Blizzard Entertainment VP of game design Rob Pardo opened the Austin Game Conference this morning with a keynote exploring the decisions and management style that has led World of Warcraft, Blizzard's flagship massively multiplayer online game, to precedent-setting success in the genre. With over 6 million subscribers around the world, World of Warcraft has become the most successful MMO ever to come out of the Western hemisphere. Pardo covered a few of the core group of philosophies and approaches that the Blizzard team took in developing their game. Blizzard, according to Pardo, envisions their market as a donut, with the core players at the center and a more casual audience around that. "It's increasingly important to appeal to both," Pardo said. Blizzard leans heavily on the "easy to learn, difficult to master" mantra, Pardo said, but also does things like lowering system requirements to build a broader audience. To maximize both depth and accessibility, the company concentrates on depth first and accessibility later, Pardo said. "First we try to come up with what are really cool things, things that will get people to play for two to three years. Then we actually start talking about accessibility, how to make the content approachable and easy to learn. But it starts with depth first." Accessibility decisions always start with the user interface, Pardo said. "One of the first pitfalls that happens in user interface development is trying to make everything visible on the interface itself. We streamline our interface. We present information that's really important in the heads-up display, but there are a lot of other things we put in that actually are a little more difficult to get at." Accessing the auction house was an example Pardo gave: rather than making it a button on the user interface, WoW's auction houses are of course accessed by speaking with an NPC, thus making the user interface itself simpler and more accessible, without sacrificing depth of gameplay. Pardo gave a number of examples of design decisions that had been made to support more depth in the game. Rather than create a plethora of character classes, for instance, Blizzard wanted to make each class distinct and recognizable, so that players would get deeply involved with each one. Instanced dungeons were also designed to provide additional depth. "The dungeons are there to serve the core markets, not the casual players," Pardo said. "But they do become the bridge for casual players to become more hardcore." One of the other mantras Blizzard works by, according to Pardo is that of "concentrated coolness," which helped lead to the smaller number of character classes. A lot of initial class ideas came from Warcraft 3, Pardo said, but rather than have a Mountain King, a Blademaster and a Tauren Chieftain class, for instance, WoW combines abilities of all three into a single, deeper class so that players are more engaged with each class on its own basis. The company also tried to remain conscious of trade-offs in game design, and the fact that "every decisions costs" in one way or another. "All game designers are very greedy by nature," Pardo said. "We want to have every cool feature, and we want to serve every market. The reality is that almost every design decision comes with a trade-off. Nothing in game design is black and white, it's all shades of grey." "Whenever we can, we try not to compromise, but that usually results in both sides being unhappy with the result," Pardo continued. Pardo raised the example of instanced dungeons, which in World of Warcraft are not normally solo-able. "We chose to make dungeons hardcore. We chose not to have solo instances and dungeons. It would just have compromised the gameplay." Pardo listed a number of specific trade-offs that World of Warcraft had been forced to make. One was in terms of system requirements, and an artistic style that was not in line with the hyper-realistic trend of games like Crysis, for instance. "We would rather have lower system requirements to reach a broader market. This means having a stylized art style that is resistant to looking dated" even as hardware becomes more powerful. "That generates lots of negative press," Pardo admitted. "Our screenshots are never going to compare to something like Crysis. Crysis-level graphics are easier to market, and developers like to make that kind of art. Plus the press likes it. But ultimately, I think we're proven that the gameplay is really what matters, and if you have cool stylized art, that's enough." Another trade-off was in the area of transportation-related game mechanics, specifically, the need to use "taxis" to get around the world of World of Warcraft, rather than being able to teleport at will. "The use of flight taxis to maintain integrity limits random teleportation and instant access, but it makes the world feel more epic," Pardo said. "That way, you can have remote areas, like N'Goro Crater, where we very consciously did not put a flight path. We really wanted it to have this Land of the Lost feeling. But of course players get very frustrated with it. They call it World of Walkcraft, or World of Travelcraft. But on the other side, it creates a lot more social connectivity, which is what MMOs are all about." High-level gear was another area where compromises were necessary. "It was really cool in Ultima Online to make your own character or guild with your own colors and your own look," Pardo said. "But there's only so much art time that you have, and only so many artists. We chose instead to concentrate that coolness on making really cool prestige armor sets that came from specific places. We allowed raiders and more hardcore players to get that prestige and reward, so that you can recognize where someone has been. With customizable gear, everyone can look different. But you can't have both." Pardo also spoke about "the Blizzard polish," and the fact that it's important not to leave polish until the end of the design process. "The usual assumption is that it's something you do at the end, and that the reason we were successful is that we spent 6-12 months at the end polishing things. But we do polish from the beginning. It's a constant effort. You really have to have a culture of polish in place as something that everyone has bought into. If you do it from the beginning, you have a much better chance, but if you leave it to the end, it's going to be more difficult." The first phase of polish starts in the design process, Pardo said. "We consider everything. Is it fun? Does it have solid mechanics that will work in the endgame. Will it work in this encounter, in this dungeon? The art and cool factor is also something we talk about. Is it inherently cool? And contrary to popular belief, we actually do consider production and talk about how long things will take to do. We try to design smart where ever we can." Mounted combat, for instance, was an element that was left out of the design because of production concerns, Pardo said. "It's something we don't see a lot of bang for the buck for us in our game style." The second phase of polish is in the creation of content. Zones, for instance, are initially drawn on a whiteboard so they can be easily critiqued and changes. The newbie zone of Northshire was the first zone the team designed. "We tried to get a feel for what's downtime, does the combat feel fun, how's the pacing for the quests, how much experience do they get, where do we put the trainers," Pardo said. "We didn't go out and build the entire World of Warcraft until we king of knew what we were building. Our mantra is to make it fun, then make it big." Small factors also have an impact. "Control is king," Pardo said. Even three frames of mouse cursor lag is going to have an impact: "People will leave your game over it, because they can feel it, but they're not going to tell you about it in the forums because it's subconscious." By the time the game gets to beta, then, it's almost complete. "Beta for us is not really about finding bugs, it's not really about getting a lot of game feedback," Pardo said. "It's much more about stress testing the game from a technical and gameplay level. We encourage beta testers to exploit the hell out of the game: Is there a zone where they level too fast, do they get phat lewt by ripping through instances?" As a corollary to all of the above, Pardo warned developers and publishers to not ship a game until it's ready. "If you ship a game before it's ready, you cripple your changes for success," he said. "With MMOs the stakes are much higher. Because of the subscription model, people don't tend to come back and check in on a game again. People don't want a second look. You're really putting at risk the next five years or your product." Pardo closed his remarks by urging his audience to push the MMO genre in new and different directions. "The thing I really think is unique about MMO games -- in all other genres, FPS, RTS, all those other cases, the genre depicts a very specific type of gameplay. But in MMO games, all that says is that you have a game with lots of players in it. So this genre really has the biggest frontier. We should be pushing in all kinds of different directions." [Mark Wallace is the editor of 3pointD.com, a widely read blog covering virtual worlds. His freelance journalism on technology and culture has appeared in Wired, The New York Times, PC Gamer (UK) and many other publications.]