perhaps that should be given to the Experimentators^^
seriously there's a good game out there ("Eador") where "the ability to go back in time" is actually the special-ability of the player

Moderator: Oberlus
Depends on how exactly you'd expect that feature to be implemented, and what you understand by "easy".Kassiopeija wrote:Wouldn't that be easy to do?
Well, that's not only a matter of the personal preferences of the active devs, but also one of individual strengths - coding an AI requires a specific "skill set" (so to speak), and at least some love for all the advanced math involved specifically with the field of artificial intelligence. That's not everyones cup of tea. I, for example, am useless on that front. I did some coding (migrating universe generation to Python has been my largest coding contribution so far), but AI is simply far beyond my skills.ovarwa wrote:I admit disappointment regarding the focus on adding new features instead of on improving the AI
"Not Human" should work. Make sure to exclude unowned objects, though.Morlic wrote:Geoff did some work to give human players on beginner difficulty some advantages: https://github.com/freeorion/freeorion/pull/928
I suppose you could use that feature to give the AI some advantage on the higher difficulties (I guess substituing Human -> Not Human should work, Geoff?).
[...]
Nothing for the "official" branch, but this way players who are looking for harder challenges may create a "cheating" AI for themselves modifying only the FOCS scripts...
Part of the problem is that you can't really make an AI that 'understands' the game unless you 'understand' the game. If the game is changing because of the addition of new features, the AI will be out of date (since it is hard to tell what effect a new feature will actually have on the game)ovarwa wrote:Hi,
Playing against an AI that 'understands' how to play a game is a very different experience than playing against an AI that gets (big) bonuses but does not. Both can be challenging, but the latter tends to be far less satisfying.
Sort of like playing against a chess program that starts with four queens.
I admit disappointment regarding the focus on adding new features instead of on improving the AI, tempered by the awareness that *I'm not the one implementing any of this*, and that the devs are writing the code they enjoy on their own dime.
Anyway,
Ken