Civ IV

[Darth_Schmarth]

I do know that Starcraft's AI learns.

<{POST_SNAPBACK}>

Yeah, I once beat the AI by killing my own troops.

[ShadowPaladin V1.0]

I do know that Starcraft's AI learns. They've started to rush me and my friends with lings.

<{POST_SNAPBACK}>

 

Can it remember what it's learned?

[Calax]

yep, every single time we play they block the entrance to their base with Supply depots or somthing similar. (I said they learn, but not very quickly) so the lings don't get in. Also the lings don't go after the depot's because they keep reargeting the bunkers behind the depot's.

[ShadowPaladin V1.0]

yep, every single time we play they block the entrance to their base with Supply depots or somthing similar. (I said they learn, but not very quickly) so the lings don't get in. Also the lings don't go after the depot's because they keep reargeting the bunkers behind the depot's.

<{POST_SNAPBACK}>

 

Is that in skirmish mode ?

[Calax]

ja.

[alanschu]

Ooh, this sounds familiar, but I can't put my finger on it.  I'm sure there was a strategy game released some years ago which recorded all your games and sent them over the internet to a central computer, which would then analyse the best strategies and constantly update your computer's AI, making it increasingly more challenging at the higher levels of play.  What was this game called?

<{POST_SNAPBACK}>

 

Galactic Civilizations

[SteveThaiBinh]

Galactic Civilizations

<{POST_SNAPBACK}>

Which I've got, and I can't say I ever noticed it doing this.

[alanschu]

I don't know if it's dynamic with respect to how you specifically play.

 

I believe it may have just been modified by the strategies of the top players in the Metaverse.

[Deraldin]

I never noticed the Starcraft AI learning. However I did notice that whenever they came out with a new patch on Battle.net that the computers would magically become smarter. "

[metadigital]

I do know that Starcraft's AI learns.

<{POST_SNAPBACK}>

Yeah, I once beat the AI by killing my own troops.

<{POST_SNAPBACK}>

I think that is wonderfully innovative. I wouldn't recommend it IRL, but if ever Skynet does appear, we've got a walk-up start ...

[Skynet]

*appears*

[metadigital]

Your occipital lobe must have been burning, because I thought of you when I wrote that, but I didn't expect the Spanish Inquisition ...

[Calax]

Ha HA!

[EnderAndrew]

I love the little red box.

[213374U]

*appears*

[Guest GroinOfDespair]

Skynet mocks the governent computers like Echelon that track us all, but they are reading how you are mocking them right now.

 

Don't be suprised when Skynet *disappears*

[metadigital]

So your pessimistic ability to "see" through the authorities and their all-knowing, all-powerful and all-controlling influence on us will mean that you, too, are soon to be *disapeared * ?

[Fionavar]

Derailment seems possible ... let's hope such the AI can reinitialise on topic discussion ...

[Darque]

Hopefully the AI won't suck.

[alanschu]

A simple and blunt way to bring things back on topic

 

 

I always found the AIs in Civ to be a bit too war mongering, even if they weren't militaristic or expansionist.

[metadigital]

Agreed. Even the "Americans" led by Abraham Lincoln (back in Civ2) used to get expansionist on my border ...

[alanschu]

Perhaps he thought that slaves were in your borders? :D

 

It'd be neat if the AI leaders had certain opinions about the slavery that will exist in the government dynamic.

[metadigital]

If the game handles multiple variables, then it isn't much of a stretch to have the AI make decisions based on those variables.

 

 

 

 

Unless the developers don't programme it.

[alanschu]

Good point. It's kind of a given, and not too terribly difficult to do.

 

I imagine it'll be some of the big differences between the different leaders for the same country and whatnot.

[metadigital]

Good point.  It's kind of a given, and not too terribly difficult to do.

 

I imagine it'll be some of the big differences between the different leaders for the same country and whatnot.

<{POST_SNAPBACK}>

Well, it's not difficult if the developers planned to use the variables from the beginning, and set up some sort of Object Oriented Design. I would always do that, because the probability is very high that the variables used to identify / differentiate / leaders, as well as discriminate their actions and even measure their progress.

 

That way, any attribute can be quickly added to any of these AI task centres, without needing to add the entire processing thread (as in the case of non modular design).