Massively multiplayer (MMP) games are built on a social foundation. They differ from other games in that they encourage people to do fun things (adventure!) as part of a group, and to keep up their group ties over months and years. These games lure players in with visions of heroic battles and individual character advancement, but ultimately succeed when they provide meaningful, interdependent roles for people to play as part of a group.
In this article, I'll provide some design principles to follow based on how current MMP games are encouraging (or not) players to be part of groups within and around the game. In particular, we'll look at temporary and permanent groups, designed and emergent roles within groups, and where this might lead in the future.
The core idea is that anything you can do as a designer to encourage players to form groups and take on meaningful roles within them will increase the likelihood that those players will enjoy themselves enough to play your game for months or years to come. Conversely, anything your game does that makes it more difficult for people to get together or form social bonds will drive players away: For most people, the only thing lonelier than playing a game by yourself is playing by yourself while surrounded by a virtual world full of other people.
Support Temporary Groups
Temporary groups -- adventuring parties in most games (fellowships in Asheron's Call) -- are the most common catalyst for making friends in the game, and in many ways form the social backbone of MMP games. Players need to be able to find each other easily and quickly, "buddy up" to go out into the dangerous world, and complement each others' characters' abilities. Adventuring parties throw players together in a low-commitment context, setting the stage for more permanent connections later on.
Before players can get together to go adventuring, they have to be able to find each other and decide whether they want to join up. The first obstacle here can be simply geographic: most MMPs have large worlds to explore. Unfortunately, this can make it much more difficult to find people to explore it with. In Asheron's Call, the time it takes to run from one city to another has been a persistent problem, especially for low-level characters -- the ones who most need the help getting together with others. If you have only an hour or so to play and it takes half that time to reach your friends, you can very quickly become frustrated.
The next problem is knowing who else is open to being in a group. Dark Age of Camelot addresses this by allowing players to flag themselves as looking for a party or not, and to automatically look for others nearby of approximately the same level. Even with this, many new players find it difficult to locate others with whom to form a party, but at least they aren't reduced to shouting that they're "looking for a 3rd level fighter" as in other games.
There are many potential solutions to the issue of players having trouble finding others to group with. The best of these make it as easy as possible on the players -- help them find others with complementary characters who are nearby and who are also looking for a group. Ideally, players should be able to find out this information before logging in to the game and running around. Asheron's Call provides ‘lobby' chat rooms for players, but this seems to be insufficient to help players get together. Among other things, you can still log in and find yourself across the continent from the person with whom you were just chatting. A better solution would be the ability for players to state where (i.e., what city) their character starts play in for this session. To avoid cheating, the game could allow an expanding list of session-starting cities based on how long the player had been logged out (that is, how far could your character have traveled during the time you weren't playing). There are many other possible solutions, all having the common thread of making it as easy as possible for players to find each other and form a temporary adventuring group.
Once two or more players have found each others' characters, it's typically easy for them to chat about where they want to go, where they think some good treasure is, what quests they have to complete, etc. This is a process of the players evaluating each other's character and whether their abilities are complementary. There's typically little cost in forming a party though, so most people aren't too picky unless they have a specific, difficult goal in mind. However, it is important that players be able to uniquely identify other characters not just by name, but also by ability and even reputation. Players want to know that the others in their party aren't going to take them out into the wilderness and leave them for dead (a popular tactic early in Ultima Online's history), and that in general the other character approximately matches their own play style. Games that flag characters as good or evil, or that have a more fine-tuned reputation system, help players decide whether another character is someone they want to join up with or not. (A full discussion of robust reputation systems is beyond the scope of this article, but this is something to consider seriously in any game where people interact closely.) After a player has identified others with whom they enjoy playing, being able to place their name (character or player) in a buddy list helps make future connections easier and more likely.
When the party's been formed, one central but easily overlooked aspect of their play together is the ability for the party to communicate privately. Most games today have private chat channels for temporary groups. Providing this enables the party members to talk amongst themselves without worrying about extraneous chatter around them. In socially active areas this can make it easier for the players to keep up their conversation, and provides a sense of solidarity within the group.
In addition to chat communication, most current MMP games enable party members to see each others' health and other vital statistics. This makes it easier for players to help each other out (with a timely healing spell, for example). In addition to increasing the party's efficiency in combat, this also creates social bonds between party members who save each other in the nick of time. But if your game allows player-vs-player combat, one member of a party should not able to harm another when operating together as a party. This is the mechanical implementation of "all for one and one for all," whereby no one character can -- mistakenly or not -- harm another party member. This helps reduce tension when fighting monsters in close quarters, and prevents party-destroying sneak attacks between party members.
In every MMP game, players create characters that specialize in some way. Building on now-classic fantasy character archetypes, you're typically a fighter, wizard, healer, or some variation on those (including science fiction and other genre variations). Each character fills a specialized role from the beginning. Put a few different character types together, and you have an effective adventuring group -- a fighter or two to whack things and absorb monster damage, a wizard to provide high-damage attacks, and a healer to keep everyone alive. In some games such as Everquest and Dark Age of Camelot, these roles are not only well-defined, but accepted to the point that coordination occurs via highly condensed jargon. For example, a statement like this would not be unusual and would be understood by everyone in the party: "Thane pull, tanks rotate aggro, runemaster nuke but do not draw aggro, watch out for adds. Protect healer; rezz only after fight over." These are battle orders issued just before the fight begins that give everyone their part: the thane (a specialized fighter-mage) makes sure that only one monster out of a group ‘goes aggro' (aggressive) and approaches the party; the damage-absorbing fighters work to split the monsters' attacks; the spell-caster does as much damage as possible without having the monsters target him; and everyone protects the healer (the only one who can bring those who die in battle back to life by resurrecting or rezzing them after it's all over).
Such characters' abilities are useful on their own, but when paired up with other members of a party they become part of an overall strategy available only to diverse groups. These strategies act as a multiplier, increasing the effectiveness and benefit of being in a group. Players have been quick to explore and exercise these game mechanics, and to assemble groups of complementary character types where each player understands their role. By creating such complementary roles in the game mechanics, you give people multiple ways to group together effectively. Each character type handles situations the others can't. Of course, this means that the world has to support these roles: not every monster can be killed by whacking it; some have to be dispatched with magic, or have a rogue sneak up on them.
If in the rest of the game you provide opportunities for each character type to shine, you'll keep the players satisfied and feeling like their character is a vital part of whatever mission or adventure they attempt. This extends beyond just fighting styles too. Characters with crafting skills should find that they are able to produce higher quality items when they combine their abilities together rather than trying to do everything themselves. For example, blacksmiths can make basic swords and armor, but require the services of carpenters and leatherworkers to make higher quality items. These kinds of interdependencies can also extend beyond crafting to other types of group efforts as well (see the discussion on New Types of Groups below).
Rewarding Players in Groups
Even though people will naturally want to adventure together, enhancing this experience will encourage them to do so. Everquest encourages players to form parties by basically making the game too difficult to play otherwise. While some parts of your game should be difficult enough that they require a group effort, this is more of a ‘stick' than a ‘carrot' approach. Asheron's Call introduced the novel approach of rewarding players in fellowships by increasing the experience points earned when fighting together. This is a positive-sum rather than zero-sum approach that promotes working together as part of a fellowship; that is, instead of me earning 200 points for killing a monster, or each of us earning 100 points, we might each earn 150 points if we kill the monster while grouped together. This encourages players to take on greater challenges as part of a group, and has proved to be an effective reward for adventuring together.
Other games such as Diablo II provide examples of other forms of rewards for playing as part of a group. In that game, both the number of monsters being fought and the amount of treasure gained increases with the number of people playing together. More interestingly, this game also scales the type of treasure gained to how many people are playing together, so that the rarest items are available only if you are playing with others. This kind of gameplay would no doubt be welcomed by the players in traditional MMP games, and would further encourage them to form temporary groups.
Support Permanent Groups
Temporary groups often lead players to want to join together more permanently. Once you've been adventuring with someone and seen that they're easy to get along with and react well under pressure, you're more likely to want to get together with them again in the future. Permanent groups give the players a way to create long-term associations within the game, provide a strong sense of belonging, and create their own forms of gameplay.
The most common form of permanent groups in MMP games are ‘guilds' -- loosely based on medieval craft associations, and more directly based on social or character-class-based guilds in text MUDs. All of the MMP games today allow players to create and join guilds or similar groups. In Asheron's Call, the guild is called an allegiance, and differs from guilds in other games in specific, novel ways. In particular, AC allegiances rely exclusively on a patron-vassal pyramid system, which has some interesting effects on how the groups operate (we'll come back to this later). However, both guilds and allegiances offer design principles for all sorts of permanent groups.
A Sense of Belonging
One of the main things that guilds provide is a sense of belonging, of being part of something larger than you. This feeds a powerful and deep desire most people have. Membership in permanent groups is the main reason why many people remain in MMP games after they've exhausted the adventuring gameplay.
To aid in this sense of belonging, players typically name their guilds -- that is, the person who starts the guild gives it a descriptive name. This sets it apart from others and gives the members an easy handle to use in identifying themselves. In addition, visible methods of identifying other guild members are an important aspect of creating a shared sense of belonging. For example, in Ultima Online, many guilds adopt a certain dress or certain colors (clothes can be dyed to suit the player's wishes). In Dark Age of Camelot, players are able to create a guild crest, or colored symbol for their guild. This has real-world analogs ranging from medieval heraldry to modern sports teams, and is a great way to allow each guild to have its own visible, shared identity. Being able to easily identify others in your group is an important way to foster a sense of belonging.
Communication and Infrastructure
Guilds also benefit from having infrastructural supports within the game. Most games provide guild-based chat channels, for example. Some, like Meridian 59, provide guild-specific bulletin boards and coveted ‘guild hall' locations. Guilds in Ultima Online and Asheron's Call often use large keeps and manor houses for this same purpose. These in-game structures acknowledge the guild's permanence and importance in the world, which in turn increases the players' sense of belonging and commitment to the guild. In M59, the guild halls are so desired by the players that they have been the focus of many guild wars -- a good example of gameplay arising out of the importance players attach to their social groups.
In addition to in-game infrastructure, it's common for guilds to have their own out-of-game web pages where the members congregate, post messages, and plan meetings and adventures. These are often elaborate sites attesting to the group members' dedication. Providing in-game links to these sites adds to the identity shared amongst members of the group. These links might be displayed in a central registry or, as in M59, on the individual characters' identification panel. Linking to these pages from the game's main web page also shows that you take the players' guilds, and thus their contributions to the game, as seriously as they do.
Forming Groups and Changing Leadership
All the current MMPs allow players to create new guilds or other groups, if you can meet certain requirements: in Ultima Online, you have to own a house first, for example, and in Dark Age of Camelot you have to have eight people to join your guild. In Asheron's Call, anyone can be a patron with vassals, but in practice most monarchs (group leaders without a patron, and with several levels of vassals beneath them) are powerful, high-level characters.
However, in most of the current games once a guild is formed it's internal structure never changes. In these games, if the guild leader decides to disband the group or simply leaves the game, the entire group falls apart. In such cases, the other members are left to scramble to re-assemble their group, which may or may not work out. This makes the permanent group structure brittle, leading to situations in which a guild fails because one person -- the founder -- leaves the game (or even just the guild). This causes frustration and dissatisfaction amongst many players, and weakens the utility of permanent groups in the game.
In contrast, in Asheron's Call it is possible for any character with vassals to sever his ties with his own patron, and thus become monarch of his own allegiance pyramid. Conversely, a monarch could quit the game, leaving her vassals to re-form their allegiance as they saw fit. In such cases patron-vassal relationships lower in the pyramid are unaffected, and the allegiance as a whole continues on. In a similar way guild members in Meridian 59 can elect a new guild leader at any time, deposing an absent or unpopular leader whenever more than 50 percent of the members put their political support behind someone else. The game automatically bestows all the leadership abilities on the new leader, who can then demote or even expel the deposed one.
Methods like these, where the overall structure of the permanent group or guild is preserved across leadership changes, makes the group much more resilient to change, and thus more likely to carry on over time. In other words, it's not enough to let your players form their own groups; they have to be able to manage them as well. Providing simple but powerful methods that enable the group members to do this gives them a greater sense of ownership, binding them more closely to the group and to your game.
Guild and group management is a fertile area for finding new ways to improve the players' experience. Asheron's Call's allegiance structure as a substitute for guilds is an interesting innovation. It makes the entire group hierarchical, and essentially institutionalizes the practice of ‘twinking' -- that is, giving a lower-level character money or equipment in exchange for something; in this case the exchange is that a patron earns a small amount of all the experience gained by their vassals. This is a novel form of reward for entering into a group relationship that should be explored further. However, AC's allegiances also have significant downsides: all person-person connections are vertical, going from patron to vassal, rather than being horizontal, from group member to member. This means that two people who are vassals to the same patron or monarch are essentially strangers with little shared identity, and without any reason to help each other. They often do not develop the same feelings of cohesion commonly found in guilds in other games.
Another area of innovation in games such as Ultima Online and Asheron's Call are provisions for player-owned houses. Housing has turned out to be phenomenally successful for UO, and slightly less so for AC. The difference may be that in UO, players can design and place their own homes, while in AC you can only buy those that are pre-existing, made and placed by the designers. This may not seem to be that big of a difference, but players feel much more control and ownership of their houses in UO, and they have thus become a more desirable part of the game.
On the other hand, pre-set limits in the game are not necessarily bad. Everquest has innovated in this area by dividing characters by race (troll vs. elf, for example), with biases for or against other races, and with a limited ability to speak to other races. This creates externally imposed divisions between groups of players, but encourages intra-group roles to emerge.
This type of grouping-by-division is taken even further in Dark Age of Camelot, where characters from each of the three Realms (Albion, Hibernia, and Midgard) cannot communicate with -- and are necessarily at war with -- those from the other realms. This "Realm vs. Realm" (RvR) play is the core of the game once your character has more or less exhausted the fun of killing monsters and completing most quests (level 30 or so). Players slide naturally into this as part of their character's identity is as a member of their Realm, even though there is little to be gained by ‘winning' this part of the game. Still, groups of characters form in the frontier-lands, waiting to either defend a keep against foes from another Realm, or sallying forth to increase the glory of their own Realm.
Finally, there are other possible ways in which players can interact within permanent groups, as for example via political meta-games. Both M59 and UO attempt these using political factions that are separate from player guilds. In UO the faction-game has not gathered a strong following as it has no real effect on the world other than allowing for guild-wars. In M59, the political game has been much more popular. In this meta-game, players are able to interact with NPCs to indirectly affect their opinions. This eventually puts one faction or another ‘in power,' resulting in in-game rewards such as better prices or cheaper spells for that faction. While the gameplay and rewards are relatively simple, players have banded together and spent a great deal of time and effort to ensure that their faction -- and thus the permanent group of which they are a member -- remains in power as long as possible.
Support a Broad Range of Group Roles
Within both temporary and permanent groups, players need a variety of valuable roles to choose from that fit their goals and play style (note that these roles need not correspond to character skills or classes). Games that implement only typical fantasy roles (fighter, wizard, etc.) are quickly seen by many who are not die-hard fans as being narrow in their scope. Ultima Online has retained many players who see their characters as tailors, carpenters, or miners -- roles that have value in the game, and which extend beyond the typical fantasy archetypes. This breadth supports more players than do more narrowly defined roles, and increases the chances that people will form strong social bonds.
Each role that a character may take on must have a valid and valuable function within the game; those that have no discernable function will be largely ignored. Giving characters the ability to mine ore is irrelevant unless the ore has some value in the game -- such as being turned into ingots which can be hammered into armor and weapons. Not all roles have to have overt gameplay value such as this, but each must have a direct perceived value for the players. In some cases this can be a social value: being a guild leader, tavern owner, or even just an accomplished story teller are all potentially valuable functional roles. The broader the set of functional roles your game supports, the denser its gameplay and social web will be.
Avoid the Uber-Role
In-game character roles must remain synergistic throughout the game. In UO for example, many players eventually gravitate toward what's called the "tank mage"-- a high-level character with enough skills to throw spells with the best of them, and then pull out a big weapon and wade into combat. The problem with this is that it reduces the need for getting together with others; such a character is essentially his or her own group. In contrast, Everquest players are keenly aware of the continued need for grouping together, as there is no one super-character type. However, even here, the game too often doesn't present monsters or quests requiring a wide range of different character types, maintaining instead the "whack monster, get gold" method throughout. By focusing not only on the combinations of character types and roles, but also on how they are used in the game (situations that are impregnable to fighters or magic users, but where a thief or alchemist or even a healer might be just the thing required), you can encourage players to explore all the possible roles, and include them in their groups.
The types of roles described above work as part of groups, but are still focused on individual gameplay. Other roles focus more on maintaining the groups themselves -- both temporary (parties) and permanent (guild) groups. Adding these roles and connecting them to the rest of your game will give players another often neglected area in which to shine.
In adventuring parties and other temporary groups, many players like having a designated group leader. Others like a more ad hoc approach, but adding leadership capabilities to the group can enhance the players' experience and add new dimensions to their play. For example, the group leader might be the only one who can let new people into the party or toss them out. This formalizes what often happens in parties anyway, where one person takes the lead, gives people their battle orders, etc. By allowing (not forcing) parties to select a leader in a way that's known to the game, you can also make this part of the game: a traditional paladin-type character might have greater leadership skill, which, when he or she is the party leader, adds to everyone's attack bonuses and/or morale. Or perhaps when a thief is the leader, everyone gets the benefit of moving a bit more silently. Improving individual gameplay via group roles is a great way to add to the effectiveness of groups and move beyond alone-in-a-crowd MMP gameplay.
There are many opportunities for group roles in guilds and other permanent groups, most of which are still untapped in today's games. As discussed above, allowing the guild or group to designate their own leader (rather than just having the leader always be the person who started the guild) creates many more dynamic group interactions. This helps the group be more resilient and last longer. It also gives players something to focus on beyond just killing more monsters, since those who want to can take a more active hand in helping to manage their guild. You can also allow the players to designate ranks or titles to go along with group functions. In M59, each guild leader can designate up to five guild ranks (each with its own title), and each with its own functions in the guild. Some can promote or demote others, some can bring lower-ranking members into the guild's hall, and only the leader can change the hall's password. Such group roles within the guild add an entirely new layer of gameplay to the game.
So far we've focused on roles that you as the designer put into the game for the players to fill. While you could just let all the group roles emerge based on the players' actions, designed-in roles are valuable because they channel the players' efforts and actions to make the game more fun for everyone. Players will inevitably do what they are rewarded for (whether or not those rewards are intentional on your part), so carefully designing in a variety of group roles makes for better gameplay that is more appealing for a longer period of time.
At the same time, as players become ensconced in the game world you've provided, they will find and create their own types of roles. These are valuable to them as they provide a sense of ownership and thus attachment to the world. They're valuable to you as they show you what the players consider important, and thus what you might want to focus on expanding in the future.
There are a wide variety of social and group-related roles that have emerged in MMP games. Some, such as UO's tavern-owners, appear because of the unique combination of social and gameplay factors in that game. But such roles are also pointers to more general types that appear in all games, and which your game should support. This may require little action on your part, but on the other hand, you may find novel ways to include these emergent roles in your gameplay.
The first of these roles to consider is the newbie -- the new or unfamiliar player. While this type of person is almost by definition not part of a group, how well your game accepts them in and provides methods for them to access meaningful group roles will be a gating factor to long-term play. People who come into an MMP game and don't have an opportunity to interact positively with anyone else in the first few minutes are likely to just leave and not come back.
Newbies need information and items -- they need help becoming familiar with the game and becoming confident that they can have fun within it. Some games like M59 and UO provide an entire "newbie experience" in a relatively safe area, to give new players a gentle introduction into a complex game. Others like Asheron's Call make helping out newbies an important part of the game for anyone who wants a strong allegiance (though this has to some degree backfired as new players get pelted with unsolicited "do you want a patron?" messages). Since every player will go through a newbie stage, this is an area where future games can focus to greatly enhance most players' experiences.
This also leads to the next set of emergent roles filled by veteran players. The first of these is the docent, the type of kind, patient, informative guide that every newbie hopes to find. Some people gravitate toward this role on their own, and get praised for it by the otherwise hapless newbies whom they help. MMP games are beginning to include this role -- AC patrons are often docents, for example -- but it is still largely unexplored as a formal part of the game. There are many opportunities to strengthen the in-game society by recognizing and rewarding those who act in ways that help and guide others.
Another veteran role similar to the docent is the trail guide. This type of person is typically an "explorer" in Bartle's typology, one who knows how to get into and out of difficult and dangerous places in the world, who can provide guidance on quests or other tasks in the game. In some games like Everquest, this is the kind of person who will often help you retrieve your corpse (a necessary task after you die in the wilderness), or who will lead your party through a dangerous zone. As with the docent however, this role is almost entirely emergent. The trail guide is an important role that crosses between many temporary and permanent groups, and yet one that has not been codified in a way that makes the players' natural inclinations more meaningful (that is, directly rewarded) within the game.
Finally, there is a broad and diverse class of innkeepers -- social instigators who run make-believe taverns in UO, set up scavenger hunts in many games, perform marriage ceremonies between characters, or find other ways to bring people together in some emergent fashion not related to a guild, standing quest, or other designed group element. These people are incredibly valuable as they really make your game come alive for other players. They involve others in ways that are both unique and are still part of the game world. This increases other players' immersion and connection with the world, and adds a valuable social role for those who may want a change of pace from the primary gameplay. While various games have had centralized volunteer programs, innkeepers (as a role, not an occupation) have not been supported with in-game rewards. One way to do this would be to expand the set of functional roles, as described above, to include economic, political, and entertainment roles. Various skills and reputation systems could also add in-game visibility to the innkeeper set of roles.
Look to the Future
As MMP games innovate on current gameplay mechanisms and group roles, what can we expect to see? Clearly parties and guilds work for many of the existing players, but this and other aspects of MUD heritage have probably been taken about as far as they can go. As MMP games try to attract and keep more players, and as they move into the mainstream of online entertainment, we'll need a wider variety of groups and roles for players to fill and new synergies among them.
As mentioned earlier, there are many ways player interactions within groups can be made more rewarding, and thus more attractive. AC's positive-sum reward for fighting together as a fellowship could be extended to include guild members who group together, giving them a bonus to their experience, skills, or treasure. This would tie the two types of groups more closely together, and encourage players to use their guilds as a springboard for adventuring together.
From the guild point of view, players could be encouraged (via in-game rewards) to select their guild not only for social factors, but for the benefit the group brings to their character. In the case of AC's hierarchical allegiances, they might be more successful if characters gained a discount on skills they could earn based on the monarch's skills, or the aggregate of everyone in the group's skills. This would encourage the creation of groups focused on archery, Life Magic, and other skill areas, and give players an in-game reason to consider joining one group over another. Turning this around, skills for running groups would also be useful. Using an earlier example, a paladin could have leadership skills that increase his party's morale or attack bonuses. The same principle could be applied to permanent groups: a guild leadership skill could increase the group's effectiveness -- in this case, the skill discount each member receives. A guild leader with guild leadership skills could increase each member's skill discount. Thus one character's skill benefits everyone in the group.
Another idea along these lines is to provide skills that can be gained only from within a group. Suppose that once the members of a guild attained a certain average level of skill (perhaps even coupled with a minimum level of group-leadership skill), new skills become available to members of the group. So when members of your archery guild show up and start firing guided-missile arrows, others immediately know your group (which they can identify by the symbol on your surcoats) is made up of true bow masters -- no one else could have such a devastating skill. This sort of innovation in skills and group management binds groups more tightly to the game, and makes them much more than just another form of in-game social club.
In terms of inter-group dynamics, there are many ways to think about how groups might interact with each other beyond the simple ‘guild war' model seen now. It's easy to imagine quests or other tasks that require a party to complete, but which pit two or more parties (or guilds) against one another to see who can complete the task first. Or, from the guild point of view, by making trade and crafting skills more a part of the game and then linking these to groups and group skills (analogous to the archery example above), groups could enter into symbiotic relationships: one guild of master smiths makes exemplary weapons and armor for others -- and in particular for another guild that has agreed to protect them against all foes for a deep discount.
Enabling symbiotic and other relationships between groups can also help avoid some of the problems seen in current games with guilds that overpower all others. That is, just as you want to provide multiple character roles (to avoid the ‘uber-role' as described above), gameplay that allows one guild to outpace all others can limit the enjoyment for everyone who isn't part of that select group. Currently in Everquest for example, success in the endgame depends entirely on being part of a large powerful guild. Players simply cannot undertake a high-level raid on their own or with less than a large, powerful group. Naturally enough, these raids tend to yield some of the best, most powerful loot, which then makes the guild even more powerful. This can create a positive feedback loop that creates what players call an ‘uberguild' that dominates all others. If instead guilds depended on each other just as characters do, rather than being entirely self-contained, there would be more opportunities for different forms of success, and less likelihood of one overpowering guild emerging.
New Types of Groups
With the example of the smiths' guild in mind, consider new types of groups that MMP games might support. Right now a guild is a guild; there are no real functional differences between them. Enabling the players to focus their group on one skill area (as described above) would allow them to differentiate their guild from others -- thus also leading to guild specialization and symbiosis. But beyond that, why not allow players to make permanent groups with explicit business or political purposes? For example, in UO an individual player can own a business that's run out of their house. But this could be a group-level affair as well, with the skills of multiple characters (the group members) affecting the quantity and quality of what's produced. This would open up new kinds of trade and crafting skills and add entirely new aspects of gameplay.
In the political arena, various games have tried player-run towns with limited success. This is a complex area, but one that can benefit from looking at neighborhoods, towns, regions, and other political constructs as groups (and even as groups-of-groups). Consider for example the group-related gameplay (and new opportunities for players to interact) that would come with a town-management group structure: the group leader is the local mayor or baron, and the members of the group are the citizens of the town. Among other duties, the mayor sets the tax rates on the citizens, which in turn determines the number of NPC-town guards protecting them. And of course, the citizens can remove an unpopular (or over-taxing) mayor, just as they could remove and replace any guild leader. This is just one possible example. There are many ways to explore applying groups and group roles to the political and social landscape within the game.
Beyond Individual Adventuring
A powerful way to move beyond current gameplay is to consider the group, not the individual, as the main unit of the game. That is, make the group the protagonist, not the individual character. You can see parts of this in some of the ideas above, such as skills that can be learned only through your group. But think of all the aspects of gameplay that are currently focused on the individual character and how they might apply better -- with more engaging and meaningful results -- to the guild or other group of characters. Can a guild gain experience via the actions of its members? This might be a good vehicle for deciding when a new guild-skill becomes available to its members. This also introduces a new dynamic in players evaluating which guild to join ("nope, your guild is too low level") and in guilds evaluating characters who want to join ("great, you have just the archery skill we're missing!").
Guilds could also be given quests to complete, just as individual characters are now. When parties of guild-members resolve the quest, the guild as a whole is rewarded. Similarly, the guild as a whole could be attacked -- this could lead to interesting new ways to think of guild wars, putting more at stake than just the individual members' (recurring) lives. A political group (representing a town and surrounding area) could lose territory and resources it controls if it cannot defend itself against raids from a rival clan -- just as an individual may lose possessions to the victor of a fight.
Finally, there are other unexplored areas such as skills that can be used only as part of a group. Some fighting styles or skills might be usable only with others -- consider the ancient Spartan phalanx as a multi-person battle tactic, a skill that cannot be used by an individual. You could also implement certain magical rituals that require one mage from each school of magic to complete. Other skills or abilities might be usable only in concert with other members of your guild (along the lines of The Three Musketeers who were great on their own, but dazzling together). These represent just a few examples of how groups and group roles may improve MMP games in the future.
Current MMP games struggle against turning into "massively single-player" games -- that is, games where there are a lot of others around, but you don't pay much attention to them. Creating and supporting effective groups and group roles is a primary way to address this issue. Better groups make for better social bonds and increased loyalty to the game. We are still just scratching the surface of how, with better designs, more varied roles, and broader group support, players can create great social and group experiences in MMP games. By applying and developing the principles discussed here, you'll be able to find even more ways to make this happen.