The topic of today’s overly long blog post is ‘permanence in online worlds’. A few months ago, as a response to the blog post I made about morality, I was quizzed on why there was a necessity to formalise a morality system at all in a game. Many games don’t bother with such consequences, and where they have some kind of moral impact for what you do it’s based on having done certain things within the game. If you kill the little fairy princess, then people respond to that specific act rather than as a result of you being ‘chaotic evil versus fairy princesses’. If you can do that in games, why is there a need to introduce an artificial mechanism for tracking morality at all?
Permanence of action in online worlds is the answer – the need to create a world that is (largely) ‘same as it ever was’ has some pretty loopy implications for the way in which game systems must be represented. Online worlds require a sense of continuity (in some sense, a Groundhog Day of continuity) that single player, ‘once through to the end’ games don’t. The problem is that players can come in at the start of your game at any time, and likewise bow out at any point they like. In a single player game, the world ends with you. In an online game, the world continues whether you are there or not.
You can’t really tell long, sweeping stories in an online world. At best, you can tell small stories that interlink. Moreover, you can’t be ‘the guy who did the thing’ in an online world, because there will be a thousand other people who are also ‘the guy who did the thing’. Likewise, the need to provide for people to ‘log in’ at any point they wish means that you need to have a consistent game world – the guy with the starting quest always has to be there, or you need to completely drop the idea of having fixed game content at all. The boss at the end of the instance always has to be there. The components for the spells you want to cast always have to be there (or at least, they always have to be replenished). There are limits to how dynamic a world you can create in an MMO or similar without having to dramatically redefine what kind of game you are offering.
So, the actions a player takes within an online world lack real persistence, and that is the biggest limitation with regards to the kind of stories you can tell. Mostly people are prepared to buy into the conceit that they did impossible things and that makes them mighty, but the suspension of disbelief that you can count on rapidly fades away when you have people saying ‘You! You’re the monster who killed our princess’ and she’s fluttering away RIGHT OVER THERE, GUYS! SHE’S ALIVE! You *can* do that, but to my mind it horribly breaks immersion and the conceit that people buy into is fragile enough as it stands.
Your actions have no permanence, and that also means your actions do not necessarily have consistency. You might kill that fairy princess a hundred times, and save her a hundred times. How then am I supposed to use your previous actions as a baseline for how the game world responds to you?
Warcraft uses ‘phasing’ to get around this to a degree – zones change as you perform mighty deeds within them. It’s very neat to see a zone altered by your actions, but again it’s a conceit and it’s very fragile. It’s also very technically limited because of where players within different phases of the content must interact. Either they are ghosts you cannot see or interact with (which is weird if it’s a friend of yours and you are supposedly adventuring together), or they seem to be fighting imaginary monsters while you sit there sipping on a cool glass of soda (which is confusing). Multiplayer games require multiplayer approaches. While it’s possible for everyone to have their very own instance of the world in which they adventure, at that point you cease to have a multiplayer game – you have a single player game with some shared content.
‘Why can’t you just have a game world that genuinely does maintain permanence’, is the question that raises itself next. I say, as if they are universal truths, that NPCs must respawn and quests must reset. ‘Is there a basis for assuming that must be true?’
Well, let’s look at it via a thought experiment. Let’s call this ‘ConsequenceWorld’, a persistent, multiplayer online game designed around real permanence of your actions. There are certain things we can infer from this, but they are not necessarily core to the idea. ConsequenceWorld *implies* a world of permanent player death, but it doesn’t have to. We’ll assume it doesn’t, just for the sake of this discussion. ConsequenceWorld implies a fixed economy, but again it doesn’t have to. We’ll assume again that it doesn’t.
What ConsequenceWorld does have is a finality to your game actions. If you kill an NPC, they stay killed forever. If there are respawns, they have a fixed limit – ‘there are 1000 people in this village. When you’ve killed them all, the village is gone forever’. Perhaps respawning villagers will take on the mantle of their predecessors, so you won’t necessarily always get your quest from ‘Bill The Farmer’, because if he’s dead ‘Dave the Blacksmith’ will take over, and so on and so on until everyone in the village is dead and the quest becomes impossible to attain.
ConsequenceWorld also means that if you steal something from a shop, it stays stolen – it doesn’t respawn. If you destroy a building, it stays destroyed. If you hunt the wildlife to extinction, they stay extinct. If you empty a mine of its ore, it remains empty. If you run out of food, everyone starves to death. In short, every action you take in the game is likely to be negative sum.
There’s a phrase in economics that beautifully describes what happens next – ‘the tragedy of the commons’, and it describes the situation where exploitation of a shared limited resource will eventually lead to its permanent depletion to the detriment of everyone. It’s in nobody’s interest for this to happen, but simultaneously it’s in nobody’s interest to be the one person who abstains from exploiting the resource. In our scenario, the common resource is the pool of harvestables such as XP, materials and quests. If someone kills all the quest givers, then that quest is no longer available for the rest of the people exploiting the resource, meaning the pool has become depleted. This in turn creates an additional pressure on other harvestable resources. If you’ve killed all the orcs, you need to go somewhere else for your XP – but you’ve just depleted one of the XP sources, which means that if the same number of people are looking for XP their attentions are going to be directed towards a more limited set of possibilities. This in turn depletes those resources that remain more rapidly. As each resource is used up, it speeds up the rate of depletion of the others.
The end result should be pretty obvious – you end up with a ghost world where there is nothing to do and no way to advance. New players log on and say ‘What? Where is the actual game?’, and older players say ‘A newbie! Kill him for his delicious sweetbreads!’. If you’ve ever played Fallout 3 in Mass Murder Mode, you’ll find that the world becomes pretty rapidly a very lonely, very dull place. There are *some* respawns in Fallout 3, but not enough to really make up for the rate at which you consume. Likewise for Human Revolution – the dead stay dead, and while that is a great reminder of the way in which you chose to play the game, it does guarantee that you fast become a very *lonely* psychopath.
I’m not saying that such a game would be unplayable, I’m just saying that such a game model doesn’t lend itself to the kind of online worlds that most of us enjoy. It would be a horrible gameplay experience to start up Deus Ex and be left with the carnage of the last person who played it. You wouldn’t be able to do the quests, you wouldn’t be able to gear up or get ammo – but it wouldn’t be a problem, because there wouldn’t be anyone left to kill. If you want to have these kind of permanent outcomes in a multiplayer, online world then you need to be developing something substantively different from what everyone else is doing. You’d need to be working on creating something like Project Zomboid, which is literally about creating a scenario whereby you try to stave off absolute depletion of all your available resources for as long as possible. Project Zomboid has the tremendous tag line ‘this is the story of how you died’, which sets the tone perfectly – it’s not a game where you are expected to emotionally invest in a a long, persistent session. There might be mileage in an online game that has permanence of action but resets when everything is gone, but personally I think that kind of design is great for experimentation, and great for single player games, but makes for a shitty persistent multiplayer game.
So, the lack of persistence of the consequences of actions requires a different approach, and that often works via a proxy system. If I can’t base gameplay systems on your previous actions, I should be able to base them on what your actions *imply* about you. I might not be able to say ‘you killed that princess’, but I can say ‘you’re a pretty mean-spirited killer’. I might not be able to say ‘You have always treated our faction fairly’, but I can say ‘You have earned this faction rep as the cumulative total of your actions’. These things may be an artificial abstraction of the deeds you have performed in the game, but they serve the purpose of letting you ensure game actions have consequence. For some games, that doesn’t matter – for others, the theme may demand some measure of adherence to a system like this. The Star Wars MMO won’t feel like Star Wars if you’re not feeling the impact of choosing to do light side versus dark side actions. If you can kill, murder, rape and main without it impacting on your ‘Good Jedi’ credentials, the game is going to struggle to maintain your buy-in to the underlying game philosophy. At the same time, for all the reasons outlined above, the game can’t simply base your reputation on the fact you went into the temple and killed all the ‘younglings’ – because they’re going to be back for the next respawn.
Abstract proxies for ingame actions then are the tool that you have available as a designer for multiplayer online worlds. They may not be a more elegant weapon for a more civilized age, but they allow you to cover quite a lot of the gap between the desire to add consequence to actions and the limitations of the format.
 Which was a pretty great game I thought, especially the soundtrack.
 Players within different phases of content often report problems such as one player seeing a herb that doesn’t exist to the other.
 Although not technically feasible for most online games given the size and scope
 Just kill everyone once you’re done with them and then steal their cola.