Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
Content Generation and AI
Aruvia
Member UncommonPosts: 86
One of the first things I look at in MMO's is how content is created. As example:
A game like WoW uses very static content, it is generated by the developers and then released to the players as is. Quests, Areas, NPC's are all laid out in a set path of progression that only changes when the developers make a change through an update or expansion.
Archeage starts with a very similar set up for the first 30 levels or so. but after that becomes something very different, while the world and locations and areas remain static the content is driven by PvP. PvP is the focus of everything post 30 gathering, trade runs, exploring, sailing, Piracy are all rooted in PvP which adds a dynamic element to all of these. Even housing and auction are PvP-centric from competition for land ownership to owning land in contested/pvp areas.
Guildwars2 I think is technically static though the "Dynamic events" do branch based on player interaction the locations are static and the loops are fairly short so repeat often. the developers also release Episodic updates regularly which kinda adds a dynamic element.
What I am really hoping and looking for is dynamic AI, that is, smart enough AI that it can create content and things for Players to do. that it can recognize needs and wants and recruit players to help, or better, AI that was smart enough to carry out their own desires.
Imagine a Troll chief set on domination and an extreme hatred for elves. What if the Troll was able to see the needs of his home and people. need to fortify his walls. he passes on the need to another NPC who then begins to oversee the fortifications. The Troll could then solicit the help of either Players or other Troll grunt NPC's to gather the materials for the fortifications. and when enough was gathered this could actually add walls to the front of the city, or new towers, or whatever. Now what if the Troll chief decided his city was safe enough and was ready to go on the offensive, he could recruit an army of warriors NPC or Player and begin a march on the Elven city. Now the Elven leader needs to decide what the best defense is. is the city strong enough to repel an attack , is the best defense an offense and attempt to take out the trolls before they get to the city. how many soldiers do the elves have how many players can she recruit? Now imaging these things could all play out with or without players. how long it all takes and the outcome could all be affected by players but I want the world to live and go on without me. Imaging logging in to find the Elven city is almost destroyed, the Elven leader is imprisoned and you as an elf are now little more than a slave to the trolls, what things would the trolls have for you to do, what tasks maybe the Elven leader ask of you if you visit her in prison, maybe you could help her escape to an outpost the surviving elves have created when they fled. Imagine what the Elven leader would need to do to begin rebuilding her nation, to build the forces to retake their capital.
So far no game exists that could do this, and the tech may just not be there yet. Or perhaps I am alone in wanting this type of world.
I look at what StoryBricks and EQN were trying to do I see a glimmer of hope. I would imagine if a game at least takes a step in this direction with any success it could encourage other developers to expand on this type of content generation.
This is, in my opinion, exactly what the industry needs to move forward. It’s the area that has seen the least innovation (if any)over the years.
Intelligent characters that live in the world that have wants and needs, npc's that remember the actions of characters and react accordingly, npc's who interact with each other with emerging behaviors that generate content for players to interact with. This creates content real-time, always changing and evolving, as players fulfill needs of npc's and in turn may negatively affect other npc’s and where npc's with differing drives and motivations come into conflict and that conflict generates more content for players in an ever changing world. Suddenly we are not as reliant on the developers to spend vast resources creating content for us to chew through in a matter of weeks or days. nor are we reliant on the whims of other players and populations to generate PvP content. It does not mean that neither of these will exist, both pvp and developer content can and should exist but potentially in more meaningful ways. while the npc's and the living world continue to provide content dynamically in a living breathing world.
Comments
You would need one very complex system or scripts and database to make it happen on such a large scale.
In theory sounds doable,only until you realize the amount of horsepower needed for that kind of AI.
This is why devs stick to simpleton cheap ideas,they have the gamer's fished into buying their half ass effort so of course they will keep selling it.
Even if gamer's said enough is enough,step up or we quit buying your crap,somebody would still break rank,cross the picket line and buy it.
Then if not one person bought that game,they would simply try a new game,new genre,new angle.
Yes if i was in charge of a game,i could eventually work something decent out,but would take a lot of work and a lot of thinking and from what i have seen,Directors seem to have very little ability to think.I believe 100% gamer's could design better games,however Devs would make the excuse that our ideas cannot reasonably be done.Of course i say BS to that.
Never forget 3 mile Island and never trust a government official or company spokesman.
While I get what you are saying, Tech is changing and with things like SpacialOS as an example the horsepower is becoming a non issue. But like any R&D it would take time and recourses to develop this. And AI has been one area that no one has felt the need to spend time and recourses on for mmorpgs.
While I am talking long-term, big-picture. It would not have to be some revolutionary leap; it could be a smaller evolutionary process. Start with simple things like orcs packing up and moving if killed too much.
I don’t know maybe borrow some elements of RTS AI for controlling a clan of goblins. Where they gather recourse, defend and build their camp to an outpost to a town etc. and if they get wiped out they try again somewhere else.
Or similarly with elves, except they can pass the tasks of gathering, defending etc.. to either players or to non-players.
In another 20 years, may be. Right now, I am more than happy to play professionally created stories by devs.
I get what you are saying. there is something great about playing through someone elses story, that is there from single player RPG's to games like swtor and wow. and I am sure many more to come. it is something many people myself included enjoy.
and I am not saying to immediately, if ever, end dev created content. though it could progress to that point. I am talking bringing more life, more simulation into MMORPG's, expanding on the "Virtual World" idea rather than the static story.
While the CRPG and MMORPGs can be a great medium for telling stories, I want personally something more. I want to live a story by participating in an ever changing world, I want the story to be what I have done in this virtual world.
again though I am not expecting something Revolutionary.
I am hoping for something evolutionary.
What you seem to really be asking for is a computer game that does things that it simply isn't practical to program computers to do. There are a lot of types of procedurally generated content (e.g., mob placement, attributes, and battle tactics) that it's at least plausible that it could be done well. Procedurally generated storylines that don't just consist of copying and pasting chunks of hand-written storylines aren't one of them.
I don't see AI for individual NPCs being able to go too far yet, as you would like. But I do see the ability to give NPCs an immediate "goal" and then have a random but weighted "roll" on what said NPC does from there.
I'm talking mainly about combat. Under various situations and NPC might heal, change attack strategy, or run away. By using this idea, the NPC would usually do what makes the most sense, but might do something different.
But the same sort of thing can be done for NPC's just wandering around. They make decide to make a home in a cave. They might later decide to move, as you suggested, because of too much danger.
Then this can be taken to the next level of Group AI for social NPC groups.
The world could be built to further this. I think of this as "Dungeon AI", but it could just as easily be in particular area (like a ruins, a clearing next to a lake, etc.).
The idea being that such areas give more "choices" to the NPCs.
Once upon a time....
While I have nothing against GW2 as it is perhaps the only one to attempt to change things. Perhaps if you read it all you would have a better understanding of what I mean or perhaps If I found a better way to explain it...
If devs could pull it off we would have it. Remember the old axiom of AI: "if it work, it is no longer AI".
Epic Music: https://www.youtube.com/watch?v=vAigCvelkhQ&list=PLo9FRw1AkDuQLEz7Gvvaz3ideB2NpFtT1
https://archive.org/details/softwarelibrary_msdos?&sort=-downloads&page=1
Kyleran: "Now there's the real trick, learning to accept and enjoy a game for what it offers rather than pass on what might be a great playing experience because it lacks a few features you prefer."
John Henry Newman: "A man would do nothing if he waited until he could do it so well that no one could find fault."
FreddyNoNose: "A good game needs no defense; a bad game has no defense." "Easily digested content is just as easily forgotten."
LacedOpium: "So the question that begs to be asked is, if you are not interested in the game mechanics that define the MMORPG genre, then why are you playing an MMORPG?"
I was!
I mean we can see a sort of semblance of what I am looking to start with in an RTS like Warcraft or StarCraft as just an example.
WC example: assign random starting locations from presets on map. peons are given specific task go to mine interact got to town hall interact. go to trees interact go to town hall interact. build tower at location ...
or something like that. each npc follows very basic ai. the computer "player" has a larger macro level awareness of the game rules and objective. and assigns tasks accordingly.
Why could this computer controlled "player " not control a clan of goblins in an mmorpg. where the lose condition meant selecting a different location and starting again potentially very far away where the chances of building a full city may be higher. this of course would need to be modified for the world and for the game
you could then have many of these Controllers for many different groups including the ability for the "controller" to give tasks that a player could accept and do, it would not be much different than the tasks currently given in many mmorpgs of say kill ten rats or gather ten logs or what ever. this would be a start. and not the elimination of other types of content.
When someone makes one that is good enough, it will change things. We may have to wait.
I actually wrote something along these lines when I worked at Symbolics long ago. Random quest to dungeon generator. It wasn't good enough for my taste as I wanted something that looked hand crafted. It also suffered from too many parts looking similar.
Epic Music: https://www.youtube.com/watch?v=vAigCvelkhQ&list=PLo9FRw1AkDuQLEz7Gvvaz3ideB2NpFtT1
https://archive.org/details/softwarelibrary_msdos?&sort=-downloads&page=1
Kyleran: "Now there's the real trick, learning to accept and enjoy a game for what it offers rather than pass on what might be a great playing experience because it lacks a few features you prefer."
John Henry Newman: "A man would do nothing if he waited until he could do it so well that no one could find fault."
FreddyNoNose: "A good game needs no defense; a bad game has no defense." "Easily digested content is just as easily forgotten."
LacedOpium: "So the question that begs to be asked is, if you are not interested in the game mechanics that define the MMORPG genre, then why are you playing an MMORPG?"
I could see it being useful to get the foundations in, then human content creators could come in an polish it up, speeding up the content creation process.
EVERY single script and words that come out of the npc's will be repeated and just a long list in some database.Could they somehow have npc's recite the entire dictionary and form phrases and sentences,i very highly doubt it,but that would impress me.
Although would not impress me,i would applaud at least an attempt at effort but remeber ,loading a database with numbers and words is a LOT less work than making thousands of assets and hand crafting a world and creating systems,grouping mechanics,xp mechanics,quests,sounds,textures etc etc etc.
Never forget 3 mile Island and never trust a government official or company spokesman.
again I hear what you are saying, but what do we have now? an npc sitting in one place saying the exact same thing over and over to every player to present a static story. while this is entertaining perhaps for the first couple times through it is still just a single npc that when interacted with gives a screen of text out of a database.
I found the link a very interesting read and a great direction to try. But I was not referring to dynamic generated narrative. though that would be a great addition and it is something I really like.
I was referring to simulated behaviors more than dynamic narrative. see my above posts. more about the virtual world and the story I live, through my actions and reactions to the world around me and less about following a specific narrative.
I was imagining much smaller steps of just giving npc's better behaviors that slightly mimic life at first.. like my above examples. bringing more simulation into mmorpgs.
what's presented here, while similar to what storybricks was trying to do and I am so glad there are people still working on this type of system. it is a huge leap forward. which is great but I fear a long way off.
Again you are either not reading or just not understanding, I get you like static written narratives. that's great I am glad you have a lot out there to play that meets this enjoyment for you. To follow a narrative can be enjoyable and exists in plethora, Again I am not referring to narrative.
You seem stuck on watching a movie or reading a book type of narrative story telling, I am referring to the story I experience in the game actions and reactions to the world.. my gameplay story not written anywhere and not spoken by an npc static or random or whatever, I am referring to the adventure of interacting with a world that is more lifelike than currently exists.
yes at this point I can log into a game, walk over to a goblin camp and kill the goblins. Or I can be told to by a quest giver, whatever.
Rinse and repeat
What I am suggesting is that the goblins be smart enough to leave if they get wiped out. Or to get reinforcements whatever.
I am suggesting that maybe I walk out one time and kill goblins at their camp. The next time I or someone else comes along they are no longer there. But on the other side of the mountain or forest or whatever someone walks by and sees them and maybe gets killed by them. They are allowed to live longer and begin to build up from a few tents to a large outpost.
Now at this point a story is being told with no narrative. No text. It is playing out dynamically through actions.
Again I enjoy narratives I enjoy the stories people tell, if you have not looked at it already you may be interested in Shroud of the Avatar, Richard Garriott has some good story telling skills and Tracy Hickman has written some great novels. It is going to be a great narrative driven game I would imagine.
But once more that is not what I am talking about.
There's an additional vector to imaginative thought. This is the intangible nature of PvNP. Crossing or colliding two seemingly unconnected symbols to one idea, as adding wings to a "letter A" for an advertisement slogan, is not something we can teach a machine to do, primarily because we don't entirely yet understand how or why we do it. It's not that machines don't "want" to dream, it's that we don't know enough about why we "want", to teach something else to do it. It's quantum entanglement of completely insubstantial substance.
I don't think there "is" AI. I think there's something with which we amuse ourselves to "call" AI. I think there will, eventually, "be" AI, and it will be very strange and dark. Watching something it makes might be more disturbing than the most intense, grotesque psychonautic experience. We might be better off without it. Terminator would be nothing compared to it, because I'd rather be hunted by murderous metal golems, than have my world perspective mutated to some stark, twisted, version of electric, ersatz hell.
The term AI kind of died off with AI Winter and is now called by other specific technology names. It didn't go away, but there was marketing reluctance to use the term. It seems we are seeing marketing returning to use it though.
Epic Music: https://www.youtube.com/watch?v=vAigCvelkhQ&list=PLo9FRw1AkDuQLEz7Gvvaz3ideB2NpFtT1
https://archive.org/details/softwarelibrary_msdos?&sort=-downloads&page=1
Kyleran: "Now there's the real trick, learning to accept and enjoy a game for what it offers rather than pass on what might be a great playing experience because it lacks a few features you prefer."
John Henry Newman: "A man would do nothing if he waited until he could do it so well that no one could find fault."
FreddyNoNose: "A good game needs no defense; a bad game has no defense." "Easily digested content is just as easily forgotten."
LacedOpium: "So the question that begs to be asked is, if you are not interested in the game mechanics that define the MMORPG genre, then why are you playing an MMORPG?"
The biggest problem with creating a more complex AI to manage such things is how complex it has to be and how much detail it has to track and respond to on top of the regular game world functions. It's basically making the game analyze itself to find the next step with thousands of people constantly messing with the data.
It's still a good ideal to work towards, just it takes time to overcome hurdles that limit both the servers and the client interactions.
There are other more actionable choices that could be implemented though.
One example being "open nodes" in a semi-static environment which are for the most part the same as you'd expect as a normal node for spawning mobs, except that it does not have a specific thing it spawns.
Instead, the game world is littered with these nodes and uses a library of NPCs, mobs, doodads, and quest events to procedurally generate a variety of activities. As a standalone mechanic it would be something to randomize user experience, but can also be tied to themed libraries or a pool of quest-specific events so that players can still obtain a more free-form experience out of a technically static system.
"The knowledge of the theory of logic has no tendency whatever to make men good reasoners." - Thomas B. Macaulay
"The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge." - Daniel J. Boorstin
http://www.newyorker.com/tech/elements/ive-seen-the-greatest-a-i-minds-of-my-generation-destroyed-by-twitter
Have much bigger budget then GW2 did, employ much more coders, make 20 x more scripts than GW2 did and propably do 50 x more debugging hours than GW2 team did. Then propably have much stronger server hardware to have all those CPU time to process all those scripts constantly.
Then you could make game world have a good illusion of being 'alive' and dynamic - well for few weeks anyway until content would get used and seen by playerbase.
Until we get 'real AI' that could make this stuff instead of humans, it is "just" a matter of budget. Same issue that is in all MMORPG themeparks when players complain there is not enough dungeons released and complain that old dungeons get reused, etc It is "just" a matter of more budget and more people and more project managing.