Defining persistence for MMOs
Massively asks, “Are MMOs truly as persistent as they claim?”, prompted by a blog post over at Player vs Developer. The Massively piece actually takes off in quite a different direction than the original blog post, because the post is about how much game developer changes to balance and systems affect the perceived value of a given character. But the question that Massively asks is more direct: are MMOs really that persistent?
And the answer is unequivocally no.
Back in 2004 the IGDA SIG for online worlds put out a report on the state of the medium, and they called it “Persistent Worlds.” And at the time, I complained on MUD-Dev,. Dave Kennerly, who was on the panel, commented,
I too find the PSW thing troublesome. I inherited the title and couldn’t think of a better one. The IGDA MMORPG White Paper? No.
Well, we can see where THAT debate ended up. 🙂
Most MMOs today use persistent extended character state. You have your character’s stats, their inventory, and fake inventory such as the contents of their house (really just alternate inventory displayed differently). The extended character state developed over time from basic character state. Early combat MUDs saved stats but not inventory. Then they saved inventory and stats, but if the world crashed, you lost your corpse with your stuff. Then they saved inventory and stats, and add a way to save corpses so that your stuff would still be there if the mud went down. Then they saved inventory, corpses, stats, and the contents of your house. Then…
You get the idea. All of these additions were really additions to the state of the individual character. The world itself was not even truly dynamic, since it was lagrely generated from static flat text files. Reboot the world, and all the monsters were deleted, and everything respawned from scratch.
This is not significantly different from a modern MMORPG where that sort of content is delivered on a disc — or even streamed! — rather than loaded from flat text files. Either way, even if the world permits changes, they are really just deltas on top of a static, immutable, unchangeable initial state. And really, most games do not permit changes to this stuff, for a whole host of reasons: rewriting the map may well make it less fun, for example.
The other stream of design here, however was the true persistent state world, and it’s unfortunate that the acronym or even the word “persistent” has been co-opted to some degree by the worlds that really aren’t persistent in their state. Simply put, this is a world where the static inital data set is made dynamic in some fashion, and changes made to the world can persist independent of their association with a character. A house in Ultima Online is there whether its user is or not.
Many of the persistent state worlds share several big characteristics. They tend to be highly object-oriented environments aimed at user creativity. They often have a strongly simulationist bent to them, a MUD-Dev term that refers to the notion that much of the world can be driven via algorithms rather than statically placed data (this can relate to where creatures spawn, or to objects made of wood being able to catch fire). The big factor, though, is that they save most everything.
There’s plenty of pitfalls there, not the least of which is the reduced control the designers have over the user experience. The layering on of user changes can lead to a world where nothing is coherent anymore, thereby ruining the careful guidance that a designer may have planned.
But persistence of this sort is a sliding scale. You can have partial persistence, you can have a base state that gets modified but is reverted to gradually, over time. You can have locations that offer it and areas that don’t. And there’s a lot of possible and fun persistence that MMOs are leaving on the table that has nothing to do with whether or not the designers are changing the rules.
17 Responses to “Defining persistence for MMOs”
Sorry, the comment form is closed at this time.
Raph, are you saying that you haven’t left my house and grass and stuff in MP out there on the sim *exactly as I left it* and you have it like folded up somewhere to render only when I come back? Or somebody pops in?
Blizzard seems to be experimenting with regional persistence, as of the Lich King expansion. A few areas in Northrend change permanently, as the player quests through. The Wrath Gate in Dragonblight, the frozen lake in Storm Peaks. I want to say there’s another, but can’t recall where at the moment.
One problem that has presented itself is that the mini-map doesn’t reflect the “phased” region. Let’s say you’re tracking mining, and see a mining node on the mini-map. You fly over, and it vanishes. Not because someone mined the node, but because you’re saved to a phase in which the node doesn’t actually exist.
For me, the most annoying persistent change in Galaxies was when people dropped houses around mine, making my lovely, forested, creek-side cottage into a frigging mobile home park. 😉
I found that entire article somewhat amusing and bemusing. It was lambasting the lack of persistence in online worlds through such a narrow focus, defining persistence as if it meant completely static and unchanging. As Raph states, MMOs are more persistent today than they’ve ever been. At least in terms of persisting character and world data that is. And that’s all that term was referring to.
I’m not sure I would want to consider phasing an experiment in persistence. It’s kind of an old trick we’ve seen in several MMOs at this point (offhand I can think of EQ2 and Auto Assault) and while it does allow plot lines to resolve themselves permanently, that change only perists for that single user, as you illustrated in your node example. I guess it is persistence, but it comes comes at the price of isolating the players from each other, by way of “being out of phase”.
I think we’re going to see more “public” attempts at persistence by other games soon, however. Parsing recent press releases seems to imply that more than one company has been putting time and thought into this problem. I have to admit, that gives me more hope than I’ve had, in awhile.
@Mike I disagree .. the phasing is entirely personal, and is thus simply a further extension of personal state persistence, not world persistence.
The closest they’ve got would have been the rollout of content during the Shattered Sun Offensive, and even there it’s not that impressive (being more a staged deployment of a content patch).
You might also consider Wintergrasp and Halaa for examples of persistence though.
The definition of persistence in MMOs and VWs is a tricky one, as there is an awful lot of crossover with dynamism. Perisistence is where player actions have a lasting effect on the state of the game world, dynamism is where the world is always changing due to the actions of players. The difference may look like its just a case of semantics but there is more to it than that.
Persistence implies that a player action can change the world permanently, and once changed it stays like that. Dynamism emerges from persistence as a result of many players actions changing the results of other players actions.
For instance, lets take EVE as example (I know, but its the game/vw I know most about and have most experience of). A player places a large number of the same item on the market at a price which undercuts his competition. If no-one then undercuts their price, then that price can be said to be persistent, that player has changed the world permanently. However, that player’s competitors are not to be out done and figure out a way to undercut the price again and still make a profit, so the price changes again, this is a basic dynamic system.
It would be difficult (and inadvisable) to make every system (resource placement, MOB spawning etc) dependent on the actions of players or affected by the actions of players in this way. In a close system like a VW you’d soon end up with MOB extinction and resource exhaustion which could damage the world irrevocably. Some things have to be static, for the dynamic system to be relatively stable.
IMO, Blizzards claim of persistence in WOTLK is a false one as the world only changes for that individual player and it makes little or no difference to the rest of the world when it does.
Getting a VW world to be truly persistent/dynamic is a ways off yet and would involve every system being simulated procedurally, it would also take control of the world and its behaviour away from the developers which may lead to the VW becoming, at best, no fun, and at worst completely unplayable.
(OT, just got ATOF and am enjoying it immensely 😀 )
When I think of persistance, I think of legacy. After I leave a game, will there be any trace that I was there?
Of all the MMOs and virtual worlds that I’ve played, I can only think of a handful for which this is the case. AlphaWorld, as far as I’m aware, never wiped player creations, so the log cabin I created is probably still there after 14 years… although, having lost the coordinates, I’ll never see it again. There may still be weapons, clothing or books with my maker’s mark floating around UO or SWG, and my creations in Second Life will continue to clutter up bloated inventories after I move on.
(Of course, even if there is an opportunity to leave a legacy, it’s only as persistant as the game itself. The band object in The Sims Online was based in part on a proposal I wrote, which would be a great legacy were the game still in existance.)
Call it ego, if you like, but I think it’s a common desire in both the real world and the virtual. I’d just tickle me to think, a hundred years hence, some updated holographic warrior might whip out a flashy katana, and declare to his admiring friends, “It’s a Thorin Ironbeard. Collector’s item. They don’t make them like this anymore”.
mandrill wrote “take control of the world and its behaviour away from the developers which may lead to the VW becoming, at best, no fun, and at worst completely unplayable”
The world will necessarily lose its lovingly tailored and idiosyncratic feel, but it won’t necessarily become unfun. Randomly generated maps in Civilization and Master of Magic and various RTS games aren’t unfun. And I remember interviews with Deus Ex designers where they talked about how even in their lovingly tailored idiosyncratic scenarios, unplanned consequences of the general rules (for explosions and whatnot) tended to create more fun and interesting possibilities than they could design in by tailoring specific rules.
On the other hand, it does seem to be awfully hard to get right for a persistent world. One problem is that the emergent consequences of general rules tend to depend on the rules in a very complicated, unstable way, so that it’s hard to foresee the impact of tweaking a general rule. And for extra design challenge, sometimes the impact can’t be determined very quickly by testing because it’ll take your playtesters or playerbase a while to catch on to the possibilities opened by the rule change. This problem can be very severe in a persistent multiplayer world where causal chains can be very long and run through many, many game-theoretical faceoffs between players: it is fundamentally hard to calculate properties of outcomes of processes like that. Forcing the player experience to run on preplanned rails through hand-crafted scenarios involving custom-made rules and fixed data helps control that instability. As a consequence, even if custom-made specific rules don’t make fun more probable than general rules and emergent data, it can still be much more tractable to use designer intuition and tester feedback to hunt down the fun subset of specific rules and hand-crafted data. The fun subset of general rules and emergent data may be comparably large, but who knows how to find it?:-|
What is the color of fun?
As a game complexifies, does it exhibit coherence in the abstract sense of interferometric visibility?
Is wave function collapse a valid model of complex play? Is there a visibility ideal?
Is fun intensity?
if a tree falls………….
Still confused about this use of the word persistence; coming here with the dictionary meaning and trying to understand a seeming contradictory concept. But will give it my best shot with an example: building a house in a mmo world, that’s persistent in the sense that it should stay there with it’s content intact (like the use of persistence mentioned in the article regarding saving data like stats, items etc), determined by the rules set by the developer that nobody could break and enter. But that’s not realistic, especially in a pvp game, to not be able to ransack a house or even to have ‘bandit’ mobs programmed to loot a house that is unattended and in the wilderness. On the one hand the developer is creating a more dynamic situation, but with the other imposing some strictures on players i.e. who would be foolish enough to build a house in the middle of nowhere, away from civilisation, so you see it is wrong to say that developers programming more dynamic play lose control over the activities of players, they do it in more considered ways. Going back to persistent again, the so called legacy of a player who chooses to build a house in the wilderness in a war-torn land is a set of ruins and passing players saying noob under their breath. 🙂
My idea of persistance is for a newbie to come along a year or two into the game and explore a set of ruins that resulted from an epic player city war that happened just after launch. The ruins are a clear marking of past NPCs and PCs of note. Some veteran that’s mentoring the newbie explains that this is the site of the old city of blah blah blah. Man you should have seen the battle here… It all started with this warrior that managed to get a high ranking in the city only to betray the city…
In that regard, Eve’s about the best we’ve got in a landscape littered with character only persistence games. I want the game to be immediately recognizable as a game but open enough to allow the players to own, shape and create the future history past launch day. That rules out the cool ideas of Second Life and Metaplace, as well as the social web sites and leaves me wanting for the next level of MMO experience. I’m not interested in playing checkers anymore (WoW, EQ, DAoC, EQ2, WAR, LotRO…) and no one will sell me a chess game (Eve) set in dwarves and elves fantasy.
@c3: were you flying over it while in phase? if so, kaboom. if not, it didn’t matter.
In a persistent system, the tree is still treeing if affordable.
If you’re interested in games displaying this kind of persistence, see Wurm Online. I haven’t played in a year and I’m certain beyond doubt that you can find bits of land that I terraformed.
[…] — David, in a comment in the earlier post […]
It’s also not realistic to have the would-be thief who was just decapitated by your traps/guards come back for another go at it. In the real world, a remote keep could get by quite well in the absence of the landowner — an army could take it, but players don’t command armies, just rag-tag bands of motley adventurers. And in the real world, a single crossbow bolt to the chest can ruin your whole day.
That’s not an argument that player housing should persist forever. It’s cool that it does in some virtual worlds, but it also means those worlds aggregate over time a disjointed sprawl that’s constantly expanding.
The challenge to the designer is to envision ways that players can leave their “Killroy was here” stamp permanently on the world without disrupting the experience for others. Maker’s marks are an example (though the permanence of such is subject to destruction of the objects). Other games have monuments to players or developers who have passed on; can that be reasonably expanded to honor any player who wants to be remembered? Are there ways to chronicle adventures in game that isn’t dependant on the memory of old-timers or the longivity of a particular guild? And whatever the mechanism, is it proof against the malice of that peculiar species of subhuman who lives to destroy the enjoyment of others?
[…] course, the persistence of MMOs can and has been called into question. (Raph weighs in as well, noting that giving players power to make truly persistent changes to the world can be […]