View Single Post
  #17  
Old June 12th, 2003, 08:18 PM
cybersol's Avatar

cybersol cybersol is offline
Corporal
 
Join Date: May 2003
Location: SF Bay Area, CA
Posts: 145
Thanks: 0
Thanked 0 Times in 0 Posts
cybersol is on a distinguished road
Default Re: My AI Design Q&A

Quote:
Originally posted by JLS:
This appears conflicting, if you Modified the se4 Default setting of = False to true...
Will the AI player (1) actually? ~ “ once the first ship enters that square and does not blow up then it will most definitely clear the minefield marker and move on to the next one.”

It may be recommended with base se4, that you set Mine Sweeper availability to be a high priority tech to Research, if you wish your AI to expand?
Well, by definition in creating a AI, I am modifying the AI files of a race. So all changes to the AI files of that race are fair game. "Ships don't move through minefields" is in xyz_AI_Settings.txt so I can change it to whatever I want. But I am not changing the data files, thus in essence the goal in a competitive AI for the unmodded game. This is similar to the goals of the TDM Modpack for instance, only on the scale of only one race for me right now.

Now their is a trade-off to both settings of ships don't move through minefields. If it is true, then only "Mine Sweeper" Design types will move into marked minefields. Predictable and since they travel alone, they could be prey to an intelligent human opponent. If it is false, then the behaivor is less predictable but the AI acts as though it doesn't know about the minefield. So it sends innocent ships without minesweepers to their demise.

So to answer you question explicitly, the AI will move a "Mine Sweeper" design type into the mine field even when set to true. At that point if at least one ship survives, my experience so far is that it will clear the marked minefield. It may be smart and clear these marked minefield when making a treaty, but I wouldn't bet on it.

Yes, ship construction, an early weapon, point defense, and then mine sweepers are early research priorities for the AI IMHO.
Reply With Quote