Complete.Org: Mailing Lists: Archives: freeciv-dev: March 2003:
[Freeciv-Dev] Re: (PR#3646) freeciv-ai ai_military_findvictim
Home

[Freeciv-Dev] Re: (PR#3646) freeciv-ai ai_military_findvictim

[Top] [All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index] [Thread Index]
To: ue80@xxxxxxxxxxxxxxxxxxxxx
Subject: [Freeciv-Dev] Re: (PR#3646) freeciv-ai ai_military_findvictim
From: "Per I. Mathisen" <per@xxxxxxxxxxx>
Date: Fri, 7 Mar 2003 08:03:55 -0800
Reply-to: rt@xxxxxxxxxxxxxx

On Fri, 7 Mar 2003, ue80@xxxxxxxxxxxxxxxxxxxxx wrote:
> When looking at some AI-games (with diplomacy patch) is saw that AI is
> better in landwar thenin seawar. Really bad in expanding, bad in
> cityplacement, not optimal in the building it builds ...

True. Mostly because the settler code sucks. Secondly because the ferry
code sucks.

> > Can someone give sketch some ideas on how the AI _should_ use paras? Since
> > the AI is omniscient, its use of paras can be deadly, if coded correctly.
>
> Jump into empty cities or on adjacent tiles and conquer it after.
> That is not more or less deadly than normal players do. I thought we
> want an AI which can beat players like K and not an AI which is
> hopeless.

I'm afraid it is not that simple. You have a para in city X, there is no
target nearby. This is the most common case - you build them one place and
use them some other place. What do you do? Do you let the standard AI
routines handle it until you are in a city and there is a jump target
somewhere? Then it will be unlikely to activate often. Or do we give it
over to special para code which attempts to find cities from which it can
jump? Here we risk not using paras as regular troops when they could and
should be.

  - Per




[Prev in Thread] Current Thread [Next in Thread]