Raphael Koster is a prolific theorist in the young field of massively multiplayer online game design. For many years, he has developed his design philosophy on his blog, in lectures, at conferences and in his 2004 book, A Theory of Fun for Game Design.
Unlike some high-profile thinkers, Raph Koster actually ships product. “I do all this writing to clarify things for myself,” he says. “I put it out there afterwards, figuring maybe it’ll help other folks, but the initial drive comes … because I am banging my head against a design problem. So, the theory is a tool. You write it down so you don’t forget it – it’s like having a toolbox full of screwdrivers, wrenches and whatever.”
Koster started designing for MUDs in 1992, while working on a Master of Fine Arts in Poetry from the University of Alabama, Tuscaloosa; he was “implementor” “Ptah” in LegendMUD. (Check out an in-game interview with him here.) Joining Origin in Austin, Texas, in 1995 as a designer, Koster worked as creative lead on Ultima Online and, after launch, was lead designer on the Live Team until 1999. He joined the Austin office of Sony Online Entertainment, where he was creative director for Star Wars Galaxies: An Empire Divided. Shortly after SWG’s launch, Koster moved to San Diego to become SOE’s Chief Creative Director.
Though Koster left Sony in March 2006, his contract expired in June. In this interview, conducted by e-mail soon afterward, he talks about past and current projects, and about what went right – and wrong – in UO and SWG.
In A Theory of Fun for Game Design (you can find some excerpts here), Koster defines “fun” as a function of learning and mastery. As we explore a new game, we learn to recognize its challenges and exploit the tools offered to overcome them – that is, to gain mastery over the game environment. “Fun” (which Koster distinguishes from aesthetic appreciation, visceral responses, delight and other forms of enjoyment) is “the act of mastering a problem mentally” – the endorphin reward feedback the brain gives us when we are absorbing patterns for learning purposes.
“Overall, I was thrilled at the reception the book got. I’m really honored and pleased it seems to have become a useful part of the overall landscape of thinking about games. It was interesting seeing the criticism, too. Many academics wanted much more detail in the book, for example. It’s a mode of writing I am not all that interested in anymore, not since graduate school, so I felt few qualms about shrugging and moving on. Lots of folks said the ideas in the book were too obvious – and certainly lots of folks, like Chris Crawford, had said large pieces of the book before.
“By and large, the general ideas have held up for me. I think the most interesting comments on the book’s ideas have come from the folks who say it did a good job of pinning down one sort of fun – a few people have taken to calling it kfun. I think there’s something there. As [player experience psychologist] Nicole Lazzaro keeps doing research into emotion and games, my theorizing may get replaced by concrete data soon, which would be incredibly helpful to game designers.”
Koster’s current book project is A Grammar of Gameplay, an ambitious attempt to symbolically describe the component “atoms” of games. He presented an early example at the Game Developers Conference in San Jose, CA, March 2005. The grammar would be a tool to reverse-engineer and notate individual game ingredients, such as topology (“the operational space for a given asset”), core mechanics (“ludemes”), depth of recursion, cost of failure and many other abstractions. Using the grammar, a designer could quantifiably assess a game’s difficulty, range of challenges and required feedback mechanisms.
The first large-scale success among graphic MMOGs, Ultima Online peaked at 250,000 subscribers in July 2003 before gradually declining to its current 135,000 (of whom half are in Japan).
“I really like my [MMOGs] to embody user creativity. I also dislike cliques, so I have tried to design so people who wouldn’t normally hang out together come to realize each other’s importance in the world, the value of their roles in the society, that sort of thing. So I try to have interdependence as a key feature – people relying on each other, not in the moment-to-moment sense, but in the sense that our modern lives would fall apart if there weren’t people in a zillion jobs doing things we never think twice about, from stocking grocery shelves to manufacturing pens. The heart of UO was in many ways the original ecology system, which I wrote about at some length on my blog. It didn’t pan out, but even what we managed to get in there did in fact open up a lot of doors. I think we hit a lot of [our design goals], and were close to having much of it working, but the PK [player killing] problem basically undermined everything.”
UO inadvertently popularized several now-familiar online dysfunctions – especially player-versus-player (PvP) griefing. Koster and the Live Team wrestled with UO‘s escalating problem of high-level player characters killing and looting lower-level characters. Koster argued against a “PK switch,” whereby players could select whether to participate in, and be vulnerable to, PvP combat. After Koster left Origin, UO instituted an area-based PK switch, segregating its shards (servers) into two facets (worlds): “Felucca” (unrestrained PvP) and “Trammel” (PvP only by mutual consent). Most players migrated to Trammel.
“A big part of why I fought the PK switch was because it meant we were trading away player self-determination for security – echoes of today’s political situation, in some ways! UO often felt like long days of taking out things we had put into the game because players found ways to hurt each other with the toys we gave them. But the goal was still self-determination and freedom. The Ultima series was about learning to live with the Virtues, and I wanted the MMOG to be about the same thing.
“We worked hard to reduce PKing without instituting the switch, and based on what I saw, we did in fact make steady progress. But after I left the team, the introduction of the Trammel and Felucca facets [settled the issue] in a very different way than I would have chosen. I would have kept to the general path we were on. The various systems like stat loss and ping-pong murder counts were having a gradual effect on PK attacks.
“If we had gotten to the natural next step, which was player cities with control over PvP within their territory, I think the real nature of PvP in the game could have emerged.
“On the other hand, in terms of what I expected players to do with it, I think [UO] exceeded every wildest expectation. The players don’t care about what you wanted there, about what the dreams were – they only care about what they have in front of them, and then they proceed to do things you never imagined. And in UO‘s case, a lot of what they managed to come up with was truly amazing and not at all something I had ever pictured.
“I literally have not logged in for several years now. It’s changed beyond recognition for me, in some ways. I am still in touch with some of the players, though.”
Star Wars Galaxies: What Went Wrong?
After almost three and a half years of development, Sony Online Entertainment launched Star Wars Galaxies in June 2003 to mixed or negative response. SWG peaked in 2004 below 300,000 subscribers. SOE revamped the game, first in April 2005 and more drastically – some say disastrously – in November, with the New Game Enhancements (NGE). Bruce Woodcock estimates SWG’s current player base at between 110,000 and 175,000.
“When I started on SWG, one of the first things I did was write down the things I thought the game had to mean: a great civil war, political battles (under the influence of the new trilogy here), the Light and Dark sides of the Force, swashbuckling derring-do, a rich and diverse world and, lastly, community – because this last one is common to all MMOGs and also because Star Wars fans are themselves a strong community. My goal was to marry the open-endedness of Ultima Online with all the content and depth of EverQuest. My feeling at the time was, ‘This is the single best opportunity ever for MMOGs to swing for the fences, damn the cost – if not now, when?’
“Fundamentally, SWG was launched too early from a game design point of view. It may not have been from a financial point of view – there’s considerations like how much had been spent, how soon it would earn back the investment, that sort of thing – but most systems in there were first-pass at best. The place where that was most obvious was in the relative lack of content at launch. The tools simply came on too late to make the volume of content needed, and even though a heroic final push tried to populate the game with distinctive content, it just wasn’t anywhere near enough.
“A large chunk of the blame lies with me, for being over-ambitious with the design. I don’t think there were all that many fundamental problems with the overall design itself – some, sure, but nothing like the closed-economy debacle in UO, for example. [The systems] were first-pass, but mostly conceptually solid. (I am sure current and ex-SWG players will want to argue this point in detail, but hey, this is an interview, and there’s no room to give my detailed postmortem on every system! Yes, I still think something like HAM [the threefold Health-Agility-Mental damage system] could work, but yeah, it was probably too complicated.)
“I was not involved directly with SWG from about four months after launch. People seem to think that as [Chief Creative Officer] I was somehow in control of all the design being done at SOE. That’s not really the case at all; I had some influence, but I spent most of my time doing pitches, R&D, publicity stuff and that sort of thing. It’s been well over two years since I did anything significant on the title. The last things really finished under my tenure were player cities, mounts and the Warren – and by cities, actually, I was barely involved. I had philosophical disagreements with a lot of the direction taken after that.
“I didn’t like a lot of the subsequent choices – thinking here of Holocron drops, [Temporary Enemy Flag] changes, group sizes, buff expansions, global market, entertainer changes, creature handler removal, levels, a lot of the combat upgrade … I could give you my opinions there, but there’s no point – even those changes have been changed. Many of them were essentially trying to dig out of the hole that launching prematurely had caused. I can’t really blame anyone for them, whether I agreed with the decisions or not. The whole time, the team worked their ass off and tried their best.
“I’ll make an exception for the NGE. I don’t think you can or should change a game that radically out from under a user base. You dance with the ones that brung ya, whether they are the market of your dreams or not. They have invested their passion and built expectations about where they want the game to go. Changing things out from under them isn’t fair in my mind, especially given how they have been loyal to you in times of trouble. It’s like dumping the girlfriend who has always been patient and loving to chase after the supermodel who probably won’t love you back.”
“I wanted to do things that weren’t really possible at SOE, for a variety of reasons. Oh, we’d had our disagreements, too – the NGE changes in SWG, for example, were a big one. But it boiled down to the fact that I had a few projects and directions I really wanted to pursue, and SOE wasn’t at a place to pursue them. They couldn’t fund the projects I wanted to do, because of other projects already going. They generously let me start working on going indie before the contract term was up, even, which they didn’t have to do – it was a very amicable departure.
“I really wanted to get back to working hands-on on a game, and I also have firm ideas about the next directions online games are going to take. The cost of making things like World of Warcraft is pretty much completely unsustainable for most companies, and it’s mostly in making endless numbers of static quests and awesome artwork. The next generation of MMOGs, outside of a few blockbusters, will have to be about something different: lower cost development, smaller mountains of static content, more embracing of techniques that reduce those costs.
“I don’t doubt the DikuMUD-based game we’re all still playing will have legs as long as there’s people who still haven’t tried it out, but it won’t keep the current players happy forever. That means new sorts of virtual worlds have to come into being, or else all those folks will just flow right back out of the market. It’s way, way past due that we get out of the tank-healer-nuker game I got bored of back in 1993.
“I am working on a startup company, but we’re running quiet for a while. I’ve got that book to write. I also have two or three game ideas banging around that I really shouldn’t work on, since I need to focus on the startup! The public doesn’t get to see it, but I usually design a dozen games a year, of different sizes. Sometimes they are trivial little card games you can play in five minutes, and sometimes they are complicated German-style board games that take an hour or more to play. Sometimes they are puzzle games on the PC, and sometimes they are online worlds. I just enjoy making games, so I imagine I’ll keep doing it.”