Welcome to Raph Koster's personal website: MMOs, gaming, writing, art, music, books.
Welcome to Raph Koster's personal website: MMOs, gaming, writing, art, music, books.

The whole Web
Raph's Website

These are full-blown essays, papers, and articles.

Slideshows and presentation materials from conferences.

Interviews and Panels
Reprints of non-game-specific interviews, and transcripts of panels and roundtables.

Excerpts from blog, newsgroup, and forum posts.

The "Laws of Online World Design" in various forms.

A timeline of developments in online worlds.

A Theory of Fun for Game Design
My book on why games matter and what fun is.

Insubstantial Pageants
A book I started and never finished outlining the basics of online world design.

Links to resources on online world design.

Another post originally from the Star Wars Galaxies boards that got a lot of reprint attention on various message boards. The first part is the post from those boards. The subsequent snippets all come from MUD-Dev discussion of the topic.

On Socialization and Convenience

We've been having an interesting discussion here at work over the last couple of days. We've been working on the layouts for various buildings found in cities, and in the process found an interesting philosophical question I'd like to share with you all.

To start with, let me ask you this question. It's a very arbitrary question, and your answer will reveal a fair amount about you as a player. It's an unfair question, of course, but ignore that, and just go with your gut.

How much time do you think the average player should spend socializing in SWG? Meaning, as opposed to "playing" however you define that--killing things, crafting, whatever. Chatting while recovering from a fight counts; chatting while forming a group counts too.

Got a number?

So numbers we arrived at among our team ranged from 3% all the way up to 50%. I think it mostly reveals things about how different people play the game--and also about how people define socializing--and also about their memories (I flat out don't *believe* the 3% people--that's a total of a few seconds every hour spent chatting with people, on average. My take--nuh-uh, no way. :) For the record, I was a 50% guy).

Why do I ask this? Because we have contradictory goals for the game. We want to reduce downtime. But people get to know people during downtime. That's when they socialize. That's when they make friends. In fact, I'd go so far as to state that it is a Law of Online World Design: Socialization Requires Downtime. The less downtime, the less social your game will be. And we ran headlong into this while discussing interfaces for common municipal structures.

Let's take a bank as an example. The question came up as to how you would use a bank. It matters because of how we lay it out. If you have to walk inside and use a computer terminal, then we need wide doors and spacious interiors and lots of terminals. But we could also make it so that you could use it anywhere inside the structure--we'd get rid of the terminals, change the layout somewhat based on the flow. Probably have many doors in, since people would tend to stop at the doorway, which is the first place they can do their transaction, and then turn around and leave.

Then we said to ourselves, "Wait a minute. We have a credit economy. We could make it so that you used the bank from anywhere via your datapad." First we talked about a radius around the bank--like say, in the courtyard outside. Then we started talking bigger radii. Finally we we said, "You know, you could just use the bank from anywhere in the city!"

And we said, "Wow, that's awfully convenient! Saves tons of time! We could do that for pretty much every municipal structure!"

But there were some nagging concerns. And it helps to think about the purpose of different types of structures.

In architectural theory (cf Timeless Way of Building or A Pattern Language) there's a lot of well-established thought about traffic patterns and the ways in which they affect the well-being of a community and the ways in which they affect the culture of a community.

Let's take the example of community building. There's an oft-told anecdote (the precise source of which escapes me atm) about a company which was suffering from malaise because people weren't coming up with good new ideas to advance the business, and there was stagnation and loss of morale. When the office building the company was located in was reorganized such that there was a central courtyard type space that served as a crossroads, and the different departments were obliged to walk through the courtyard on a regular basis to do their regular work, morale boomed, so did ideas, and so did profits. Why? Because the fact that people were interacting with people (and therefore ideas) that they normally didn't sparked both creativity and community.

The same logic can be applied to other types of desired results; if you seek convenience, then it makes a lot of sense to *avoid* crossroads. Roundabouts actually improve traffic flow precisely because you don't have to encounter other people head on. The reduced speed but lack of a total stop and waiting for turns to move forward means that vehicles continue moving at a steady pace, and there's actually a reduction in pedestrian accidents (good article recently in Discover magazine about this effect).

A second key philosophical question--I asked the team, once we'd argued these points for a couple of hours, what their preferred metaphor was for a town in the game--player-run or not, really, though we focused mostly on player-run. Many different answers came up--what sort of organization or community do you see that feature of the game as being most like?

Ponder that one for a bit.

Got an answer?

OK, so one of the most frequent answers we got was "guild," meaning people saw it as an alternate form of player association. We also got "staging area" a lot, meaning people saw it as a launchpad to the "real game." Some saw the metaphor as being "shopping mall" or "apartment complex." My answer was kind of long and poetic, and people kind of looked at me strangely as I rambled. It went something like, "A 1950's small town with a local hardware store on the corner where the shop owner knows what sort of paint you really need for your fence and an ice cream parlor where you can go to get root beer floats and sarsaparilla and a bar where everyone knows your name and where the people you see at the local grocery store are mostly people you know by sight if not by name and there's a gazebo in the town square where sometimes they play live music..."

So we stopped for a moment and thought about what sorts of downtime we were removing, and what sorts we were enforcing, and what types of community building mechanics we were putting into the game. We used examples from other games to think about the sorts of activities and locations that we saw as drawing crowds and leading to community ties.

Here's some examples from other games:

Most of us on the team had fond memories of blacksmithies in UO. You went there because you were looking for a player to repair your weapons and armor, and you needed a skilled player to do it, or the items might be ruined. They were there predictably because they needed a forge to do it. The result was a pleasant experience chatting with the blacksmith, with others waiting their turn, and a great launching pad for meeting folks and going on adventures.

On the other hand, most of us disliked the UO banks. They were the default place where there was a crowd, but everyone who was there was either there absorbed in an interface screen (i.e. not talking) or was spamming the crowd with items for sale merely because there were lots of people there. It was not a sociable place, though it was a social place. And of course, everyone stopped there at the start and end of every adventure.

I have many fond memories of hanging out at town fountains in Diku muds. Usually they were set in a town square, and the structures to the sides of the square were key to gameplay. The newbie hall, where everyone first entered the game, opened down onto this square. The inn, where everyone came to log out, and from which everyone logged in, was on one side. The guildhall where you had to come to advance a level was there. As a result, there was always a knot of people swapping stories about where they had just been, and making plans about where to go next. It's hard to imagine a more welcoming environment for a newbie to step into.

Spawn points in EverQuest are of course a much maligned source of downtime. But many people attested to the idea that that was where they chatted and talked. But the fact that the downtime was a barrier to further gameplay in their eyes (meaning, they were camping so they could get some piece of armor or a weapon that they saw as necessary to continued enjoyment of the game) led to resentment of the enforced downtime and appears to have harmed its value as a social space.

Whereas safe zones in EQ were seen as staging areas. These are places of lower risk in the midst of dangerous areas. As natural gathering places, these locations became places where you bumped into people with common interests (killing whatever was nearby) and of comparable skill (since they were likely to be in your level range). It was where people retreated to to rest up and heal, and it was where they started a big foray from. A base, so to speak.

With these examples in hand, we classified the types of social spaces into three:

Staging: these are places where you form a group, find a friend, and decide what to do.

Pit stop: these are the obligatory stops you make before you get to have fun.

Recovery: these are the places where you go after an adventure.

Here's a third touchstone question that emerged. Recovery areas--what are they for? Think hard.

Got an answer? OK.

We were divided on this one too. Many of us saw them as obligatory character maintenance--the place to go when you need to heal up. We also saw them as rites of passage--the place to go to level up, learn your skill now that you have the achievements, whatever. But the third big thing we identified, and the thing that some of us felt was the most important thing, was that they were opportunities for mythologizing. The chance to retell the story of our adventures to ourselves, the chance to establish a consensus history, relive the incidents, and weave a narrative out of what were in fact very disjointed moments with no storyline or structure to them.

As the fates would have it, we had a fantastic example of this at lunch today. We went out to Fuddrucker's for burgers, and after we had eaten (think of that as the obligatory pitstop!), some of the guys who had most strongly seen the recovery areas as being about character maintenance started talking about the previous days' game of Counterstrike. "...and then I whipped around the corner and the machine gun..." "Yeah! And the idiot kept going and..." "Yeah, it was great! And then he did the thing!" "Yeah, the thing! That rocked!"

They made no sense. :)

In light of this breakdown, it's easy to see that recovery areas are GOOD sorts of downtime. That's why the safe areas in EQ and the town fountains on Dikus work well. And it's also easy to see that pitstops kind of suck; people see them as barriers to getting on with the fun, like camping spawns or having to visit the bank before and after every adventure.

An interesting case was the blacksmithy in UO. Clearly a pit stop. But since it involved a player service, there was a human element to it that was missing from the bank or the spawn point. Waiting for another player is more palatable than waiting for the server to do something for you. So pit stops don't have to all be bad.

Lastly, staging areas seem plainly vital, because you need to have places where you form your party or group.

But here's the rub. We had eliminated almost all of this stuff in the name of convenience. You don't need to visit the bank in SWG if you just enter town and transfer credits. To pick up gear you go by your house, which means your group scatters to the four winds before setting off. You don't visit the blacksmithy to get your weapon repaired--you drop it in a hopper and pick it up later. You get a mission on your datapad. You don't need to go to the town square to get your mail, you do that on your datapad too. In fact, the more we talked about it, the more we posited that if there were key structures (like needing to visit a shop to pick up your fixed blaster) they'd be placed on the edge of cities, not in the center, so that you could "bounce" off of town as quickly as possible.

Even our recovery areas suffer from this. Yes, we pretty much make you go to cantinas and taverns, because you need to heal wounds. But that means that the only people whom you will meet in cantinas are wounded people and healers. And maybe a bartender. That leaves out a lot of types of player--the politicians, the crafters, the farmers and the animal trainers.

And that brings us to a fourth touchstone question. Do you think you will play mostly with friends you make before the game, or friends you make in the game, or with strangers?

Empirically, we know that friends made in the game are retention devices. Frankly, we want you to play with people you meet IN the game. That's because otherwise, all we have is a bunch of cliques. Hermetic organizations made up of people who mostly knew each other in advance somehow (maybe they organized their towns on a web board, like so many of the SWG players are doing now). And no easy way for a novice to the environment to make new friends. The fact that the decisions we had made meant that people would not tend to bump into strangers reveals a flaw in our thinking about managing community and downtime.

Online games have the opportunity to offer microcommunities, tight-knit groups of people working towards common causes. This is something that most of us miss in our daily lives, and it's something that is very woven into human nature and life, and has been for millenia. We speak of the dehumanizing pace of life in the cities, and the ways in which we tune out people in crowds. That's why I can speak so nostalgically of the small town experience. A large part of the attraction of online games is, to my mind, recapturing that sense of community. If we make life online overly convenient, what we may end up doing is merely recreating the experience of being a newbie in New York City.

But I could be wrong. And that's why I pose the question to you now again, after you've read this very long rambling post.

How much time do you think you should spend socializing? And where? When does convenience become dehumanization? And fundamentally, just how much downtime are you willing to take? Because it's evident that some needs to be there.

I look forward to seeing the discussion. :)

This seems to presume that you're making a monster-bashing game (which I know you are, but I got the idea you were making a broader statement). I'd add that sometimes a single social space can be all three of the above AND the place where you have your 'fun'. For instance, the politicians in Achaea might be sitting in a single room, doing everything from there while chatting. - Matt Mihaly

I am making a broader statement; there's staging areas, pit stops, and recovery phases in politics too. :) The monster-bashing just happens to have a lot of this stuff quantified so it makes an easier example.

In terms of traffic flow, it's of course necessary to talk about it in terms of locations. But your example is not location-based, it's time-based. The area changes in function as different stages of the gameplay/activity occur.

The problem with not paying attention to traffic flow is that in your above example, it is extremely likely that the politicians will only talk to other politicians, unless there's some traffic flow reason for non-politicians to come through there. And I WANT strangers to rub up against one another, because it encourages that more open 250-person community (hermetic cliques tend to be much smaller), it increases the odds of welcoming novices to the environment, and it therefore acts as a retention mechanism.

The disconnect I'm having here is that for socializers, socializing *is* playing. - Jeff Freeman

Oh, yes, of course. But they are also not incurring any game downtime in the process. To get more specific:

You can look at a game as being composed of activities tied to reinforcement mechanisms. An action takes place, and the game/opponent provides a reaction; the reaction had better be broadly predictable, or else the player will consider it gambling, not skill, and hence not a game in that sense.

In most games, the degree of activity tends to be bursty; in physical games the bursts are the periods of actual physical activity, followed by downtime for rest and recovery. In mental games, the game is either designed to be relatively brief (usually by overwhelming you) or the pace is player-directed.

In online games, you have a combination of all of these things. Take the basic hack n slash advancement paradigm as an example, since we're all pretty familiar with it (most of the list because they make them, and those diehard holdouts who don't who keep hoping we'll talk about non-GoP muds sometime, well, they've gotten amply familiar with it by now, poor guys!).

In a level rat race design, you have a player-directed pace, and also an enforced activity schedule determined by the need for stats to regenerate. It's loosely enforced, since there's usually ways to accelerate the process. While within the encounter itself, it's designed to be relatively brief; it too is also designed to overwhelm you (or for you to overwhelm the opponent). A given combat is not planned to last days like a chess match, but rather to be self-terminating in a much shorter period of time.

So I am defining "downtime" as the periods when a player is on that cycle and is not in the combat. More broadly, I'm defining downtime as "time during a game session wherein a player is not actively participating in game mechanics because said game mechanics are not viable for that character at that time."

Socializers are excluded by this definition; they are not waiting for something, so for them it isn't downtime. It's free time, which is a different beast. They have intentionally stepped off the treadmill and have all the time in the world. That's a radically different mindset.

Given that mechanically speaking, you're probably not typing while you are hitting keys to get things to happen (such as striking blows) I would guess that a lot then depends on the granularity of your perception of downtime. If you're walking from one close spawn to another, then that's a form of downtime. A zero-downtime game would be like an arcade shoot-em-up game (and even they pause when you clear a wave). The fact that you are describing downtime as a time when you read a book suggests a much longer period of time.

"I'd go so far as to state that it is a Law of Online World Design: Socialization Requires Downtime."

Argh. Resist the tempation. It's not right! - Jeff Freeman

Well, I haven't heard a good argument against it yet. :)

In a GoP environment, players who are engaging in the activity are going to want to devote their full mental faculties to the activity itself (in fact, that's how they get those endorphin highs, entering quasi-meditative states while engaging in the activity). They're not going to chat during that time; if they don, they won't be all that successful at the activity unless they are expert enough players that they can time-slice the activity so tightly that the space between one blow and another is perceived as downtime by them. (And this is pretty common, and a lot depends on the pace and nature of the activity).

Now, if we posit that meaningful social interaction requires at least a question or statement and a reply (an in most cases, will require more than that; a mere two-sentence interaction that actually reveals something about the other's personality either presupposes prior knowledge between the two, or is a dead-end) then we see that in order for there to be enough room for social interaction, the activity must have a fairly slow pace of input (and situation assessment) and both players must have some comparable level of timeslicing that so that their respective slower-paced moments coincide.

In my head I am seeing this as a graph; places where the two troughs coincide are where players might talk. The trough has to be long enough that the player's attention can be easily divided from what they percieve as the principal activity.

And actually, using the above definitions (which are brand new, I just came up with them as a typed, I am thinking out loud here) we can see why the bank sucks. Despite the fact that player congregate at the bank, the bank is actually a peak on the graph, not a trough. It's a moment when players' attention is heavily taken up by a game mechanic necessary to their goal. They cannot effectively timeslice the activity of dropping stuff in the bank and taking stuff out, and many of the decisions they make are in fact fairly high-pressure decisions that can affect the success of their characters in the future. They are self-directed in time impact, sure, but players are not likely to want to be distracted while making them.

The sole exception I can think of--if they are there with someone who is already trusted, in which case they may socialize a little, and ask for advice on the decisions. But this is not a case likely to get strangers to talk to one another meaningfully. Unless they have crossed some threshold of trust on the player's part, their advice isn't useful.

The blacksmithy example involves very few decisions. It's a required gameplay activity, but the decision is "go there" which is a high-priority decision but comes before the standing around while there, and a low-priority decision "leave early" case which burbles along in the back of the mind but doesn't have much tug to it. (I'm coming up with a different metaphor in the back of my mind now, which is a weighted AI decision; the "keep getting cheese" impulse has a far lower priority than the "fix my paw" impulse; it therefore gets pushed to the bottom of the stack until "fix my paw" is completed. Since the "fix my paw" has an enforced delay on it, impulses which are normally much lower than "keep getting cheese" have an opportunity to appear, such as the "sniff other mice's butts" impulse. Hi there, Mr, Maslow, how are you today?).

The lesson to learn from this is that the people were required to come into contact with one another due to the new building layout. They were NOT required to stand around in the foyer for hours a day doing nothing. - Jeff Freeman

I see these as two facets of the same problem.

For the layout described to work to that purpose, it's evident that some of the tasks the people had must have been lower priority than "sniff other mice's butt" or else there would have been zero interaction even in the courtyard area. Or perhaps some of the tasks were higher priority but had that sort of enforced downtime (I can easily see this layout working if every Xerox machine in the entire company were located in the counrtyard, for example).

If not, what you get is just a traffic nightmare as everyone rushes through the courtyard as quickly as possible.

HUGE side note: it occurs to me that if you know the Bartle breakdown of your playerbase by broad percentages, and you can write some basic ALife code for weighted goals and then weight additionally for each entity based on their Bartle's type, you could probably throw this at your game's dataset and get a nice overhead view of your game with dots running around that shows you exactly what zones won't get used and why.

I believe that socialization can be summed up as introduction and then companionship. - Paul Sage

This ignores the factor of where the introduction point occurs. Either it occurs in a place, or it doesn't. If it occurs without a place, then it's occurring on a global chat channel or via some sort of matchmaking service. If it occurs in a place, then we have to worry about why these different people happened to be in the same place. What's more, int he healer scenario, the healer by definition has to be in downtime, or else he's too busy to heal the other player; in the fight needed by merchant case, the fighter has to be between adventures or else he's off slaying a dragon, not helping a merchant.

Fundamentally, if the other guy's busy, then he's not coming to the introduction party. If he isn't busy, then he's gotta be parked somewhere, or on the way to somewhere. And that suggests we need to manage traffic flow. And it also suggests we need to manage downtime, or else everyone will always be busy--especially achiever types who by nature seem to be driven by always having another rung ont he ladder to reach.

On LegendMUD (a fairly GoP environment, fundamentally) we added a socialization area with a bunch of nifty social facilities. The Wild Boar Tavern offers a lounge for chatting, goofy food to buy, an auditorium, a gift shop to buy goofy items like birthday cards, a wedding shop for in-game events, etc. You can reach it instantly from anywhere by merely typing "OOC." It was there in an instant for anyone who wanted it.

It doesn't get used.

On UO we had taverns with NPCs, dart boards, chess boards, backgammon, dice. There were multiple ones in every town. You know as well as I how crowded they were.

Leisure time in a mud is pointless time in players' eyes, and only a small subset of your players will be looking to spend pointless time.

John Buehler made the excellent point that keyboard and mouse input is currently a huge barrier to socialization.
A very important point. A single point of input; adding real-time voice provides two input channels. In addition, voice in real life is a much faster input channel than typing is, so the addition of voice allows you to socialize in smaller timeslices.

...People use voice constantly in shooter games such as Counterstrike. But they don't socialize with it because the game itself is very intense and offers a lot of sustained, high pressure decision-making. Instead, it is seen as a better way to accomplish the game tasks.

As far as I'm concerned, there's no need for architected downtime periods in order to encourage socialization. - John Buehler

Obviously, I still disagree. What I will agree with is that the downtimes can be made much smaller if you support a faster means of communication; and I'd say that this is a universal principle. You can encourage more communication in any game by either a) slowing the pace of the game or b) speeding up the communication facilities. It's a simple mathematical equation. You can either make the window bigger or the throughput greater.

One of my teammates (who is on this list somewhere lurking) is a pretty goal-oriented player who has recently gotten addicted to Sojourn 3. He cited a very low percentage too, when asked the question. He decided to run an experiment with a pair of timers whilst playing a typical session. He was a group leader and he was hack n slashing his way through the mud with abandon. He didn't count stuff like "OK, go west then headbutt" as social interaction.

Turned out that his actual time spent socializing was 25-30% even though he thought it was 10%.

Sun, 29 Jul 2001

I inherently distrust the concept of virtual apartments, rather than actual houses on the land, because they are less conducive to neighborhoods, in my mind. Less odds of bumping into a neighbor, less odds of seeing a house and getting curious about who lives there, far less ability to express yourself to a passerby and thereby catch their attention.

...A long time ago on this thread, there was an observation (perhaps by JCL?) about how the fact that UO was overcrowded was what led to many of its key community-building effects. I believe the term used was "hothouse." Also related is the whole socialization and convenience thread from not very long ago. If you add more and more lubricant to ensure that a player gets a good experience without needing to rub up against anyone else, compete for resources, interact for things that they lack, then you are increasing convenience and removing reasons and opportunities for socialization.

I really ought to go ahead and elaborate the preceding thread into a more detailed theory, but oh well. In the end, "Socialization requires downtime" made it into the Laws. :)

Child's Play

A Theory of Fun
for Game Design

Cover of A Theory of Fun



Buy from Amazon

After the Flood

Cover for After the Flood CD

Available on CD

More stuff to buy

Gratuitous Penguin 2006 Wall Calendar

Gratuitous Penguin 2006 Wall Calendar

Receive CafePress Updates!


click here to visit the Legend website

"The world the way they thought it was..."