A Wish List for Massively Multiplayer GamesDr Richard A. Bartle
IntroductionThis talk concerns designs issues for massively multiplayer games.
Well, sort of. Although I do go into specific design issues later, what it really concerns is the lack of original design in upcoming massively multiplayer games. Cosmetic and genre changes are just sugar frosting, and world design is more like level design in first-person shooters - you might get a snazzier look and some intriguing new puzzles, but the gameplay is much the same as it ever was.
Massively multiplayer games are the titans of online games. EverQuest takes $4,000,000 a month, Ultima Online $2,750,000 and the "failure" Asheron's Call $800,000. Those are just the ones played in the English-speaking world.
Spurred on by the success of the current big three (i.e. the potential for untold riches), a second wave of games is appearing. Most of the designers of these games seem to have garnered much of their knowledge of the field from having played one of the big three (and that one is usually EverQuest...).
Aside: EverQuest is a knock-off of the DikuMUD server code, with a graphics engine bolted on. DikuMUD does a lot right, but a lot wrong. People who have played DikuMUDs and want to run their own often hack the code to make it work how they want it to work. Because they don't do it right either, other people rehash their code. There's a whole family tree of DikuMUD descendents, each varying in sometimes subtle, sometimes not so subtle ways from its parent. Are we going to see something similar with EverQuest?
No, we aren't, at least not to such an extent. DikuMUD is open source, EverQuest isn't, so each "child" of EverQuest must be recoded anew. Is this a slight window for creativity to creep in, perhaps?
The big three have dominated the online game scene for 2-3 years now. At last, the new wave is breaking. How big will it be? And what new treasures will it wash up on our shores?
New Kids on the BlockThe new games (just out or close to coming out) fall into one of five categories:
The ChallengeHow will the new games affect the status quo?
What new ideas are these games bringing with them? The state of our industry is like the movie industry was 80 or 90 years ago. Are we seeing equivalents to advances in technology, acting, writing and directing? Or are we just getting the same Keystone Kops sequences week after week?
What old ideas are these games leaving behind? In 1912, the Motion Picture Patents Company was a consortium that held all the major movie-related patents. It controlled over half the 10,000 Nickelodeons in the USA and was the only organisation other than Pathé in France licensed to use Eastman filmstock. They held all the aces, but none of the founding companies survived the 1920s. Why? Because they were geared to produce 1- and 2-reel movies only. Independents borrowed the concept of "feature" films from French and Italian film-makers and cleaned up. Are new games dumping outmoded ideas and forging ahead with something that has the potential to turn the market on its head?
Also interesting from the point of view of design is what the new games shouldn't have done that they have done, and what they should have done that they haven't. If they bring in new ideas, great! If, however, there were some innovations they could have made but didn't, that raised the question "why not?". Similarly, if there are some obvious bad ideas that they didn't ditch, why didn't they ditch them?
What we have NowAlthough successful, the massively multiplayer games that formed the first wave are not without their problems. They have three goals to achieve that sometimes conflict:
Functionally speaking, acquiring newbies is the most important of these because the more players you have then the stickier your game tends to be; retention is therefore helped by acquisition to some degree, which in turn helps make money (the ultimate goal). That said, games that appeal to newbies don't always appeal to oldbies and games that appeal to oldbies don't always appeal to newbies (hi, Asheron's Call!). These problems can be countered by throwing money at them, but then you don't make a profit...
Some of the issues disappear with good operations. Reliability and customer service can go a long way to spreading good word of mouth, nurturing a sense of com ... comm ... commun ... it's no good, I can't say it. They welcome newcomers, and help make you feel a part of the game world so you don't want to leave.
The big design issues at the moment, where current massively multiplayer games most painfully fall short, are:
How do the new games address these? Indeed, do they address them at all?
Coming Real SoonThe new games seem to be split as to whether to aim at existing gamers or aim at non-gamers. It's long been known that MUDs can have demographics to die for:
You might, therefore, expect most of the new games to be targeting non-gamers.
There are two reasons for this. Firstly, most developers want to develop games they'd like to play themselves. After all, why waste 5 years of your life working on something you don't like? Secondly, the developers have typically played one or more of the current games and think they can do it better. They imagine that doing it better will automatically bring in players by the trainload.
The first point is fair enough, but the second is ill-conceived. It's not like you're converting someone to a new religion and they'll immediately spread the word the moment they see the light. Rather, it's like you're a young woman attempting to woo a man away from his wife and then a couple of months later you start worrying if he's going to run off with someone prettier than you (after all, you've shown him he can do it once).
Writing games for gamers tends to colour the nature of the games developed. The games become more and more centred towards the core market - often, even when the developers believe they're spreading the net wide. What's mainstream for gamers looks way out to non-gamers. What's more, some of the things that gamers who have played a game every night for 6 months come to believe do not necessarily present a true picture.
When I was a child, I used to eat school dinners. I hated the salads. They were awful! Everyone hated them, we used to dread salads. We would plead to be given something else - anything but salads! They tasted awful. No more salads! If someone designing a new school meals menu had visited our school, salads would have been top of the list of things to do something about. If there had been other schools in the area that didn't serve salads, I'd have wanted to go to them. Yet salads are good for you!
Actually, we all hated salad cream. We didn't know the people behind the counter put it on as it was some kind of watery vinaigrette that didn't look like it was there. We were complaining about the wrong thing.
Before they cut salads from the menu, game designers need to know whether it's the salads that are the problem or just the salad cream.
Out with the Old...Here are some of the old standards that new games are throwing out:
...In with the NewThere are two main directions that the upcoming games are (collectively) investigating: new genres, and content creation by players.
The great hope with new genres is that they'll highlight new issues or suggest new ways of thinking about old issues. Perhaps surprisingly, most activity here is coming from the licensed products, as they have bounds within which they must operate and therefore solutions have to be found to problems that other games can gloss over.
Generally speaking, there's a swing from Fantasy to Science Fiction at the moment, on the grounds that the underlying implementation of a SF world is much the same as that for a Fantasy world and that lots of gamers like both. It's something that developers and programmers tend to like, too, although on the mainstream radar it registers with the same sized blip as Country Music.
There aren't many properties that would make good licences for massively multiplayer games. Too many have an indispensible main character (Buffy the Vampire Slayer), too many are niche (Stargate SG-1). Not many of them describe worlds, and the games themselves take so long to write that they might be born after the franchise has died. Sure, there will always be people who would want to play a game based on The Matrix, but not enough to justify the development costs (although it would work as a text MUD). Those properties that do have the necessary qualifications promise to be sensational hits, however, in particular Star Wars Online (which has a crack design and development team behind it.).
Some "licenses" are actually public domain, but whoever picks them up first can effectively claim them. The archetype here is Dark Age of Camelot: Everyone knows and loves the Arthurian legends, but you don't have to pay royalties to any brand owner to use them. Yes, someone could come along and do a game in the exact same setting if they liked, but it would probably do them more harm than good if they tried..!
The other area of experimentation, player-generated content, is at once exciting and depressing. It's exciting because (setting aside copyright and ownership issues) it gives a chance for new talent to emerge, adding freshness and vibrancy to a very insular industry. It's depressing because most players cannot design, draw or write, yet believe they can do at least two (if not all three) and become very uppity when presented with evidence to the contrary. By carefully limiting the parameters that the would-be designers can change (as with Neverwinter Nights) some attempt to bring a degree of hands-off editorial control to what is produced can be made, but overdone it could stifle creativity. That said, I do believe that this could be very fruitful territory for game developers to explore.
What Players Want that they Won't be GettingFirst of all, the usually cautionary note: players often don't know what they want, or refuse to accept the consequences of their desires (because of the tension between the good of the game and the good of the player). Example: the same players who demand for reasons of balance that an object's powers be reduced are likely to complain loud and long if the next object to suffer the same fate is one they actually possess.
Players want variety. If people like to live by routine, fine, let them. If people want to experience the new, though, shouldn't they be allowed to do so? Variety means quests that change every time; variety means unlimited opportunities to role-play, to experience the game from vastly different perspectives. It doesn't mean the thrill of wondering if maybe this is the spawn when the monster will be carrying the magic toothpick you crave.
Players want atmosphere. They want the game to draw them in. People who don't want it might spoil the general ambiance by discussing last night's football results in public, but unless the game itself promotes an atmospheric world there's no chance of getting it at all.
Players want wit. That's twice as much as what they get.
Players want to matter. They want to make a tangible impact on the game, and for it to be meaningful. Being hailed as a hero is tremendous if you've done something heroic, but if everyone is a hero then ultimately no-one is. Spending a day to make some craft item that nobody wants but hey, the game will buy it off you to stop you moaning is frustration exemplified - it really rubs in your insignificance
What Players Don't Want that they Will be GettingPlayers will be getting storylines. There are some very persuasive theories of drama, and influential academics have attempted to apply them to massively multiplayer games. Any modern screenwriting book will tell you about Syd Field's three-act paradigm, every TV viewer understands that if the dragon razes vast tracts of countryside it's because it was abused as a child...
Storylines are bad. Players want to change the world, not impotently watch it unfold about them in a predetermined manner (branching or otherwise). Arcs work for soaps and TV Science Fiction series, but players want to shape their own futures. A story can provide a shared experience to help people socialise, but it has to be rapidly changing and feature stars that people care about. It's far better if the players themselves are the protagonists and the audience is participatory. Besides, once the show starts, how will newbies feel when they come in half-way through?
Players will be getting politics. This is seen as the great solution to the end-game problem (i.e. what players who have been playing for 3 years do). I have my doubts. If it were the very raison d'être of a game, fair enough, but it suffers from the basic flaw that most people actively dislike politics and politicians. Besides, what does it mean for players when the game is 6 years old and there are more would-be politicians than there is politics?
Players will be getting artificial group conflict. You're a member of a faction at war with other factions for control of the country/continent/universe. This would be really interesting were it not for the fact that no side is ever going to win or to lose. Dark Age of Camelot has three factions fighting for control of Great Britain, but none of them is actually going to win it. This is a great shame. What should happen is that if a side does achieve victory then there is the possibility that players could split off to form factions of their own and begin new conflicts. If the original war is never-ending, what, then, is the point?
Players will be patronised. Half-hearted solutions to genuine player concerns will be passed off in syrupy "we listen to you" personnel manager language. "People will simply adore the pretty baskets you can make". "Cooks are just as capable of great deeds as half-elf paladins". Yeah, right...
All Things to All PlayersThe problem is that players have conflicting needs. Addressing one player's needs affects the other players' needs.
Old hat quick summary of player types:
It's possible, but not easy, to get a stable balance between these. If you manage it, the game costs very little to maintain and will "virtually run itself" (as if...). Otherwise, balance has to be forced, usually either by making continual updates or acquiring a newbie hose.
Conflicting NeedsAchievers need an end-game, but the one they've been promised (politics) is for killers.
Socialisers need meaning, but they don't accept persona death as a mechanism for delivering it, and they don't accept a free market economy.
Explorers want depth, but depth costs money.
Killers want to dominate other players (except for other killers), but no-one wants to be dominated.
Socialisers and explorers want story but achievers and killers don't.
My Ideal GameThese issues are not insoluble.
Too many designers start from the wrong set of axioms. They use received wisdom as if it were a given truth, and consequently try to find a way out of a dead end that they shouldn't have entered in the first place.
I'll now outline how some of the major problems can be resolved. This will be done very briefly or it would take forever. Note that this is from the point of view of a designer. It's not that of a player, because (as explained) players have conflicting needs; it's not that of an operator, because good customer service is good customer service (not that players currently get good customer service); it's not that of a gamer, because the gamers are already playing something else...
Persona DeathThe single most misunderstood concept in massively multiplayer online games is that of persona death (PD). Early implementations didn't work, but the whole idea was condemned as a result. Salads and salad cream...
Persona death is the permanent, irrevocable loss of a character in the game. Note that this is not the same as "player versus player" (PvP), which allows for characters to do nasty things to one another such as attacking them or stealing from them; neither is it the same as "player killing" (PKing), which is PvP that allows for the PD of the loser.
Some of the newer massively multiplayer games have PD in them, much to the consternation of EverQuest players (not that it should matter to them, it's not like they were going to leave their friends and play the new game anyway).
I want to see PD as a feature in those games where it makes sense for the genre.
Without low-points, there are no high-points. Death gives meaning to life. Player characters have to be able to die, dead dead, 100% snuffed it. It gives heroism value, it gives character value, it gives players the chance to start afresh (and yes, that is a blessing rather than a curse). However, curiously it's only the threat of loss that's of primary importance, not loss itself. So long as the threat is real (maybe 1 deadly peril situation in 20 a character dies) then the heroism is real.
Most new games start from the standpoint that players don't want to lose their characters. Well no, they don't. In practice, though, it's only socialisers who truly hate it. This offers up a solution: make PD apply only in areas where socialisers never go. Restrict it to certain well-advertised badlands, which if you don't visit then you won't be killed. Why would anyone ever visit them, though? Because otherwise there's a limit to how far a character can advance - you might have to visit a shrine in bandit country as a precondition to becoming a bishop, for example. People who want to "win" the game will take the risk, and players who don't won't - the socialisers.
Some people won't play any game that has PD in it on principle. Don't worry about them: they weren't going to leave their current game anyway.
Player versus PlayerPvP is the concept that players should be able to do to one another what they can do to mobiles (monsters and non-player characters). If you can steal from a mobile, why not another player? If you can cast a sleep spell on a mobile, why not another player? If you can give a mobile a ticking bomb, why not another player?
"Because the other player won't like it".
PvP is similar to practical jokes, in that being the perpetrator is a lot more pleasant than being the victim. Of course players don't like it when it happens to them! That's no reason to ban it across the board, though. PvP adds conflict, conflict leads to drama, drama leads to interaction, interaction is the USP of massively multiplayer games. Note that PvP doesn't have to come with PD; it's possible to allow stealing without allowing assault.
Some PvP is a good thing. Designers have in the past recognised this, but the solutions employed (reputation systems, PK switches) haven't worked. In response to this, the pendulum of opinion swung the other way, asserting that PvP was intrinsically evil and therefore all aspects of it must be purged. This includes hard-to-legislate-against secondary forms, such as riling a group of monsters then leading them straight at a group of picnickers.
The problem, as with player death, is unsolicited PvP. If people raise their hands and say "I'm prepared to risk some PvP if that's what it takes to get the treasure I want" then they have accepted that PvP may occur. As with PD, if people consciously agree to take the risk then they have no comeback (and, more importantly, feel the game is less unfair) than if it just happens when they were happily picking berries in the woods without a care in the world.
So the same solution can be adopted as for PD. In some areas there's PD, in some areas there isn't. In some areas there's PvP, in some areas there isn't. In some areas there's both (PK areas). In most areas there's neither - the default. If you don't want to risk PvP, don't visit the Thieves' Guild. If, however, you want the Spoon of Power from the arch-thief's trophy box, you have to weigh that against the risk that someone will wallop you when you try to purloin it. It's a dilemma, but that's the kind of thing upon which role-players thrive.
You can't stop negativity anyway. People will always find a way to block doorways, steal your kills, lure monsters to attack. At least with PvP you can get revenge!
Player HierarchiesThe problem: socialisers want to be part of social networks, but these are always dominated by killer types.
The solution: have multiple, interacting hierarchies.
Politics is a game for people who are either serial manipulators or are na´ve enough to believe that they can make a difference. If there is only one hierarchy, the killers will control it and everyone within will suffer. If there are multiple hierarchies, however, the leaders can war on each other, not on their own members. Members become valuable commodities, because the more you have then the more power you hold. Even formally subservient hierarchies can be independent: the king can sack the general, but if the troops decide to follow the general anyway...
I don't believe that player hierarchies are something that game designers are deliberately omitting, it's just that they're so focussed on combat that diplomacy takes a back seat. Remember, it's war by other means.
Different hierarchies that can play a part in most Fantasy style game worlds include:
QuestsExplorers need content, but writing it is expensive. Automatically-generated quests are the answer. The way that games currently do this, however, is awful - just awful! Quests, are repetitive and predictable (which leads to camping); they are context free ("Rescue your cat? When 10,000 invaders are attacking the city?"); they lack variety ("Gee, you've lost your cat too"). Designers accept that quests are of major important (as the very name EverQuest shows), they just don't know what to do about it.
This is because they're looking in the wrong place for solutions.
Quests can arise naturally from the actions and interactions of players. A player has a goal ("become mayor"), which leads to sub-goals ("get rid of the existing mayor, then get elected"). Alternative ways to achieve these sub-goals are considered ("to get rid of the mayor, kill him, get him promoted to governor or discredit him"), each of which leads to its own sub-goals ("get someone to bribe the mayor, then expose the mayor as corrupt"). At some point, the sub-goals become achievable, or can be farmed off to someone else ("give the mayor these 1,000 gold pieces and ask him to grant your pub an all-night licence").
The trouble with natural quests is that there simply aren't enough of them. More players want work than provide employment. If non-player characters (NPCs) could do the same natural quest thinking as players, though..?
Revelation: there are tried and trusted Artificial Intelligence (AI) methods that will do all this for you. Coming from a games background, though, most designers have a rather narrow view of what constitutes AI. It's what controls the enemy forces in Age of Empires II, or the party members in Baldur's Gate II, right?
Well, only in a relatively weak sense. AI can do much, much more than we see in games, but the field is partitioned up so that few AI engines integrate approaches, and those that do run slowly. Game AI is relatively simple because it has to be fast.
The sub-field of AI that is of use for quests is Planning. When an NPC's quest button is pressed, it generates a goal from the current context and its own predetermined personality. This could be practically anything, from short-term "visit my sick mother" or "repair my house after the storm" to long-term "become king" or "retire rich". The AI machinery creates a plan to achieve the goal, which consists of a tree of sub-goals, the leaves of which are the actions that have to be performed to cause the goal immediately above to occur. The machinery to do this is well understood in AI planning circles, at least for the relatively uncomplicated plans we need for quests.
[The actual mechanism is a search through a space of planning operators - of which several hundred would be needed - for those who wish to follow this up].
To get the quest, the NPC's plan is examined and a node (i.e. a goal) is selected appropriate to the player's abilities. Example: a high-powered mage approaching a noble who wishes to become king might be asked to remove one of the king's supporters from the equation; a lowly merchant may simply be asked to supply a dozen first class warhorses and be given no inkling that this is part of some grandiose scheme.
The components of the goal tree have to be programmed up in terms of preconditions (what has to be true for the goal to be achieved) and effects (everything what the goal achieves). For example, the precondition for "become mayor" might be "win the mayoral election" and the effects might be "gain money, gain house, gain servants, gain prestige, gain vote on shire council, gain ability to issue pub licences". The beauty is that these goals/actions can be programmed independently in their hundreds to offer a genuinely wide variety of coherent, intricate yet different plans that all make sense in the context of the game.
Aside: if NPCs can press each other's quest buttons, you don't even need the players..!
EconomicsSome people really like to make things in-game - the "crafters". They want the things that they make to be useful. OK, this isn't so bad; stuff can be programmed to be consumable (e.g. it wears out) so that a demand always exists.
The trouble is, crafters want their stuff to sell. Always. No matter how saturated the market is. This is so bad...
The problem is easy to alleviate, though, if:
Players will still complain that they spent all day manufacturing leather belts and no-one bought them so "the game" (or, worse, the specific customer service representative they harangue about it) wasted their time. The game did no such thing, of course - they wasted their own time. At least they only wasted a day, though. Next time, they'll know to make something different. If they only like making belts, well, the act of creation should therefore be its own reward...
As with many other of the problems I'm describing here, a lot of this is to do with designers being forced into making quick fixes in response to player complaints. Having set the precedent, changing the "if you can make it, you can sell it" rule could be perceived as an erosion of hard-won player rights - albeit "rights" they shouldn't have had in the first place. Because of this, designers are leery of doing things differently - they don't want to offend potential players. As I've said before, though, these are players they're not going to get anyway so it doesn't matter a hoot if they're offended by a return to sane micro-economics.
On the macro-economics side, the current model is the faucet/drain. Money comes in from the game, money goes out to the game, but the money that goes out has no effect on the money that comes in. In cases where closed, balanced economies have been implemented, thus far they've collapsed due to a combination of bugs, complexity and some bizarre behaviour by the players. People end up ignoring the official currency, instead resorting to bartering with goods that have some meaningful value to them.
Although a full, balanced economy remains the ideal, getting such a beast in place in a game is not something that's likely to occur for quite a while. Nevertheless, there are ways to add more balance to the faucet/drain model that should prevent any immediate massive devaluation of the currency. Primary among these is the levying of taxes on people who have large sums of money or large quantities of goods. If someone wishes to keep a collection of 10,000 shirts in their house, let them, but once a game year they're going to have to pay for the privilege.
Making objects consumable answers a lot of problems, too. Objects represent wealth, because they can be bought and sold and therefore have value. Someone who creates an object guaranteed to be bought by the game is therefore literally printing money. It thus makes sense only to guarantee the sale if the player takes some kind of risk. An example of this might be to locate an import/export market in the badlands where PD can occur. A merchant who buys all your statuettes at 10 GP each because they can be sold at 50 GP in the badlands could become very rich or very dead very quickly.
AdvancementThe established model for characters in massively multiplayer online games is drawn mainly from face-to-face role-playing games such as Dungeons and Dragons. All characters have a name, gender, race and class. Allied to these is a set of skills and abilities that state what characters can and can't do. Magic-users cannot generally wear metal armour or use swords, for example, and elves get bonuses for archery because of their invariably keen eyesight.
Although intended to promote game balance and role-playing, this approach does neither. Balance is only an issue if a game world is so impoverished of depth that everyone follows the same career path unless forced to do otherwise. Role-playing is about exploring personal identity, and therefore restricting the scope of the exploration restricts the role-playing possibilities.
There should be no character classes, just character kits that set up your initial skill set so you can start off as a monk even if you later become a merchant. Skills should be universal, so you can become jack of all trades or master of one, your choice. There should be no skill caps: if you keep learning how to make pots, you'll keep getting better at it - even if only by a tiny fraction.
Physical variations should be minimised. Games are quite at pains to point out that there are only cosmetic differences between the genders and that men aren't stronger and women aren't more dextrous, yet when it comes to race they really lay it on. There are all kinds of advantages and disadvantages to being an elf, dwarf, halfling, lizardman (or whatever names they use instead).
Firstly, these are just classes by another name.
Secondly, they aren't races anyway. What the general public means by "race" is reflected in the game only by the button that determines skin tone. Although a purist might argue that this kind of race in a game could accurately be reflected by sets of bonuses such as "+10% heat resistance" or "-10% long-distance vision", this carries an unacceptable smack of racism and would be roundly condemned. However, having "races" of elves, dwarfs and so on with hugely differing abilities also validates the opinion that stereotyping individuals by race is OK, which it isn't. For monsters and NPCs, yes, but for players, no. Drop it, guys.
Having other sub-species of humanoids is fine cosmetically, but differentiating ability and skill scores based on sub-species is unnecessary for role-playing and inadvisable for real life.
Character advancement should be based on actions, not goals. The more directions that players can take in their exploration of personal identity, the more they'll see. Character kits can set newbies off on a journey, but no-one should have to follow any preordained path.
If players can be whatever they wish to be, the issue of "balance" goes away.
ConclusionMassively multiplayer games do have problems, but there are accessible solutions.
Unfortunately, the designers of the second wave of games often don't seem to understand the problems (or even that there are problems!). They follow the wrong precedents, and they think they're infallible.
So much applies to any type of game designer, of course! Massively multiplayer designers have it worse because they often attribute the success of their games too much to the design and not enough to the players. These games are naturally sticky. Just because thousands of people play it, it doesn't mean it's any good. 400,000 people play EverQuest, but 600,000 other people who bought the boxed set don't play it.
The second wave of designers also, in general, design for themselves rather than for their players. They don't think through the full consequences of their ideas, and (under the influence of marketing and customer support people) place way too much emphasis on the opinions of gaming lobbies.
The result is that we're getting games that players think they want, rather than the games they actually do want. The more we go down this road, the worse it's going to get.
Never forget, people: these are worlds, not games.
They work differently!
7th September :\webdes~1\ og01.htm