What do you expect from the AI in Civ 4

As someone in another thread pointed out, the Civ games have always been MP games.. just with AIs as the other players. Focusing on actually doing the AI last is probably what they want to do since during that playing of multiplayer
1. There could have been large changes to game design that would have affected basic AI strategic approach
2. They would be taking time to figure out what the best strategic approaches Were to the game, after which they could begin programming a very good..scratch that ... competent AI.
 
Ozymandous said:
Hmm, well if it's quicker and easier to develop graphics than AI then why did they wait until only the past 6 months to focus on AI when they have been playing it MP style for 18 months?
Becasue that's the only way it can be done. In order to write and test the AI, the game has to be playable, the game rules have to be close to finalized, and the guy writing the AI has to know how to play the game well. Playing multiplayer games first is probably the best way to write a decent AI. It's kind of like teaching a person to play the game. You'll be a very poor teacher if you haven't played the game, don't know the rules, or aren't a good player yourself. Of course writing the AI has additional difficulties, but if the programmer doesn't know how to play the game, it's doomed from the start.
 
Thank you, I have been trying to get that through myself. As stated in previews playing MP has helped the developers to fine tune the AI. Hopefully that will give us a bigger challenge and more variety in computer player behavior.
 
To be fair, Ozy, the Firaxians have already stated that they have been working on the AI-on and off-since the game started development, with the more intense development occuring in the last 6-8 months. What they have also said they are doing differently is that, by having MP up and running from the get-go, they can see how human players react to the actions of other human players-and adjust the AI's behaviour accordingly. i.e. they hope to make the AI savvy to player exploits, close player exploitable loopholes wherever possible and create an AI which has a reasonable chance of fooling a human player into thinking they are playing another human. So, its not that they are simply focusing on MP for the sake of MP, they are focusing on MP as a means of developing a better MP and SP game.

Yours,
Aussie_Lurker.
 
Aussie_Lurker said:
To be fair, Ozy, the Firaxians have already stated that they have been working on the AI-on and off-since the game started development, with the more intense development occuring in the last 6-8 months. What they have also said they are doing differently is that, by having MP up and running from the get-go, they can see how human players react to the actions of other human players-and adjust the AI's behaviour accordingly. i.e. they hope to make the AI savvy to player exploits, close player exploitable loopholes wherever possible and create an AI which has a reasonable chance of fooling a human player into thinking they are playing another human. So, its not that they are simply focusing on MP for the sake of MP, they are focusing on MP as a means of developing a better MP and SP game.

Yours,
Aussie_Lurker.
AL, that is indeed what they say..and I for one would be much less skeptical if some reviewer out there released an independent review with DETAILS of how the SP AI side of things is shaping up. How does it play in epic mode? How is the Middle and Modern era's? Deos the AI handle troops efficiently in the new system? Can the AI use Great People effectively too.. or prevent the human from exploiting their characteristics? Does the AI follow an intelligent building strategy? Is the dreaded sub bug gone when dealing with hidden units? Will the AI build enough workers? Will the AI build enough cities?

The lack of such a review has me concerned...reportedly the game is going gold soon...I am a skeptic, can't help it ;) , and quite frankly everything from Firaxis so far has first touted the graphics, then the MP side of things...with a "don't worry we are working on improving the AI" attitude. Well, I am worried.

All of the new features we have discussed here over the last 3/4 year sound wonderful (even though I don't like a few but I can live with them) - but as they say, the proof is in the pudding.
 
Well, the true 'game reviews' have only started to come out this week, and not suprisingly the reviewers were focussing on the most new element featured in civ4-namely Multiplayer. Over the next two weeks or so, I too hope to see more solid reviews relating to the MP experience.

Yours,
Aussie_Lurker.
 
Aussie_Lurker said:
Well, the true 'game reviews' have only started to come out this week, and not suprisingly the reviewers were focussing on the most new element featured in civ4-namely Multiplayer. Over the next two weeks or so, I too hope to see more solid reviews relating to the MP experience.

Yours,
Aussie_Lurker.
You mean SP, right? ;)

My concern is the developers have not made SP available to the reviewers yet...and I wonder why....being an oldster, and set in my ways, I really have no time or desire to play MP. My feelings are because I am feeling like a member of the abandoned segment of the 3 part demographics Soren talks about in his development strategy of designing games....the ones who get left behind in search of new customers.
 
Aussie_Lurker said:
Well, the true 'game reviews' have only started to come out this week, and not suprisingly the reviewers were focussing on the most new element featured in civ4-namely Multiplayer. Over the next two weeks or so, I too hope to see more solid reviews relating to the MP experience.

Nitpick: previews. These aren't reviews. These are Firaxis-conducted, -managed, and -controlled dog-and-pony shows. What and how much they see are up to Firaxis. Real, actual reviews won't happen until near the release date, when the media get CDs to install on whatever machines they want and can do whatever they want with them and actually spend more than an afternoon investigating the game. Until then, all we have is fragmentary knowledge, which is inherently unreliable.
 
Sorry, you are right Apatheist (and Oldstatesman!). Although they were a 'dog-and-pony' show, I do feel that there was a lot more 'hands-on' experience of how the game actually looks, plays and feels than in 'previews' for previous civ games at this point in the development cycle. That said, though, I am taking much of what is being told to us at this point with more than a small dash of salt ;)!

Yours,
Aussie_Lurker.
 
oldStatesman said:
the almost total silence on DETAILS of the SP side has me extrremely concerned.

Soren (lead designer on Civ4) is a man of few words. Have you seen some of his posts? :lol:

When they put the AI guy (read: single player expert) in charge of the whole operation, that speaks louder than anything else as to how important the company feels SP is. At least for me, that's been obvious (and implied) from the start.


We simply won't know what's in the game until we play it or get more complete reviews and information. In the mean time, I wouldn't read too much in to what is or isn't being said yet. The game will be out soon. (A matter of weeks now? Halloween, or thereabouts, the publisher claims.)

I'm looking forward to playing it. In fact, the hibernating grognards from Realms Beyond (Ozy is one of us) will be waking from our long slumber soon. There are Epic adventures waiting to be had! :cooool:


- Sirian
 
Sirian said:
When they put the AI guy (read: single player expert) in charge of the whole operation, that speaks louder than anything else as to how important the company feels SP is.

This is a very good point. Soren is considered one of the strongest AI programmers in the games industry.
 
Thanks, Apatheist, oldStatesman and Ozymandous, for pointing out and expressing quite some of my concerns.
I agree and second most of your comments in the above posts.

warpstorm said:
This is a very good point. Soren is considered one of the strongest AI programmers in the games industry.
Warpstorm, first one would like to know by whom Soren is considered to be one of the best and second, as far as I know he still is quite young, isn't he.
My experience tells me that programming (not from a pure technical point of view, but from knowing how the combination of certain algorithms will work) has MUCH to do with experience.

Now, I am not in the position to blame Soren nor do I have the intention to do so, but the conclusion that a good programmer for AI will make for a good SP game seems to be a bit risky.
As others and I have pointed out already, there MIGHT be a difference between a SP-AI and a MP-AI.
A good car designer may come to different results when his main task is to design a SUV than he would have come if he would have had to do a sportscar, right?
 
Commander Bello said:
Warpstorm, first one would like to know by whom Soren is considered to be one of the best and second, as far as I know he still is quite young, isn't he.
My experience tells me that programming (not from a pure technical point of view, but from knowing how the combination of certain algorithms will work) has MUCH to do with experience.

By the other programmers in the industry, of course. Okay, that's a small incestuous sample set. :)

By quite young, yes, compared to Sid (or myself for that matter) he's young, but he's getting up there for programmers in the industry. Most people burn out early from the stress (my company is very odd in that the median age is over 30). As far as experience goes, who has more experience actually programming an AI for a Civ-like game? Brian Reynolds at Big Huge is probably the only other person with more.

Compared to the top AI researchers in the field how is he? I really don't know. They haven't made any Civ AIs (yet) to compare them. With the SDK being released maybe we'll see some.
 
Commander Bello said:
Warpstorm, first one would like to know by whom Soren is considered to be one of the best

All you have to do is play a wide variety of games and look at their AI. Performance speaks for itself. Identify the best AIs, then extrapolate to the best AI programmers.


- Sirian
 
Take a look at his website. He's won awards for his programming skills.

As for how he compares to AI researchers... I think there's been a misunderstanding, here. Creating a game AI has very little in common with genuine Artificial Intelligence at the moment. You might as well ask how Soren's ability to program compares to the work of the top mathematical theorists in topology.
 
I predict that the Civ4 AI will be worse than the Civ3 AI. It's good to keep your expectations low. That way you won't be disappointed.
 
Don't forget that they're releasing an AI development kit. Expect to see any failings of the AI shored up within a month and fixed as completely as possible within 6 months to a year.
 
Mentioned this in a couple of other posts but i will re-mention it again. I suspect that
all the programmers are white males. (Don't know this tho) This means that the mind set is homogeneous. If they varied their programmers i wonder what effect it might have.
 
troytheface said:
I suspect that
all the programmers are white males. (Don't know this tho) This means that the mind set is homogeneous.

Most game programmers in the USA are white males. Based on my experience it does not follow that they have the same mind set. Some of the ones I know think very differently. (Some so much so that you wonder which alternate universe they came from).
 
i suspect ur correct there...but from a biological perspective it might be interesting to mix it up a bit-course i have met women tougher and more dogmatic than men and men who were "softer" than many women.
 
Back
Top Bottom