One of the first things we're going to be beta testing for Elemental is a new way of doing computer AI. As some of you know, the main reason I got into development in the first place was to write computer AI for games.
Elemental is an opportunity to try some new things. For example, the initial betas of Elemental are going to require the user to be on the Internet to play. The reason is that the computer AI won't be inside the game itself but instead be implemented as AI players on the Impulse game cloud that we're building.
So you would, in effect, start up Elemental like you would a multiplayer game except that the people joining your game wouldn't be people. They would be AI people. And the AI work won't just be in how they play the game but how they behave -- in the lobby and elsewhere with the goal to make them feel like real people (except not jerks).
These AI players will have their own Impulse accounts with their own standings in the rankings, their own achievements, etc. The code won't execute on your machine, it'll be all server-based.
The final version of the game, of course, will not require the user to be on the Internet. We will bundle a number of the AI players in with the game to play single player, off-line games.
But what I hope to do is to build a cloud of multiplayer AI players that I can constantly be updating and enhancing in near-real time based on the data I get back from them on how they were beat, what exploits were used against them, what chat messages were sent to it that it needs to respond to (I'll probably have it respond "no speako English" for the most part).
In the final version of the game, these AI bots will be out there ready to join on-line games. We will give players the ability to discriminate against the robots if they have some sort of anti-robot issue and don't want them joining their multiplayer games. But the match-making system otherwise won't know which players are controlled by humans and which are controlled by the robots.
The nice thing is that for people who are used to playing with strangers on-line, the robots will help fill out the on-line community with people to play against who are explicitly designed to play like people (well like decent people not the people who drop immediately).
I expect the first version of this to be included with Beta 1 of Elemental (i.e. this isn't some "future" thing) which we expect will be out this Spring.
Please let us know ASAP what the qualifiers will be so we can try to make sure we meet them.
Sammual
Brad,
I love the aim of the AI for Elemental... and that alone might get me to buy it. However, as a gamer I tire of both losing and winning. The entertainment is in the challenge and in emergent gameplay. So, while you have the ability to write an insanely powerful AI that knows my every strategy and its counter, what will you be doing to ensure the gameplay is a challenge and not an overwhelming defeat every time?
I'm no masochistic player, and all I really want is a decent challenge. Usually the AI is either beaten swiftly or overwhelmingly, or it takes a long time with a decent amount of thought to defeat them. To lose to an AI isn't so bad, if it's close.
So in the interest of not shaming players by being repeatedly beaten by masochisticly powerful AI opponents:
A lot of developers restrict their betas to North America only. That sort of thing always upsets the rest of us. I don't know if you were thinking of it, but please don't.
I'm pretty sure they wont. At least they've never done it with previous beta tests.
Imho it's much more likely that they'll look for active people who are willing to invest some time into early and unstable game versions and who will write lots of feedback about issues they'll surely encounter.
Support for this one. Please make the difficulty in single player adjustable in game, if possible.
I'm curious how this will work for saved games, and how multiplayer will work for saved games.
In my multiple years of playing Age of Wonders and Heroes of Might and Magic, seldom was a game completed in one sitting. In fact, the method for saving and reloading said games in such a way that everyone who re-joined joined back up playing the same players was a headache.
I usually played these games with friends, so we knew when we'd get together to finish the game.
How will these internet-based games work across multiple sessions, or is it expected that many Elemental games will be able to finish in one session?
I sincerely doubt geographic location will have anything to do with it. Traditionally, Stardock has never discriminated based on location
It's an interesting way of getting feedback from the game logs, probably better than what submit to metaverse could do in GC.
Unfortunately, not being able to play when not connected is a big no for me, as I'd be unable to play at all during wek-ends when I have the most time.
Somewhere along the beta, you may want to switch to a more open mode where it's possible to play locally.
I wonder how many people played the beta of GalCiv2?
Edit: quoting on these forums is so broken. I had to go find a different post of frogboy's, quote it, and copy the tag from that because quoting the OP caused Bad Things to happen.
It's funny, people have been talking about doing stuff like this for so long and finally a company is taking AI seriously. Count me in for not playing against humans - It will also be nice to play against an AI that isn't a human playing 24/7 but claiming to be a newbie.
This seems to be the core of your post. Much of your surrounding comments are about win-lose mechanics, but I agree that 'next-gen' game AIs should be include story/personality parameters in their decisions as much as possible. I want a single-player TBS game where the computer players are at least as much concerned with building a story line as they are with any given zero-sum game checkpoint.
Re the informal announcement that not all pre-orders might qualify for the first 'public' beta, I don't like the sound of it, but I can quickly start talking myself into the idea on grounds of the net-based AIs being a fresh notion that might need heavy work before we start seriously polishing what I still hope will be the heart of the game, offline single-player 4/5X.
But I'd also be happy to see a first beta that might have a 'select' crowd for the online components while giving all pre-order folks free access to the single-player, offline functionality.
Yeah, I know, I was just being paranoid, they've been good to us. Actually I assumed for quite a while that Stardock were German - doesn't make sense thinking about it now, but Galciv just seemed...German. Like Gothic or X, except easier to play.
man, I see so many uses for this. It is really awesome. (easy to help with play testing too I would guess)
Alright, I can't wait to play the beta with mr. robot
I'd also love for AIs to be able to recognize situations within the game, and respond/react/comment accordingly. I don't mean in ways that are directly aimed at making them better players. What I mean is it would just be nice if the AI occasionally commented on specifics of the current situation, and commented on things that have happened earlier in the game.
"Considering we just annihilated your army in the Greatwood, and are now marching on Fairview, we think it's time you talked peace concessions."
"We have never forgiven you for killing our great hero Sir Lawrence in the War of 1266, and now intend to take this opportunity to avenge ourselves upon you."
Obviously this would require a lot of context-specific conversational text be added to the game, but I would think that would still be orders of magnitude less than was required for the epic generator that Stardock once attempted to create for GalCiv2.
Along those lines, I'd love to see an AI that can be petty and spiteful. It may not make the most strategic sense to come after me, but he still holds a grudge against me for something I inflicted on him many years earlier. Or, conversely, an honourable AI remembers that I specifically came to their aid many years earlier, and will consider that in current circumstances (and mention it during negotiations).
This reminds me of something else. From screenshots, obviously terrain features can be given names (either by the player or automatically by the game, it's not clear which). Can wars also be assigned names? That would be cool, particularly if the AI would occasionally refer back to them.
If I'm not mistaken the names on the paper map are put in during map generation. Somewhere, I think one of the devs said that you will be able to put names into the map generator for different types of features, and the generator will then draw from those (and presumably from its own set of names, too if yours are insufficient).
Dunno if you'll be able to rename terrain features on the fly in the game. I kinda hope you can't, excluding some major event (maybe even player-caused event - like burning down a forest with a giant spell). But honestly it's not really a feature I'd want the devs spending their time on, unless it's a trivial matter for them to implement.
Will Elemental be multi-threaded to let the AI take advantage of multi-core cpu's? I've heard that the more cores, the more human-like AI can get. So, in the near future, we'll see 100 or even 1000 core CPUs that will be able to mimic human behavior completely.
Don't get your hopes up on anybody being "able to mimc human behavior completely" any time in the near future. Yeah, maybe in the near future we'll have the technological potential to do so, but I think we're still quite a ways from being able to program such an artificial intelligence.
I've never played GalCiv multiplayer for one reason: I typically only have a half-hour-to-hour a day to play. That's not so unreasonable a gaming period, but the setup issues that come with multiplayer games make such an option impossible for me in the normal case.
That said, the Impluse Cloud changes this scenario. If the game persists on the Cloud, and has a "please play at least a turn each day" requirement... Then when I start the game, I can just click the game to enter it, play my turn and go back to doing whatever. If other players are around, they can play their turns and maybe I'll be able to play another turn that night. It becomes more like a hosted PBEM than a multiplayer game (unless all of the players are logged into it). And that I could do. Especially if the game made it possible to play multiple such games at once.
I hope this is the plan... Otherwise I won't even be able to use the Beta.
I imagine that like all public betas, the more time you have to volunteer, the more you'll be able to contribute to the initial dev effort. But I don't think you need to worry about being excluded from the 'full' beta if you can't commit to a heavy multiplayer schedule. It sounds like the server-side stuff might need what amounts to a 'selective pre-beta' so the devs can work with a small group to do some initial polishing for the online parts of the game. I'm sure they also want plenty of input from folks who are interested in single-player, and hopefully long-term "hosted PBEM," scenarios.
Just don't log on as Sarah Conner and you'll be fine...
So whats going to happen to me (the player) if we get disconnected from the AI Bot servers?
that may be true. But you can get a lot closer than most strategy games do. Even just with a crazy long series of if statements (if this then that logic for programs)
See now, I almost want some of them to be jerks too. And I don't want to know which ones will be, but let there be subtle hints (rather than random attacks). Things like the AI saying when you tell him to bugger off his army outside your capital "oh, don't mind them. They are just passing through" when they are obviously not going anywhere. The not-jerk would be like "sorry, I'll move them out right away." Then the super jerk would say "sorry, I'll move them out right away." and continue to advance. Then when he declares war he will say "oh, I guess I lied. Sucks to be you. I hope your city burns as well as it looks".
I like AI taunts
so... 'IN URE MP, STEALING URE SLOTS?'
That is to say, will an AI joining a game with 7/8 players (for example) cause that game to be full, so Joe the player in the lobby sees it as full and can't join?
Also, jerk AIs would be fun. Its not as fun to revel in your glorious victory when your opponent was a nice guy (or an AI in disguise) as it is when your opponent has really been a pain in your butt the entire game. If you are being told in a very undisguised manner by the AI "You suck, and I will own you every time," taking one of its cities will be much more satisfying than if the AI is either completely polite or attempting to suck up to you. As long as to much harsh language isn't used, jerk AI could add a very interesting aspect.
Also, having one or two which are overtly AI would be funny... as in AI which address you as if they were HAL 9000...
Try and make them as diverse as possible. AI ranging from homicidal jerks who raid you constantly and dig up your cememtaries just because they can down to passive, slightly neurotic AI who focus on economy and technology until provoced... One of the most distinguishing factors in today's game AI which labels it as AI is how homogenous they are; every AI in most games will react to stimulae in similar, if not identical, ways. Try to avoid that if possible.
Key point right there. No strategy game comes close to actually behaving like a human. Getting a lot closer is only one more small step . And I'd argue that a crazy long series of if statements* would result in an AI that is about as predictable in a given situation as any of the ones we have now. To make an AI that behaves like a human, it has to be adaptable. It has to be in the same exact situation multiple times and respond differently every time. And it has to make mistakes, to boot!
No, while I think that we probably are getting real close to AI's that can present a really strong challenge in a strategy game, I think we're still a ways off from any AI really acting human.
*Also, if by a series of if statements you mean something like "if () else if () else if () else if ()..." then from what I understand of parallelization, 100 cores wouldn't really be able to execute that much faster than 1 core. On the other hand if you mean "if () if () if ()" then sure
I haven't been very active, so I'm sorry if this is blatantly obvious to everyone else. Will there be singleplayer? This thread almost makes it sound like there will only be multiplayer outside of the campaign. I can't imagine a real time strategy game requiring an internet connection, so I'd be shocked if that was the case.
There are many great features available to you once you register, including:
Sign in or Create Account
