Raph Koster's Home Page

This is a series of posts to the SWG message boards describing our design process for SWG. I have posted it here because I know there's a lot of folks who are probably curious about things like this.

Star Wars Galaxies Design Process

Well, as expected, the majority of the posters (who answered the thread, anyway) were among the hardcore. That's what I figured we'd see; since being on this board in the first place is a self-selection mechanism--it's the hardcore who tend to hang out on game boards anyway!

You know, everyone wants different things out of a game, especially a massively multiplayer game. At least in a single-player game, there's a prescribed narrative, usually--there's a way to win. There's a goal you try to reach, and specific, easy-to-spot milestones along the way. We strive to put some of these things in MMOs as well, but there's something about saying "here's an online world" that makes people not only crave the open-ended ness, but also try to impose their particular playstyle upon it. Players feel very let down if the game doesn't support THEIR mode of play. We've seen that already here in some of the threads, with some people saying, "I won't play unless there's PvP" or "I won't play unless I can run a shop" or "I won't play unless it has action-style combat."

This isn't a list of expectations that would normally be applied to a standalone game. It gets applied to massively multiplayer because people's expectations of virtual worlds are higher than their expectations for single-player games--or even higher than their expectations for real life, on occasion!

But there's also a powerful counterargument to be made--that without encompassing a wide range of playstyles, a virtual world can't be successful. People need to mingle with people who do different things from them, argues this theory, because otherwise they won't be happy in the long run.

Probably the best articulation of this theory specific to online gaming is Dr. Richard Bartle's division of players into four types. Richard wrote a paper called "Players Who Suit Muds" using his accumulated insight as someone who has run and developed muds since 1978 (Richard is generally credited as the co-author of the original mud, which means the original massively multiplayer virtual world, albeit text-based).

The basic premise of his paper is that there are some people who are there because of the world, and some people who are there because of the other people. Based on this division, some of you are here because it's Star Wars; and some of you are here because it's an MMO. As a second variable, some of you are the kinds of people who enjoy taking action and making significant choices and knowing that you have an impact; and some of you are the sorts of people who value interaction, and like fiddling with things, building, exploring, tweaking, and trying new stuff out.

Based on this, he ended up with four types:

ACT on the WORLD: Achievers, who are out to beat the game. They tend to be powergamers, they want to master the mechanics that the designers have set up for them.

ACT on the PEOPLE: Killers, who are out to establish dominance. Not necessarily PvPers, in my reading, but rather people who like to be in charge, like to have control--leaders and lawyers as much as playerkillers.

INTERACT with the WORLD: Explorers, who want to master every detail, not for character advancement, but "for the sake of science." These are also people who enjoy crafting and creating.

INTERACT with the PEOPLE: Socializers, who are there primarily because their friends are. People who build social structures, who spend most of their time chatting or roleplaying.

Bartle suggests that every successful online world has a balanced proportion of each of these core personality types (and it's interesting comparing these to the Kiersey temperament test btw).

Everyone is a mix of these traits, of course. But games can most definitely cater more strongly towards one or another. For example, many say that EverQuest is primarily an Achiever's game, and that Ultima Online attracts more Killers. There are "Bartle's quotient tests" out there which hundreds of players of these games have taken that demonstrate that this is actually true (though the percentages according to the results differ by fairly little).

What sort of gamer are you? For what it's worth, according to the tests, I am very high in Explorer quotient--most developers who have taken the test tend to score high on that, probably indicating that it's a trait that leads them to eventually try to make the games they like playing. :)

MMO Design and Satisfying a Varied Player-Base (posted December 13, 2000)

Good morning everyone... sorry for not posting yesterday, but as you may have seen on the news, we've had a spot of bad weather here in Texas (and probably also where many of you live). At the moment my house is without power...

Since in the last couple of notes I talked a bit about the audience we expect to get, and what sorts of different gameplay people in that audience want, I thought that today I could maybe talk a little bit about how we go about designing an MMO, and how we try to satisfy that range of people.

To start with, I happen to believe that the four play styles I mentioned should all be catered to. We play massively multiplayer games in large part because of the other people. And if the game's population consists entirely of other people just like us, well, it's not going to be a very interesting world. Other people like us won't challenge us in any way, won't expose us to new ideas, won't surprise us with actions that we ourselves wouldn't have taken. And that would make a massively multiplayer environment a lot poorer, because we rely on the other players to provide unpredictability and interest.

Yes, there's plenty of room for niche products that cater to one play style or another. But if you want to try to convey a universe (especially one as deep and rich as Star Wars) you have to have all sorts of people. On top of that, all sorts of people love Star Wars, and they're all going to try the game. I'd hate to disappoint everyone except one particular play style.

So how do we go about trying to ensure that the game offers enough for each of the types? Well, it all comes down to the design, like so many other things. Our first step was to create what we call a vision document. This started out by identifying some things about who we thought we were making the game for. Right off the bat, we established what sort of audience we thought we were looking at, and that led to establishing a whole bunch of other things. As we came to each statement, we put it into the vision doc.

For example, knowing that we were looking at a broader audience than MMOs had likely seen before meant that we couldn't demand as much time per play session or as much time per week as other MMOs did. As a result, a bunch of design choices went right out the window: we knew that we couldn't have game design elements that involved spending tons of time online. No macroing, no camping, no lengthy corpse recoveries, no long waits for public transportation.

This also led inescapably to the logic that a player who had less time to spend per week couldn't be put at a severe disadvantage. Now, obviously, someone who has more time to spend on the game is going to accomplish more. But we CAN do things so that people don't feel left behind. So into the vision doc went a statement like, "players who play a lot and players who don't have as much time to spend should still be able to play together fruitfully."

We also had more concrete stuff in there. There are things we have to put in to attract people to the game, like great visuals. There's things we need to have in order to make money, like having enough depth in the game so that players continue to subscribe for months. We ended up with six pages of these one-liners. Some samples:

It should have an intuitive user interface. If the choice for initial interface is between flexibility and ease of use, we will choose ease of use, and make the flexibility an option.
It should be possible to play a satisfying game session in less than an hour.
It must be easy to extend and add content to the game after ship.
Star Wars Galaxies must provide multiple ladders of advancement so that new ways to play the game become apparent over time.
The game system will encourage specialization.
Players should never feel ripped off by the game mechanics.
We must encourage a sense of player ownership in the world.

Tomorrow I'll talk about how we translated that into game features. Stay tuned�

Design Process, Part Two--What is the game ABOUT? (posted December 15, 2000)

Hello everyone. I'm suffering from a nasty cold that made me miss a chunk of a day of work, but I still made it here to the boards to post my promised next installment of how we tackled the design.

I mentioned yesterday that we had built around six pages of "vision statements" that reflected what we felt the game had to accomplish, and I posted some examples. Now, I left something out there: the context in which these visions had to be reached. The key element there, of course, is the fact that this game was set in the Star Wars universe. And that requires some vision elements of its own! :)

I basically took the design team back to high school English class one day, and had everyone sit around and think about "What is Star Wars about? Thematically, and emotionally?" Everyone by now knows the whole deal with Joseph Campbell and the Hero's Journey, so we breezed by that bit. One big reason for breezing by it is that the Hero's Journey is not very compatible with a massively multiplayer mindset. Not everyone can be the predestined hero that will save the world (or galaxy, in this case).

After discarding the Hero's Journey, we ended up with a short list of key thematic elements that we thought every Star Wars fan would have picked up on either consciously or unconsciously. We knew that if we failed to represent these things in the game, then players would (correctly) feel that the game wasn't truly Star Wars. Just as you can't imagine The Sims without consumerism, just as you can't imagine Ultimas without ethical dilemmas, you can't imagine Star Wars without these things. These, then, were the things that the game mechanics had to be ABOUT.

Exotic worlds and epic adventure.
This meant that we had to strive to reduce tedium and repetition from the game mechanics wherever possible. We had to strive for as dynamic an environment as possible, while still faithfully reproducing the SW settings players would be familiar with. The game had to be filled with discovery and surprise.
The struggle between the light and dark sides of the Force.
This meant that not only did we have to have this in the game, but that we had to allow players to feel like they had a meaningful impact on the struggle. Through their choices, they had to be able to make a difference. At the same time, people needed to be able to opt out.
Political battles to take control of the galaxy.
We don't tend to think of this one often, even though it's a clear backdrop to everything (expressed as warfare in the Classic trilogy, but more directly seen as political maneuvering in Episode I). There are plenty of people who are fighting in the Civil War not because of the Force, but because of more mundane political reasons. Again, we had to allow players to feel like they had a meaningful impact on political struggles in the game through their choices.
Interactions with other players.
This one is more driven by the needs of an MMO design than by the specifics of the Star Wars setting, but even in SW we repeatedly see the theme that people need each other. In an MMO this becomes even more critical, and we knew that this had to be one of the main thrusts of the game mechanics. Interestingly, this isn't something that MMOs thus far have done a very good job of rewarding with in-game mechanics. Sure, games have required you to group, and that sort of thing, but the game mechanics haven't explicitly rewarded all the people who make the game world a more livable place in other ways: the ones who are good roleplayers, the ones who provide entertainment, the ones who are in support roles or peaceful roles or economic roles. We knew that we had to make our game mechanics reflect this.

With these items, we had what the game was going to be about. And with that, it was easy to see what sorts of features we would need to have in the game, right? I think I'll be mean, stop now, and leave it as an exercise for the reader. :)

Balance and More: Making Economics, High Level Content, and Advancement Work (posted December 20, 2000)

Whoo, this is gonna be a long one, I can tell. :)

I saw this topic & posting when it first appeared on the Lum the Mad message boards, but it's been reposted here at [link lost, unfortunately].

It's full of excellent questions, many of which I can't answer in detail because we're not ready to talk about how we intend to do certain things yet. In what follows, I've simply taken the principal issues raised and explained roughly what our approach is to solving them:

Balance

Yes, balance is hard. One thing we are taking as an approach is to be very very obsessive-compulsive about it. Here's some of the specific methods that we try to use:

Step one, use your math. I am a big fan of systems that balance mathematically, because they make good starting points for REAL balance. Just one example is the magic box: if you have twelve stats and fourteen attacks, and you want each attak to vary per stat, or vice versa, you make a grid, and ensure that every stat comes out at a total of zero modifiers across all fourteen attacks, and that every attack comes out at zero total modifiers across every stat. This then serves as a starting point of theoretical equivalence between every stat vs every attack, across all players. Now you can tweak from this solid base to account for much fudgier factors like "how expensive is attack 12 versus attack 9" and "how hard is it to get a weapon that does attack 3" and "how useful is stat 11 in other, non-combat settings anyway?"

There are other mathematical approaches to getting things balanced, and we try to use all of them that we can. You will notice later on just how much we will tend to use the number 3 in our game system, and that's because it lets us set up nice rock-paper-scissors relationships to ensure that in any given set of figures, everything has at least one vulnerability...

Step two, use inheritance. A lot. This means that as you create your database of items, you don't create each item in a vacuum. You base every item on a pre-existing item. That first item is ideally a generic "level 1 blaster" "level 2 blaster" etc. That way, if you determine that all vibroblades are weak, you can modify just the base vibroblade and get consistent results everywhere in the database, you can adjust all critters of a given level of difficulty, etc etc.

Step three, log everything. And I mean, everything. Track who kills what and how with how many people where. Track who spent what to buy what where how often. Etc etc. And analyze the results!

Step four, rely on self-balancing mechanisms. This is the default anyway. When an MMO starts to go out of balance, you see players pick up compensatory mechanisms. For example, when currencies in an MMO deflate, players start to shift to barter, and eventually designate a new item as an alternate currency. You can make use of players a lot with this. It's particularly applicable to economic issues, and I'll explain how when I get down to the econ answer. :)

Making the game enjoyable for casul gamers/those with less time

We've actually talked a fair amount on this one already. Our principal approaches are these:

reduce the amount of time you need to have fun in a session. This is because casual gamers don't have as much time to play in six hour blocks as hardcore gamers do. ;) So we strive to reduce or eliminate the time involved in doing traditional time sinks like corpse recovery, macroing, forming a group, and getting to where the fun is, so that the amount of time a session will take is more predictable.
provide a wider range of activities. Fact is that the larger, more casual audience wants to do things other than just kill. And we intend to provide ample opportunity for that, and what's more, we want to explicitly reward it within the game. Traditionally, those sorts of activities don't fit into the experience-based standard RPGing systems, and that's why we're not using a system like that.
don't make the game centered around power differentials. This is what makes harder core players outpace more casual players, what makes it so that a high level player can't group with a low level player, etc. Both of those effects are things we want to avoid, so we're going to the root of it, and ensuring that our game system isn't based on massive differences in power. Instead, advancement is skill-tree based, which means you acquire abilities, not just bigger numbers.

Dynamic content

Yes, it's good, yes, we're doing it as much as possible, and a lot more than you've seen in other MMOs to date. A lot of it will be player-driven dynamic content.

Advancement

Multiple paths of advancement, and what's more, paths of advancement that use different advancement engines, are a key factor in several of the things I described above. Again, this is why we don't use XP-XP is not well suited to advancement in the context of roleplay-based, economically-based, exploration-based, politics-based, or any other non-killing-based advancement ladders.

Support

This is a problem the industry as a whole is still wrestling with. We intend to keep improving, and to do the best we can. We are investing a significant amount of time in creating good support tools and backend infrastructure so that we can help you more and more effectively.

Lack of high level design and content

This is actually a double-edged sword. The way I like to think of it is this:

too often, we as developers spend too much time adding new content only at the top of our advancement ladders. This makes the games unfriendly to newbies and reduces the overall difficulty of the game as higher-level items and whatnot trickle down to lower-level players.
At the same time, we usually don't invest enough in what I call "elder games." These are the systems that give you something to do, new objectives, new ways of playing the game, after you have "maxed out."

Our entire game system is built around the premise that every different profession can serve as an elder game for someone who has maxed out whatever they started out in. (Hmm, am I revealing too much about how it can work? Probably�) I am a huge believer in elder games, and what's more, I believe they should change the HOW of how you play the game-sort of like how advancing through the military means you have to learn completely new skill sets, not just "be a better private."

Repetition

This is a tough one. Some people like it. In fact, a lot of people like it. What we can do is provide diverse enough activities, and enough change in the world, that you can always find something different to do. But it IS natural to run out, eventually, and there's nothing much we can do about that.

Game economy

The biggest thing here is that we intend the economy to be player-driven as much as possible. This is because the single biggest factor screwing up in-game economies in these games is us, the developers. We put in hard-coded value numbers, we write unsophisticated shopkeeper AI, we try our hand at macroeconomics in spawning & whatnot, etc etc, and as a result we create subsidies and taxes and hidden costs and all sorts of problems. To quote Zach Simpson's excellent paper on "The In-Game Economics of Ultima Online" (you can find it on the web), "it's not a virtual economy. It's a real economy with virtual objects." We are mucking things up when we try to "simulate" it, when the best thing to do is get out of the way as much as possible-we're not economists and we're not expert at writing econ sims. For that matter, economists aren't either. ;)

Whew. That's the top issues, I think. Hopefully it's enough for you all to chew on and speculate with!

The last part is a reply to an essay that was circulating around the online gaming community at the time. But I can't remember who wrote it or what it was called, and so the reference is lost.
We've ended up going with XP anyway because they are easier to quantify and track than a zillion individual events. But we retained the fact that you can earn those XP in a myriad of ways.