Slashdot Mirror
The Future of Persistent Worlds In MMOs
Zonk did an interesting interview with Ed Stark and Dave Williams, employees for an MMO developer named Red 5 (and experienced tabletop game designers). They talk about their ideas and plans to bring about the next step in MMO gaming: increased persistence in online worlds, where an objective, once completed, stays completed. Williams said, "Right now for most of these games, when the player saves the princess and he starts walking away from the tower — if he looks back he's going to see the princess at the top of the tower again." Regarding their current work, he continues:
"If you save the village, it stays saved — you saved it! But maybe now that village becomes an objective for another player; maybe something has to be done now because that village wasn't destroyed. And so on, and so on, and so on. Building those mechanisms to make it a world that reacts to a player's actions instead of existing in a static state. That's the world we're talking about."
I'm not a programmer, so I don't really understand, why is it so difficult to have objectives that stay completed after you've completed them? Can someone enlighten me as to why that's a step that's still forthcoming?
I've learned that they're worthless, so I don't read AC comments anymore.
With all this talk about doing away with instancing, I'm surprised they didn't mention EVE Online. EVE has *ONE* world for all the players. Granted, it can make it mighty laggy for large engagements, but most of the time it's fine. Missions are "instanced" insofar as they are randomly created when you get them, but they can be discovered by other players using scanners, so you could conceivably have complete strangers swoop into your mission and rob your loot. Annoying, yes, but it adds tremendously to the feeling that you're part of a larger world.
3. Profit!
2. ???
1. On Soviet Slashdot, a Beowulf cluster of alien Natalie Portman overlords welcomes YOU!
It's hard to make this work in a way that doesn't allow the world to deteriorate. How does all the stuff that players destroy get repaired? Probably by a huge number of NPCs working very hard around the clock. The NPC AI's need persistent state, too. They need to learn from experience, so they will rebuild better defenses. Walls are built stronger. Weak points are plugged. Overlapping fields of fire are set up. Obstacles to slow up assaults go in place. Towers are built to be mutually supporting. Checkpoints where players must disarm are put in place. NPC guards discover flanking tactics.
The day will come when the NPC AIs get smart enough to realize that the players are ruining their world and band together to exterminate the players.
The solution is algorithmic in that these games should be able to support a non-entropic reality and introduce things on their own.
I know that's much easier to say than do, but perhaps the bar is set too high to do this now - particularly on a planet wide scale. Maybe it should be tried at a city or small town level first before trying to do it all at once.
If things were dynamic enough, the developers wouldn't have to plan huge expansions of meaningless quests - ideally, these quests should form on their own based on the changing social-political situation in-game. Solutions to the computer-generated quests should also be up to the players at hand. Oh no, there's a Big Magic Dragon! Should we use magic or spears to kill it, how many people will we need, etc.
One of the most disappointing things about MMO's to me is the fact that NOTHING matters. It's an empty experience but for the social interaction with the other players. Most of the quest solutions are online anyway, there doesn't seem to be much of a sense of true adventure. True adventure involves risk of the unknown and there's damned little of that in an MMO game.
I think that games like Spore will prove (at least to some extent) that this is possible now. The first company to apply Spore-like persistence and algorithmic flexibility to MMO's will do incredibly well.
"...Well, there's egg and bacon; egg sausage and bacon; egg and spam; egg bacon and spam; egg bacon sausage and spam..."
The biggest problem is, they're underestimating just how fucking fast players will progress through the content. If there are milestones like quests to be met, the players will figure out a way to pass by them at light speed. Even if you have ten thousand canned quests ready for deployment as soon as the first batch is completed, they'll be gone in a month at most.
Someone pointed EVE Online out as a truly persistent world. I'd like to add Ultima Online to that list. Both have only the barest NPC interaction: there are mobs to hunt and kill, and merchants to deal with, both of which can be bypassed by a mature-enough player-base. The only 'quests' are specific GM-run events that are more complex than 'harvest five bear asses' or 'serenade the Princess for Cyrano'. Everything else is a sandbox. Players create their own storylines, fight their own wars, and build their own merchant empires. There's no need for a traditional dungeon master because the server is smart enough to handle simple math like combat, and the only real social interactions occur between players, not players and keyword-driven mobiles.
But there should be unique items, a few of them so rare and so powerful that whoever comes to wield one of them will be like a demigod in that world. Players will form factions, armies, to steal just one of these items from whoever currently has it. (And it will take an army to get it.) Players will betray each other to get control of one of these items.
Also, it should be possible for a character to really die, and stay dead--not just respawn. Once dead, a character should become a ghost, who can roam the world, chat with other players, and contribute in minor ways to a party, perhaps as a spy. Then governments will form. Good players will unite to protect newbies and others from player killers, and a justice system will spontaneously develop.
The problem with the "everyone is a hero" mentality is that it not only makes for a wildly inconsistent and boring world, but it also fails in its objective. Seriously, can anyone out there claim that they felt like the hero after doing a quest that a thousand other people have done? Do you really feel the hero when you ask on chat how to save the princess and someone LOLs about how easy that quest was? The "you are a hero" quests are as mundane as any other quest.
Personally, I think MMORPG should realize that the MMO part is supposed to mean something greater than making a really tedious single player where other people also play. I personally think that many people would be far more interested in a changing and dynamic world than a world of dozens "you are the hero!" quests that everyone and their dog has done.
I think that the biggest problem is that MMORPG makers are afraid to have people lose. They want you to always feel like a winner, and as a result the game becomes very dull for many people. There is absolutely nowhere to go in an MMORPG but up. I am not advocating massive exp loss or anything of that nature. People hate that sort of thing because they hate to grind. What I am advocating is a world that can turn for the worse. While you are at it, maybe it is time to rethink the absurd exponential power curves that forces content segregation.
Consider:
Forget the mechanics for a moment; just imagine an MMORPG built upon the principle that all people should be able to enjoy the content. That means instead of having to make content for each 5 level slice, content is there for all. It might mean that you need to rethink "power" and âoeprogressâ in the game.
Now, you have a game where everyone can participate in content. Now imagine a threat arises that is dynamic and moving. Instead of the "threat" being a new area spawned in that you can go to and spawn camp at, imagine if it was a living and moving thing.
So, letâ(TM)s take the classic zombie horde. The threat is a zombie horde. It starts at one end of the world and moves to the other end. As it kills it grows. It moves slowly, but it clearly moves. As it moves into an area, zombies wander in slowly. When resistance is met, zombies start heading that way. Any prolonged resistance results in a horde concentrating. So, if you defend a town, you can hold it for a while, but after some time you get swamped and either need to flee or get reinforcements. Even if you do not resist, at some point the zombie population gets thick and everyone dies.
Make it so that there are no-win scenarios. You can hold a town for a time, do so damage to the horde, but in the end you WILL lose. The best you can do is do some damage and fall back.
So, the players keep fighting and falling back. Perhaps they make some valiant last stands in various popular cities, but in the end the cities are conquered one by one. If the players fall back effectively, do damage as they retreat, than at some point they might thin the horde enough to actually hold a city. Instead of being swamped in a few days, they might just find themselves in a long term siege that lasts weeks or months. Other players might try and fight supplies in, while others fight from the walls, clean up sewers, and clear out zombies that slip in. Maybe after a time the momentum is reversed, and the players are able to push back the zombies and reclaim land.
Of course, things could go the other way. The players could be pushed back and pushed back until there is nowhere to go. The world could end and the game starts anew with some different challenge facing it.
Some people will hate this type of game play. Some people want to win every time. Other people will love it. I don't know about you, but the idea that you could actually lose is thrilling. A desperate retreat fight back to the center of the empire, losing city after city sounds a shit ton more exciting than farming NPCs or doing save the princess quests. Do I g
Good post!
I've thought about this myself and came to exactly the same conclusion. I'd put the matter only a little differently: The question is: When will we write AI that can do all the work of a competent human game content developer? I expect it will take a while, but not forever.
Surely, there will initially be some cheating. So maybe when you kill an orc, it will indeed stay dead, but another orc will spawn elsewhere in the forest, waiting to be found by another adventurer. I would love it if we tried to simulate an actual ecosystem that simulated (among other things) the conception, birth, feeding, etc. of orcs, but that would have three problems. One: It's just asking too much of the AI. Two: It would reveal how ecologically incoherent most of the classic "dungeons" in rpg's really are (even by standards of fictional ecology, orcs must consume so many calories each day, shit somewhere, etc. They're not going to be having lives in some single room in a dungeon.) Three: Such a system, even if it were relatively stable without PC interference, could easily be corrupted hopelessly by the actions of some powerful player characters. Destroying is much better suited to the activity of a PC in a CRPG than is building. Destroying is faster. Cutting down a tree is much easier than making one grow.
All current and future CRPGs must find a way to set back up the knocked-down bowling pins, or set up different bowling pins instead. If a group is bent on deforestation or depopulation of a country, and the game must replace what is killed in a natural way, there will soon be no more trees nor NPC humans.
Of course, systems could be introduced that prevent such actions. Repairer druids might magically regrow lost forests, but who will generate replacement villagers?
What's attractive about "adventure settings" is that they are in a context of very weak central institutions (so they leave space for adventure) with potentially powerful individuals. There are no "adventurers" in Singapore, because there, even spitting on the CCTV-watched street gets you in trouble. That's one way to prevent chaotic rampagers, but the four classic role-playing world types (middle-age w. spells, western, war & post-apocalyptic) are not chosen by accident. They're settings where individuals are not under the yoke of a central authority. For fans of Firefly: The protagonist adventure-group does their work on the outer planets exactly because central control doesn't extend that far.
My point is that it wouldn't be an adventure game if it were in a setting that prevented individuals from devastating rampaging. This means that such settings are inherently unstable. (Usually, strong governments elbow in and stamp out the "adventure space" - for the most part.)
So even a perfect AI would not be able to impose stability on an inherently unstable, fully simulated situation. Adventure settings are paradigmatically not in equilibrium.