in ai.c
where it actually performs the resignation, but I can't figure out where
it's called from.