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?"
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.
I agree with you and also would like to see more games with automatically created and evolving content. Unfortunately, game studios still seem to shy away from dynamic content because the behavior of dynamic systems is generally hard to predict. Some might fear that the game world suddenly becomes unstable and drops into chaos. But the game studios could hire more people with a strong physics/dynamic systems modeling background to deal with these problems.
Another problem is that games with good dynamic content have a very high replay value, whereas it seems that most game studios would prefer people to buy a new game or expansion pack right after they have finished the old one---or even earlier, as one might infer from the sloppiness with which later levels are often designed in comparison to the first few levels.
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.
I think the best way to do dynamic content in an MMO is to build a static framework (e.g. a world) and allow people to expand and create content from there.
Now, in my world, it would go like this: Your class, relative wealth, and profession would all have attendant properties and buildings. Some of them you could purchase for limited money generation: you're a smith, you need a smithy and a mine. You build the smithy, because you have to have one to make items, and then you prospect until you find a good spot, and you buy/build a mine, because it's cheaper than buying all your own ore.
Now, obviously, these buildings are vulnerable. You put the smithy and a shop (to sell your wares) in a big player-run city, where everyone who has buildings pays taxes that pay for NPC guards and defenses, so that's taken care of, but what about the mine? The mine is (obviously) outside the city, and not protected by the city guards. So you skim off some of your mining profits to pay NPC mercenaries to defend it.
Voila. You have player cities with guards, and dungeons with mobs, all at once. Make the shops able to be stolen from, and you have room for thieves, and let the players put in traps, fancy locks, etc. There is tons of stuff you can do. Obviously you're going to want to strike a balance. No fun to be a shopkeeper if you get cleaned out all the time. No fun to be a thief if it's too hard to break into a shop, and then there is nothing worth stealing...But that's just fine tuning.
See how logical and easy that is? And the person who built the buildings has a vested interest in keeping them going, paying for upgrades, replacing guards, etc. Everything can expand from that. Different types of buildings for different types of benefits to different groups. Military buildings for military bonuses, commercial buildings for commercial bonuses, etc, etc. You can throw in some PVE content: military group builds a building on the border with a non-human race, and kicks off a war with the orcs, or whatever.
Limitless possibilities, and everything that you do in the game matters. You clear a dungeon, it's gone, or empty, until someone rebuilds it, and then it won't be exactly the same...Depending on who builds it it could be completely different. You'll have done something unique, and how often does that happen in an MMO?
ad logicam Claiming a proposition is false because it was presented as the conclusion of a fallacious argument.
Being a smith in that world sounds incredibly boring. Does he have a computer in his smithy that he can play Tetris on?
Nerd rage is the funniest rage.
Shrug. Some people like playing crafters, and you can add a frickton of strategy stuff to it.
I'd love to see a crafting system than was complex and open-ended. Hell, make it so that you have some sort of skill-based mini game (like Tetris?) that effects the quality of things you're trying to craft, so it's not just about the level of grinding you've done on the skill, but also on actual skill.
On top of that, you have the whole "defend my stuff" part of the game, which moves toward traditional strategy elements. You're recruiting and training units, you're building defenses. You're making alliances with other players to defend each others stuff.
That sounds a hell of a lot more fun to me than just another MMO level/loot grind, where crafting is something you do when you're tired of killing stuff, and has very little actual effect on the game. I'm so tired of that I can't even convey it.
ad logicam Claiming a proposition is false because it was presented as the conclusion of a fallacious argument.
The problem with an open-ended system is that it is always unbalanced. At some point, the system evolves to the point where it makes more sense to be a smithy than to be a baker, or whatever. One profession/class/rank/item always tops out and becomes unbeatable. The only way to balance this is to either a) have mods who arbitrary slap people down by pushing the values this way or that or b) introduce a changing ruleset to balance things out through game events or some other "magic" process.
Either way, the players will find these changes "unfair". "I put all this effort into making this high ranking person, and your changes made me lesser!" is the basic gist of it. The problem of it is that they're correct, the changes did make that person lesser, in order to balance out the game.
- Give a man a fire and he's warm for a day, but set him on fire and he's warm for the rest of his life.
This is something Guild Wars tried to address in their approach. By making gaming areas instanced and meet-up areas MMO, a player's actions can effect the game world on a slightly grander scale. The state of the instance is decided based upon the previous actions of those entering it.
The thing is, I feel they didn't capitalize on this opportunity nearly enough...
09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0