This is the mail archive of the
xconq7@sources.redhat.com
mailing list for the Xconq project.
...and the AI's decision-process around resignation
- From: Lincoln Peters <peters2000 at mindspring dot com>
- To: Xconq mailing list <xconq7 at sources dot redhat dot com>
- Date: 08 Jun 2002 12:32:11 -0700
- Subject: ...and the AI's decision-process around resignation
In the same "Colonizer" game, I have found that if I play against the AI
mplayer with the "Last side wins" scorekeeper, after about 10 turns, the
AI is willing to declare a draw, and then usually resigns by Turn 12.
Once the game ended, I looked at the AI player's plans and the tasks
assigned to units, and it seems that the AI had assigned a "Colonizing"
plan to all the settler and engineer units, but none of them had built
anything yet. I also noticed that the other units always had either an
"Offensive" or "Defensive" plan, and most of the units that required
fuel has used up most of their initial supply.
My best guess as to this behavior is that when the AI sees that it's
running out of fuel and it's not producing a heck of a lot, it fears
that it'll starve and surrenders. I'd guess that it looked on the
situation similarly to medieval and classical warfare, in which a
heavily fortified location often had to be starved into surrendering if
it couldn't be captured by military force.
The strange thing is that the mothership (which every side starts with)
*can* produce 1 fuel per turn, but it must hold its position to do so.
Furthermore, that should be enough because no units in the game consume
anything while they're occupants, and there are no threats early on that
would justify any of the fuel-consumers being active.