[freeciv-ai] Re: long-term ai goals

["Ross W. Wetmore" <rwetmore@xxxxxxxxxxxx>]

At 06:52 PM 02/05/25 +0100, Gregory Berkolaiko wrote:
>On Sat, 25 May 2002, Ross W. Wetmore wrote:
>
>> Making sure that any system is not just AI-specific is I think the best
>> way to deal with the cheating issue. This means ruleset, and/or console
>> control with the ability to turn it on/off securely, possibly on a per
>> channel basis. 
>> 
>> Note, instead of a constant (TRUE) or FoW boolean (bool get_known()), you
>> need to introduce a weight in the range 1...0 and a pseudo-random selector.
>> The current FoW sort of does this with a time dependent grey-level, but
>> a weight-level could do the same thing.
>> 
>> If you assigned a float or scaled int instead of the TRUE/FALSE bit, then
>> aging of any values less than 1 can be done on a turn update. Any values
>> reported on map updates or queried during player-specific server
>> computations would be gated by having get_known() or get_known_and_seen()
>> return the condition (known_rand_computation() > weight(pplayer)). For
>> weights of 1 and 0 this will be just the current trivial FoW boolean or
>> omniscient TRUE.
>
>Alternatively you can introduce a delay on the FoW onset.  The 
>weaker player will be allowed to see the tile he has left for N turns and 
>then it will become shaded.  I much prefer it to the random model.
>
>G.

But being strict about the "tile he has left" doesn't help with the case 
where the tile has not yet been seen by the strict FoW rules. This is the
part that must be addressed, as the move from omniscent to FoW is too big 
a jump for the intuitionally challenged. They need "news reports from
distant lands" to trigger them into recognizing that there *are* distant 
lands. Alternatively, you can think of this as a pseudo-random explorer 
agent that handicapped players can use unitl they learn to develop the
appropriate levels of intiution and technique to manage their own 
explorations.

But it is certainly the case that you could use the existing time model
as a form of monotonically decreasing weight function, with certain
elements becoming unseen or subject to misinterpretation at a given point 
in time to get a smoother transition.

I actually like the random aspect because it may cause things to change
in more varied ways reflecting the fact that secondhand reports are not
always that accurate and are constantly updated. This helps with the 
omniscent aspect in that you can no longer count on correctness unless 
you actually see something under the strict (firsthand) rules. FoW and
omniscience are still binary - I like fuzzy which is that data may be
neither TRUE nor FALSE but actually partially or completely wrong.

This also serves notice to code like the CMA trade route analysis that
gets hung up on absolute accuracy concerns and forgets that this stuff 
is all fuzzy and subject to random change under normal game flow anyway.
These are only estimations - code and coders should treat them as such.

Cheers,
RossW
=====