Skip navigation

In-Game Professions in MMOs

Andrew Kozloski

It is very easy to lose sight of the fact that video games, as an art, are still in their infancy. Like film, and unlike books, video games are completely technologically dependent, and each new wave of technology brings exponentially more possibility. Looking back at the amazing shift in depth that we’ve seen coming from Pac-Man to Heavy Rain, it is easy for non-industry folks to imagine that the current state of video games is the pinnacle of the art form. In fact, it is probably more comparable to the moment when early films added audio to their repertoire.

With this paper, I’d like to look at one aspect of what is yet to come. I’d like to make the case that the next great leap forward in the art of video games will be the concrescence of MMOs and Open World games. This is a big claim, but I think I can show why this is not at all as complicated as it sounds and how it is actually a pretty logical progression from what is happening today. Lastly, I’d like to make a few stabs at how it could be handled in the first wave of attempts.

I propose to start out by examining two of the more serious problems existing at the current state of the art of video games, and then I’ll explain how the concrescence of MMOs and Open World will reduce or even nullify these problems.

Artificial Intelligence

Artificial Intelligence still sucks.

Despite the truly staggering forward advances of technology, we are still a long way from making computers think like people. We provide new challenges for hardcore gamers by giving the AI unfair advantages, such as perfect knowledge of the entire map. If the AI were limited to the input available to a human player, it would perform terribly in comparison. Even with these unfair advantages, it is not possible to make an AI opponent that comes remotely close to the challenge of playing against a skilled human.

These days, Starcraft II has replaced chess as the ultimate benchmark for computer AI. The Expressive Intelligence Studio from University of Santa Cruz in California created a year-long challenge in 2010 to build better AI for the game[i] (as many of us know, the AI that comes packaged with the game is not very good at all). While many bots were created that soundly beat the average player, none of them were capable of beating pro players, and to this day the world’s top programmers are still trying. An in-depth discussion[ii] of the various AI attempts can be found online at Team Liquid (a site run by professional SC2 players).

But that’s just the beginning of how terrible AI still is! Making an AI that can beat professional Starcraft II players is a remarkably simple task in comparison to creating AIs that can believably pass themselves off as real people. To this day, many games are released where AI enemies have trouble avoiding obstacles 100% of the time, never mind trying to engage in realistic dialogue on the fly. We are still a very long way from having NPCs that are fun to talk to, or even who can respond intelligently to simple questions without relying on pre-written branching narratives (which themselves are rife with problems).

Since we can’t fast forward AI technology advance in the real world, is there a way that we can get rid of the heavy dependence on AIs altogether?

Narration vs. Player Choice

It has been pointed out by numerous authors (and it should be self-evident anyway) that narration and player choice are at odds with one another. However, it is not the place of this paper to become bogged down in the famous narratology vs. ludology debate[iii]. I merely want to point out one subset of this vast group of issues—the logistical fact that the more possible choices the player can make, the more complex the narrative must be, and that beyond a certain point, the famous combinatorial explosion[iv] makes further progress impossible.

In his online thesis, Jakub Majewski offers a very thorough exploration[v] of the difficulties in video game narratives and many of the techniques used to create the appearance of more robust player choices. Though slightly dated, this general problem has not significantly changed since 2003.

In many cases, the narrative even constrains the physical path of the player, because the events happen in a particular place, and the player must go to that place. There are all sorts of tricks for making the world appear like the player could go anywhere while actually shuttling him inexorably down the One True Path. MMOs and Open World games avoid this one limitation by letting the player go anywhere.

So in Open World games the player can go anywhere, but in most of those places, nothing happens. Due to the combinatorial explosion problem, the game cannot be made chock-full of places where meaningful things happen. So the developers must use narrative pressures to make the player eventually go to the next spot where the next piece of narrative is doled out.

After that, developers these days almost universally sidestep the problem of too much choice by recombining their branching narratives to reduce the actual choice allowed to the player, as pointed out by Lucien Soulban[vi] (among others too numerous to mention). The player in modern Open World games, regardless of choices, must somehow still be shuttled into one of several pre-baked endings, in which case, the choices she was allowed to make were not very meaningful after all.

In MMOs as well, the player can go anywhere, but MMOs then have to contend with other restrictions on a narrative level. Like Open World games, because there is no rigidly-controlled path, they have to have a set of tricks for doling out the narrative one step at a time from meaningful narrative hubs. But then, on top of that complication, hundreds of thousands of people are playing the game at the same time, are involved in the same over-arching linear storyline—but always out of phase with each other.

It is taken for granted by modern MMO players that the same event is happening over and over millions of times but is intended to be understood as happening for all of us collectively, all at the same time. This becomes especially pronounced when friends play together and are out of phase on in-game quests. Like children playing make-believe in the back yard, we just have to pretend that the game-world events are happening to everybody at the same time.

MMO + Open World

The MMO and Open World game types are two distinct halves of what, I believe, will one day be a leading paradigm for video games. With the ever-increasing trends in crowd-sourcing and client-provided content everywhere on the web, it is easy to imagine that, perhaps in 20 years, the narrative aspect of MMOs will be either almost-entirely player-driven or created as collaborations between players and designers. The ideas that I will present here are only a few examples of the first movements that could be made in this direction.

We can take a major leap forward into true branching narratives in video games if we simply share the narrative process a little more with players. At the same time, we can eliminate the need for a significant portion of the NPCs that populate the game-world by having real players act out these roles in their “downtime” between raids and serial questing. What’s better, we can implement a lot of the early stages of this progress without additional work or monetary investment, simply by shifting the way we look at the design of these games.

The examples I’ll be giving will be most easily applicable to World of Warcraft, because I know this MMO the most intimately, but with a little imagination these principles could be applied to any MMO.


One of the more simple and straightforward ways to accomplish this merging of genres is by allowing players to seek out and accept in-game professions. In doing so, we eliminate the empty AI placeholders that were fulfilling these functions and we simultaneously create numerous opportunities for the players to become more directly involved in shaping the narrative of the game world. We can greatly amplify this secondary effect by shifting some of our quest-writing resources from generic quests to quests that are tailored to the experiences of being in those particular professions that are now held by living players.

I’d like to go into a bit more detail by proposing two possible professions and describing how I see them being implemented in future games.

Guards on the City Watch

In addition to whatever class a character has in the game (if any exist), she can sign up to become a city guard. Obviously, this already is a great opportunity to create a small handful of quests that actually contribute to the in-game narrative rather than an opportunity to exchange 35 Silken Horsefeathers for 10 gold. At the end of this small suite of quests, the player is named a city guard of the lowest rank and is given a small package of appropriate benefits. The benefits could largely be things like improved defense while in the city or similar relevant perks, but there is also opportunity here to tie the players more intimately into the role and the game-world at large.

One possibility is a private chat channel that is only accessible by the City Watch, and only while in the city. With it, all of the people playing at being city guards could coordinate their defense of the city (and good designers would find ways to make sure that the city is almost never entirely safe). At the same time, the city guard players could fulfill the secondary functions previously handled by the AI city guards, such as giving directions. They could use a system whereby they highlight the location on the other player’s map, or they could lead them in person.

This creates all kinds of opportunities for increased social interaction, and a good idea could be to add a social vote where players asking the guards for defense or even just directions could nominate the guard player for faster advancement and increased social perks like status titles and special clothing/mounts. This way, guard players are encouraged to contribute an added layer of social interaction to the game but can still advance slightly more slowly by doing guard quests if this does not appeal to them.

With all this, a skeleton crew of AI guards could be maintained in the city, with some at fixed posts for those who just want to know where to go rapidly and avoid this social layer of the game. The number of AI guards could also increase and decrease reactively to balance out with respect to the number of player guards present in the city at any given time.


As a logical progression of the extensive crafting systems available in many MMOs, we could offer the players hubs in which to ply their trade. There could be in-game guild-houses organized by profession (i.e. The Smiths Guild), where players band together to pool resources, craft items, and sell them to other players. Players would be given access to the pool of resources as a function of how much money and/or crafting materials they have contributed to the common cause.

As with the city guard example, existing quest designers could be shifted from creating generic quests to creating branching sets of quests that relate directly to the advancement of the crafting professions and lead to advances in status as an in-game merchant. Not only does this increase the number of meaningful quests in the game, but it encourages participation in the crafting systems, thereby protecting and adding value to the considerable time and design resources that went into creating those systems.

Socially-minded players could peruse the guilds individually, buying directly from the guild’s stocks through an AI (because in this case, it’s just easier to use an AI), but they could also place orders for items not in stock. These items could go into a queue, to be crafted by guild members and then delivered to the mail of the player who ordered it. Happy customers, perhaps those who receive their orders really rapidly after making them) could nominate their player-merchant counterparts for social advancement. Players disinterested in this social dimension of the game could still go to the mass auction house and just rapidly get everything they want without fuss (although it should be noted that it is already the norm in MMOs that players must seek out player crafters to obtain things that are not already present in the auction house).

One great example of a quest/event built around this structure is a “battleground” where a large quantity of goods could be loaded onto a ship or caravan for transport to another city. (Perhaps copper is to be found in abundance around the rocky cliffs near city A, but almost nonexistent near city B). The quest would be to protect the caravan from another faction of players acting as pirates or bandits. If the bandits win, they win the goods, which get divided among them. If they lose, the shipment makes it safely to port and the merchants/guards are paid handsomely for the transport (also winning a huge stock of reagents or materials that are difficult to obtain in the city where they work).

Any guild member can put goods into the caravan, which would also be stocked with a bare bones shipment by default in case no one put anything in. The loss or success of this operation would directly affect the materials available to the guild to which it was being delivered, and those players who contributed goods would get additional rewards as function of what they contributed. In a sense, it’s a way for players to place bets on the success or loss of a battleground, increasing their feeling of participation.

Also in this way, the battlegrounds tie in directly with player-controlled crafting guilds, the city watch and the thieves’ guild or a faction of pirates/bandits, thereby adding value to all of these systems. The merchants, guards and pirates would have the most riding on the outcome of these events, but they would obviously still be open to any players in the game. And even merchants/guards could get their chance to be bandits by attacking caravans from rival crafting guilds!

As in the previous example, a potential crew of AIs would exist on each side, increasing in moments of low player participation and vanishing altogether when both teams are full. Hubs in the city could transport players directly from downtown to below decks, like in existing models of in-game battlegrounds.


With these professions in mind, we can create a new class of quests that brings players directly into more involvement with each other and away from grinding AIs. Most of the battlegrounds in the game could be tied to specific game-world effects, for example, control of a copper mine could be a constant contention between the Smiths Guild and the Jewelers Guild. Then, every win or loss of this battleground is tipping the favor in one direction or the other. Non-guild players who just want to do a battleground could pick a side based on which goods they are more interested in buying at a given time, or based on the in-game urging of their partisan friends. Whichever guild controls the mine will get a modest bonus on copper collections for all its members during that time.

Rather than dungeons always being 5-people against a slew of AI monsters, certain instanced dungeons could be players vs. players from another faction. Perhaps there is a large castle at the edge of the city with unimaginable wonders in the vaults below. 5 members of the Thieves’ Guild come in a little-known back entrance and must find their way through the passages, past a team of 10 city guards and the traps that they have laid for them (having received forewarning of the breach). Again, any player could play this as either side, but those actually belonging to the related professions would get added benefits for doing so.

The castle could be composed of many passageways which are randomly blocked each time (so the thieves don’t know where to go). The guards are given a mini-map showing the blockages and one trap each from a list of different types, that they can place anywhere in the dungeon. The thieves can disarm traps (possibly a permanent ability belonging to in-game bandits but given specially to any players while being thieves in this dungeon) and they have access to a small set of ventilation ducts that are inaccessible to the guards, either for escape, or for getting around traps and blocked passages. Like the passages, the ventilation ducts will not always be the same, and they won’t be shown on the guards’ maps. This could be an instanced dungeon that is done as many times as desired with an AI staff at the ready to fill in for incomplete teams during times of low traffic.

There could also be open world events, where guards are protecting a stretch of forest from poachers connected to a guild of leatherworkers. Every killed deer is a mark against the guards and a small contribution of resources to the leather guild. Every killed poacher is a small bounty for the guards.


With a few additions to the current standard format of MMOs, it is possible to move one step closer to truly Open World branching narratives, all the while enriching the MMO genre as well. The games will be enhanced by fostering more involvement among players and by involving them more directly in mutually relevant narratives, driven by themselves rather than static interactions with AIs. All of this means more player involvement, more possibilities, and with these things come a longer life for the game.


[i] Studio for Expressive Intelligence Website (UCSC)

[ii] Discussion of SC2 AIs @ Team Liquid. 2012.

[iii] Frasca, G. Ludology Meets Narratology. 1999.

[v] Majewski, J. Theorising Video Game Narrative. 2003.

[vi] Soulban, L. Writing for Video Games.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: