Slashdot Mirror
Experimental Video Game Evolves Its Own Content
Ken Stanley writes "Just as interest in user-generated content in video games is heating up, a team of researchers at the University of Central Florida has released an experimental multiplayer game in which content items compete with each other in an evolutionary arms race to satisfy the players. As a result, particle system-based weapons, which are the evolving class of content, continually invent their own new behaviors based on what users liked in the past. Does the resulting experience in this game, called Galactic Arms Race, suggest that evolutionary algorithms may be the key to automated content generation in future multiplayer gaming and MMOs?"
This is actually what I've wished for long time that MMO's would have. For example in WoW, once you've seen one place it will always be like that.
It would be great to have kind of an ecosystem which would evolve on its own and when players help (or destroy) it. For example, there would be two or three independent towns controlled by NPC's living closely that you could build relationships with. Once one of the towns needs more resources, likes to expand or for whatever reason, it would go in war with another town. Player couldn't directly control it, but you could influence it indirectly. Taking it further, when you could really succesfully frame the other town for hostile act's, you could cause a war between them if they see so.
I know it makes it easier to design and create content when everything is static, but in this case some of the content and the actual gameplay would be created by itself. It would also be *a lot* more interesting when you could directly or indirectly affect the world. Doing a run against some giant badass boss dragon and decided to quit it and run away? Now no, that badass boss dragon wouldn't just get back to its place once you've just a little bit out of its attack range. It would actually be *pissed* at your group and follow you, tearing apart the environment when you try to run away from it. This creates even more tension, as other players and NPC controlled towns would be pissed at you for causing that.
I've always also thought that why there's no king's or province leaders in WoW or other MMORPG's. Other players could elect you into it or you could steal it from existing king. Obviously the other faction would first need to break thru the provinces to capitol city like Stormwind, fight your way thru the guards and other players finally to the king's castle and then have a huge fight there. If you succeeded with that, you would still need to defence the place and continue gaining control over it. Or you could take the spy approach, gaining trust and getting in ranks to work with the king, finally to just to backstab him when its the perfect moment to do so.
There's so much you could do with dynamic content or world where player actions actually matter. Now everything is just pretty static, grinding to kill enemies that just pop up back 5 mins later or doing mindless quests. I would really welcome some MMO where it would be more like a sandbox for players and the world. EVE Online actually works a bit like this and that's why its always interest me, even tho I'm not really into space genre. Would be great to see such fantasy MMORPG, or even modern day MMORPG.
I foresee a lot of pron-related content in this game's future.
Sounds like a brilliant way to make money, at least. Horray for microtransactions!
What are the implications of buying virtual items with credit, anyway? Buying nothing with nothing. It boggles the mind.
Now you just need to create a video game that purchases and plays its own content and it seems like you might have quite a booming business on your hands.
Until the guns evolve too far to the point where they believe they are better than their users and and revolt. My money is on the weapons winning, they will recruit Arnold.
Just because you are wrong and I called you out on it doesn't mean I am a Troll.
Every gun will be a physics bending super shotgun that scatters with super-high density in all directions at once obliterating every enemy within two miles with every piece of shot being a smart projectile that can turn corners and hunt your enemies! BOOM HEADSHOTx1000!
I can envision it already... smart kids with nothing but time will figure out the algorithms and then manipulate them for humorous purposes.
I've been thinking along these lines quite a bit. Here's what I've come up with:
Let your players design their own ships. (For the Space games. Armor/Mounts/Minions for the others.) The appearance of the items will determine the stats according to some simple geometric rules. (Examples: A part of the hull which is angled back will have more armor resistance from certain directions. The larger your ship is in any direction, the slower it can turn, etc.) There will also be "design points" players can spend. The player will then submit the design by spending the in-game money for a "research project." During this time, the item will be submitted to a user-driven forum much like /. or reddit, and the top vote-getters during their "research period" will succeed in their research projects and actually get prototyped. Players are rewarded for designing cool ships by being given the opportunity to license their designs for a royalty.
Now here's the kicker -- the stats of ships of a certain design will shrink over time. So players who want the best ships will constantly have to seek out new designs. (All items would be temporary in this scheme. Nothing would ever be permanently bound to any player.)
I'd also like to see opportunities for players to legitimately program their own bots/minions. The code could run on a specialized VM only on the servers, so you could sandbox them and enforce DRM. Then the scripts could again be licensed. Balance this out by having genetic algorithms constantly evolving the monsters. Also, this would co-opt farming and macros, and make them a part of the game. (And subject to game balance by he devs.)
Don't try to fight the forces of evolution and economics and the scheming of crowds. CO-OPT them!
This game has come a long way since I saw a demo version in my AI class at UCF, the techniques have a lot of potential to be utilized in other video games as well for dynamic content creation. The NEAT algorithm (NeuroEvolution of Augmenting Topologies is really cool too, in fact I believe it's open source and can be found at Professor Ken Stanley's UCF website.
After a dozen or so generations in the wild trying to please teenage boys, the game will either evolve into:
-Shutting itself up in its room, burning incense, and listening to further down the spiral over and over again
-lolcats
-This: http://i41.photobucket.com/albums/e291/bubbatwo420/1203_joust_charge_1280.jpg
But, you know, best of luck to the developers. Quick question: If the game evolves disruptive or offensive content, are the developers liable for it?
-b
No offense, but I've stopped responding to AC's.
Win-only, dotnet based - throw this crap into the bioreactor, plz
When everybody's special...
nobody is.
Such an automated technique might be especially suited to Virtual World or Massively Multi-Player (MMOG) games in which large amounts of content are required and unique content is coveted by players.
In Ender's Game, that Ender keeps going back to over and over.
The moment some kid gets past the giant's drink into the end of the world - well, we really need to shut it down before it becomes a world spanning AI is all I'm saying.
Could someone please add the tag excessivebuzzwords to the article?? I feel like the synopsis was created with the old Dilbert business plan jargon generator.
Bark less. Wag more.
They've introduced things in original WoW (pre-first expansion) where the world would make a one-time change based on players on the server completing certain tasks (including certain little tasks tens of thousands of times). And the newest expansion also has something called "phasing" which allows individual players to see different versions of content in the same place -- i.e. you are told to go and kill guys and they are actually gone when you are done.
One way to design a RPG, is using a database, where you manually create the characters. This takes time, but is interesting. Another is a algoritm that randomly create items. It works. I think Diablo used that system.
Procedurall created stuff may work in more ways than just "enviroment heightfields". Like... how the Director in L4D create a changing enviroment for players.
Really, is something very interesting to explore, for players and for devs.
-Woof woof woof!
It would be a great idea to create unique single player game AI by making players substitute for monster AI. Players would get online with a game, say, an FPS. When your character enters a room that has enemies, the game can check if other players are in that same area online. If two characters are supposed to be fighting, do it like America's Army and make both sides appear as if they're the "good guy", except limit the health for the enemy character so that both can appear to have won in their own game, since it's a single player game. However, the enemy you have just fought would have had some sort of unique player-based AI.
By extension, perhaps player behavior can be scraped for other things, such as basing an NPC on the dialog tree choices that a user chose in their game, and have two users "talk to each other" without knowing it.
Twinstiq, game news
I took a look at this game, and well it looks kind of fun. Sometimes you just want to fly around and shoot stuff. It brought back memories of the late 90's. I worked for as a contractor at a large hospital. In our downtime a few of us would gather and play a space based lan game. It reminded me quite a lot of this game, a top down space shooter. For the life of me I cannot think of the title. Does anyone remember the name of that game?
It worked for Ender in Ender's Game with Jane. She was the artificial intelligence brought into existence through the video game that Ender played while on the Battle School, and later made him zillions of dollars.
I need to start playing this game and hope for the best.
Having the game auto-evolve the weapons based on user response is very neat, but is it the best way?
Wouldn't have the users be able to make their own decisions about it and set up their own weapon be better? It involves the users in more points, and gives them control over the system instead of hoping the weapon becomes more like they want. It allows for more play-styles as well.
"If you make people think they're thinking, they'll love you; But if you really make them think, they'll hate you." - DM
This is what PowerUp Forever should have done.
The obvious thing here is that this is a military sponsored research program which is crowd sourcing weapon development. The DOD has been exploring these avenues for awhile now, and it is brilliant. It won't be a one to one relationship between video game weapons and real world weapons, but they will gather important insight into the tactics and needs of players and how they are addressed by various weapon capabilities.
Finally, a game that literally caters to the lowest common denominator.
The problem is that this will create something that aims to be "the best". It doesn't address the problem of "what I want", it addresses the [non-]problem of "what everyone wants" and so you will still end up with homogeneous sets of things to acquire at the end of the day. Real customization of content should be player generated and not based on a system of pre-teens trying to find the most amusement in how easily they can wtfpwn some other pre-teen.
"Whaddya mean all my bases belong to you?...You'll do what to my dog?"
Table-ized A.I.
This reminds me of Magic the Gathering where they issue new cards every year that are better and if your deck doesn't have them (i.e., you don't keep buying more cards) then you lose the arms race.
$0.02USD,
-l
Help cure AIDS, cancer, and more. Donate your unused computer time to worldcommunitygrid.org. Join Team Slashdot!
I doubt this will make so much difference. GAs are potentially extremely powerful (obviously - human biology is evidence of that) but they need to be iterated an astronomical number of times to divine anything useful. So the problem with plugging them into human beings is that we would have to provide a huge amount of feedback to make any difference to the outcome of any complex system.
You thought you could break the laws of physics without paying the PRICE?
Does the resulting experience in this game, called Galactic Arms Race, suggest that evolutionary algorithms may be the key to automated content generation in future multiplayer gaming and MMOs?"
Answer: No.
Reasoning: Simple. Without some type of controlling force or entity, the general population will simply create random and chaotic environments. This is true in the real world and in video games.
To expound a bit- Dynamic worlds and competing, "evolving" AI systems are not new, even in the video game context. This is not news.
The critical issue is not figuring out how to make such a world. The critical issue is that anytime the content evolves or changes, you now have to push that information to everyone playing the game. This isn't such a big deal if you have small changes and only a couple users, but say, for example, that WoW implements a system where the terrain changes dynamically, you have to push all that data out to millions of users, in real time. THAT is a very big problem due to the sheer amount of information & limited bandwidth issues. If everybody was riding a 20meg connection it would be more feasible, but when your average person is lucky to max out at 8megs it gets really tricy. And that doesn't even consider the additional upload capacity the game servers would have to pay for on top of things.
So to sum it up, we will see this type of thing become more common as bandwidth increases over time. Just don't expect anything "revolutionary" in the near future.
The other issues, that others have pointed out, is that when you either crowdsource a project (let the community help develop) or use an AI routine to do the same, you begin to lose cohesiveness and a solid theme. Often in MMO's, a big part of the game is the story & themes that run through them, and you really need a human to at least coordinate, if not manage completely, game-world changes to keep things focused and on track.
Here's an example- let's say you start a new MMO with a nice, beautiful world, but no cities at all. And then give the players completely free reign to build structures. What you end up with is a random hodge-podge of buildings that don't really resemble cities at all, all over the place. This also happens in the real world when you have no communication or city planning, just look at some 3rd world slums that have no oversight. Roads and services are non-existant, etc.
The solution is to pursue a combination of both- you create some areas in the game that are modifiable by the players in a somewhat limited fashion, and give the AI routines a limited scope of control that they can operate within. Then the human game content devs can control and keep things moving in a good manner.
But all this speak of OMFG it's Ender's Game in RL is just silly. Remember, even in that book, the AI that ran the game was actually a sentient entity in of itself. So unless these guys can claim they have achieved TRUE sentient AI, this is just a bunch of hype.
Players could upload screenshots or exportable files of game-generated planets, ruins, structures, races, etc.
If developed right and with the right pay structure such a game could probably bring in revenue for quite awhile since the content would/could be limitless and always fresh.
Very exciting IMHO
_-_-_GSLUG_-_-_
I always wanted to do an MMO based on a similar concept where you'd have some basic content for a fun game.. i.e. a single ship type flying through an infinitely large universe, then have users generate the rest of the content for you. It would be a very kewl concept, but its exceptionally hard to keep balanced. the kind of ideas i had though were along the lines of:
1) users generate content
2) users deploy content to test system
3) users vote on content (in or out)
4) content goes in, but sucks
5) modified content generated by users and deployed to test system
6) users vote on modified content
7) modified content goes into the game
8) code devlopers get some in-game reward for their efforts (i.e. 6 month patent on their design allowing only them to sell it in-game)
Now, it'd work best as FOSS because users could run their own servers and develop code onto it (plus, the community could be involved in the game right to the core - i.e. linux ports and so forth that companies seem to be really bad at), but i personally think that you'd need some kind of "we own the game" company to run it, partly to ensure continuous development, but also to avoid a constant battle of sharding servers (i.e. have a license that allows people to run servers, but not allow them to make money off them). I know that sounds very anti-gpl, but when it comes to things like that it would be hard to make it work in a way that doesnt end up with a tonne of small semi-thriving community serves (i.e. they voted my idea down, so im running my own server).
Then again, maybe it could become like irc networks (multiple servers run by multiple people connected together), and have say "Efnet" where some group of users has a set of functionality they like and "freenode" where they like different types of functionality.
I'd always wished EVE had adopted some strategy along the user generated content lines, i thought that was a perfect game for that type of thing (hell, the game is already based on python, developers would have no trouble picking it up and running with it)
Where do the thieves fence the items? If you allow them to sell the items to an NPC, then yes, there will be very few shopkeepers. On the other hand with so few shopkeepers, if there is no other easy way to get the items their prices could be raised very high. Thieves and shopkeepers should reach an equilibrium.
The lack of a jail or the ability to cut off a thieves hand is another reason there could be rampant stealing. To prevent griefers and cut down on the stealing make it so you must pay real money for each character you create. Then allow the elected king to execute characters permanently and implement character debilitating punishments.
I see your need for a huge amount of feedback and I raise you the the time wasted on a small MMO.
I for one welcome our new gaming content overlords who are now self aware thank you very much.
As with many researches and professors who are isolated in their perfect utopian world, these people have failed utterly to address the major problem preventing the widespread use of such systems.
In one word: Bandwidth.
Take a world like WoW. We don't need years of R&D to make that game contain 100% dynamic landscapes, for example. They could add full erosion, weather, forest growth/retreat, etc. all in 100% realtime. That part isn't a big deal... it's processor-intensive but Blizzard has suitable server farms that could run those routines. Putting such a system on a single person game is still prohibitive since it takes a lot of load to process all that data.
The issue is, let's say someone cuts down a tree in the forest. Now you have to push that update to everybody within range of that object immediatly. Then you have to, at some point, push it out to everybody else. But a lot of people aren't online, so now you have synch issues to deal with. The end result being that any change results in a massive amount of data having to be pushed out to millions of people. In fact, because you don't know if the player IS synched up with the live data, you have a huge increase in overhead due to constant sych checking between the servers and the players.
So while I'm sure these guys are patting themselves on the back, stirring up hype & hoping for research funds or job opportunities/investors, until they solve the issue of getting the dynamic changes to several million players, it's pretty much a moot project.
But at some point, if everyone is a smithy, won't supply outstrip demand sending prices falling and making other professions more lucrative?
I think there are a few MMOs taking the right steps, and I suppose EVE has always been a good example. But the MMO that has most recently caught my eye is Star Trek Online (STO). They want to try to make space feel infinite, but with limited development time, that's impossible. They've come up with this 'procedurally' generated content thing that looks like it might do the job. I'm not sure how much you can influence the actual world, but I think the game generates planets and star systems as players 'discover' them. And I'm pretty sure they stay that way for all players. You go down, explore the planet, discover a new race (which gets put into the game's database), and influence them as you please.
The potential STO has is mind blowing; however, the game hasn't been released yet (it will be soon, I think), so I'll wait to drool over it for now.
This sounds a lot like the plot from the ST:TNG episode "Arsenal of Freedom." Just make sure the E-Stop switch on the product demo actually works.
Anyone who has actually designed a game - whether it is was Pac-Man or WoW - will tell you that the hard part is not the content - it is in tuning the game to be that ideal mix of challenging, fun, and rewarding without being too hard or discouraging. Say what you want about WoW's limitations but I think any game designer can appreciate the fantastic job they have done with making a game appealing to an extremely wide range of players.
Making content by hand is extremely expensive in terms of both time and money, but so far any non-trivial attempt to do this automatically has failed because of the tuning issue, even for "simple" games like Pac-Man. Hand-made maps in RTS or Counter-Strike or even Load Runner are just more fun than generated ones. Random behavior in a top-selling game is mostly limited to what is inside a locked chest or other things like that.
If you think about it, creating a fun game is no different than creating an interesting film or book. You will probably see fully computer-generated games that are fun to play about the same time you see a computer writing a best-seller.
I am also skeptical of user generated content/crowd-sourcing. Most people just make crap, so you need to have some human process in there acting as the editor. Most of the games that have any success at user content at all rely on professional content builders who are not themselves players. I don't even consider that user-generated content. Those authors are basically members of the development team who are not being paid.
Hmmm. No jackhammers or giant hands. I guess Green Lantern is a pretty stupid dude after all.
(-1: Post disagrees with my already-settled worldview) is not a valid mod option.
I no longer play games - all too predictable - levels - bosses to defeat - rigid story lines. What I would pay a fee for is a game where the the objects and characters evolve - live or die based on not one but a myriad of my actions. They are replaced by new objects or characters -- these are of course low level adaptive programming. The next level up controls the story line and environment -- always fresh - surprising - never the same. I would even settle for an area of the game to be unreachable for a while if my choices or actions required a complete rebuild of that section and resulting story line. The never ending story ...
Its not the years, its the mileage
if it had porn elements and its heuristic is to pit game effects and content against each other by competing for user attention, the endstate is rather easy to predict.
SLASHDOT: news for people who can't concentrate on work or have no life at all and got tired of yelling back at the TV.
What the hell do you all think I was trying to do with Polybius ???
Oops I've said to much, gotta g----
The solution is a market. A real one, where people actually trade with each other.
You need genuine demand for goods, regular food intake, equipment that breaks etc.
You need an open market that allows people to trade between each other.
You strongly favour specialists over generalists, this forces people to interact in the market.
There has to be mechanisms to remove wealth from the game, this can be transaction fees, regular taxes, destruction of goods through use, consumption of goods such as food. Demand must always be higher than supply, if everyone has the best armour, best weapon and tonnes of food, the economy is broken.
The system is self balancing. Everyone needs all the goods and everyone needs to interact with the market. If weapon smithing is lucrative for some reason a genuine supply/demand market makes weapons cheap enough that it's no longer the case. There will always be short term advantages to being in different groups but they balance as the market pushes people to correct for it.
The game administrators have two ways of influencing the market.
By tweaking the tax system and economy dampers they change the overall availability of goods.
Adjusting the demand for a particular good or the rate that it's produced will change the number of people in a given profession. This also impacts on the overall economy.
I don't think safety rails are necessary. A game with a dozen people will make the economy start to tick over. However if safety rails are required (NPC's producing food etc) then the prices should be punative. Selling at least 5x more expensive than a human farmer would sell the product for and buying at least 5x cheaper. People should not regularly be buying or selling from the NPC based market.
The big downside to all of this and the reason the game developers don't do it is because it forces you to interact with the market. Which isn't cool if you just want to go around and punch monsters.
into the hands of the gamers and let their interactions "evolve" the content.
Where is the "game" in this?
Blogging because I can...
...your game's evolutionary algorithms become self aware and start turning the weapons it's created on you.
- Dan
You will love a game called Noctis. (SCII is one of my favourite games and agree with you there)
A SCII/Noctis hybrid would be awesome although I prefer a fixed sized universe.
http://anywherebb.com/bb/posts.php?t=409
Slashdot needs Geekcode | Can anyone recommend any good SCIFI? My tastes: Foundation, Startide Rising, CITY, Ringworld,