WoW Insider had correspondents there on the site, and they sent back audio of Kaplan's speech. We've paraphrased the salient points, and you can find them all after the break. There's some really interesting stuff in there, including the fact that in the past two years, 80 billion quests have been completed in North America's Azeroth alone, and just who is behind the frustration that is The Green Hills of Stranglethorn (hint: it's Kaplan himself).
Hit the link below to see what Kaplan told the crowd at GDC.
"Cruise Director of Azeroth" -- doesn't refer to me, it refers to all the guys directing gameplay in World of Warcraft. Directing gameplay can be used to show off mechanics or force the player to do something. But it can also be an underlying tool to immerse the player in your game. Bioshock does this by using a yellow arrow, TF2 puts objectives on the screen, WoW uses achievements to send players to do new things. With some Xbox 360 games, some players even use the achievement list ahead of time to decide how they're going to play the game.
Level design is another way to direct gameplay -- Kaplan showed off a shot from Half Life 2, where the only way for a player to move through a level was through a set of gates, and there are even two guards flanking the gates, which is a clear signal for players to go that way. Lighting also helps a lot as well -- "Valve does a really good job" with directing gameplay. Quests are another way -- "if you show everybody The One Ring and Mount Doom, everybody gets it."
From June 2007 to March 2009, over 80 billion quests in North America were completed, a daily average of 16 million. Early quest goals for WoW: they wanted players to enter the world and explore with free will, but they weren't standing by user-generated content (Warcraft III maps is an example) or randomized content. Blizzard would hand-place and create everything in the world. Early on, they looked at Everquest, and counted up about 1200 quests through Shadows of Luclin expansion (possible they have way more than that). So Jeff Kaplan and Pat Nagle aimed to create about 600 quests to compete with the 1200 quests. Internal alpha really emphasized the newbie experience (and had awesome stuff like a Shield Bash, pre-level 10, that stunned for a few seconds). But they found that the game felt empty any time there was an empty quest log, so the shipped version ended up having about 2600 quests, shifting the focus of the project to being extremely quest-driven, so that from the first steps, you're following a quest. Burning Crusade had about 5300, and Lich King had about 7650 quests.
Interface for quests had to be really obvious -- big exclamation point over someone's head. Hardcore MMO players didn't like that because they wanted quests to be much more subtle and hard to find, but it wasn't very accessible. Quest log was also an innovation -- players used to play with a notebook (and in the geekiest setup, two stopwatches) next to their computer, but Blizzard wanted to improve the experience. And they focused on the quest log's usability -- put clear goals and objectives at the top, leave the story to players who want to get into it. Also make sure players knew the game "recognized" any collectibles they picked up, immediate feedback when players complete part of an objective. And in patch 3.1 (which should come out "any second now"), Blizzard borrowed from the addon community and will add a hover tooltip to the objective tracker.
The final decision was to make completing quests "the smart way" to play the game -- even if you don't pay attention or need the armor upgrades, doing quests is always a great way to advance and progress. Experience and gold rewards turned out to be the most important rewards for players. Tuning is often the most important part of game design -- you, as a developer, are much better at your game than everyone else, so it's better to take the time to go back and make sure it's right for players.
WoW made plenty of mistakes (he doesn't want to give the impression that Blizzard is perfect), and he went over a few of them, to keep everyone from repeating them. The "Christmas Tree effect" (a common term around the Blizzard office) is the first one -- when you first go into an area, and your minimap is "lit up like a Christmas tree" with new quests. Players like it, but as designers, you've lost control over what players are doing -- they're just going to vacuum up quests without paying attention to story or which questgiver gave you what to do. And the quest log is full and you don't do them in any order -- you run around and just start mousing over things (if you have Questhelper installed), or just killing stuff and hoping for the right drops. It's not "don't give a player choice," it's "be smart about how you give player choice." In Lich King, Blizzard slowed down the amount of quests you have access to at any given time.
The second mistake is a phrase from the forums: TLDR (too long, didn't read). Designers sometimes think their work is so awesome that they don't realize just how boring it is to someone who doesn't really care. So they have a limit on quest text to 511 characters (not words, characters). New people at Blizzard have said "let's make it longer," but Kaplan likes it short -- people don't want to read anything, you need to show the story, not go crazy with the text. "Medium envy" is when videogame designers think they should be writing books or movies when they should really be making entertainment. Deliver the story in a way that is "uniquely videogame." "Shakespeare couldn't 3D model his way out of a paper bag, Scorcese couldn't program ragdoll physics, and The Beatles are pretty lousy at balancing three unique races on an RTS battlefield."
Mystery is something Blizzard gets wrong as well. Players want goals, not mystery in the action of what they're supposed to do. Even on a mystery story, Blizzard shouldn't ever say "something's wrong in Elwynn Forest, go figure it out." Always have something concrete for players to do next. "The mystery can't be in what to do." Plus, there's spoiler and helper sites out there, not to mention General chat anyway. Make the action in the gameplay, not in figuring out what to do next.
Poorly paced quest chains are another issue -- a quest that starts at level 30, spans 14 levels, and ends with you killing an elite mob. So basically putting a brick wall in front of the player to bang their head for a little while. It's cool to have expansive quest chains, but bad pacing loses the player's trust that it's worth it to do.
Gimmick quests without enough polish is another mistake -- gives an example of a popular shooter with a gimmick vehicle level. Focus on the core of your game, and then direct the player to it. Don't make the game more fun for the designer than for the player. If it's only fun for the gimmick, it's probably not fun at all.
Bad flow is another quest design mistake -- quests clustered up, with kill quests all together, and collection quests all together, and not a smooth flow between them. He showed off a suggested new flow for Loch Modan, which is a place that is not well designed at the moment. "I always tell people, if you do one thing in World of Warcraft, do the Death Knight starting experience." It's one of the best quest design places in the game to date.
He talked specifically about collection quests -- the Shimmering Flats, when you first show up and you have three collection quests all together. There are too few creatures, too few items, and too many to collect. Creature spacing also is an issue, and too much creature variety can be a problem as well -- fighting a ton of different creatures in a row will throw off your pacing.
"And the worst quest in World of Warcraft:" The Green Hills of Stranglethorn. "I made it. I'm the asshole who wrote this quest." He thought there would be a whole economy to itself, and it was the best idea ever that turned into a terrible quest. "Utterly stupid of me." Inventory management is an issue as well -- at all times, players are making decisions about what goes in the bags, and for a single quest to consume 19 spaces in your bag is ridiculous.
You have to think, as well, about why we're collecting these items -- why do I have to bring back Gnoll paws in Westfall when I have to kill bandits as well and don't have to bring anything back for them? Not to mention that shouldn't every Gnoll have two or even four paws? The player loses trust again in Blizzard as designers. You never want the player to even think that somebody made the game. If they're asking why something happens, then you've failed. But you can pull off a collection quest as long as the payoff is worth it -- example of the poisoned pumpkins in Brill, or inserting them into quest chains.
One mechanic used to solve collection quest issues is the progressive drop rate. Usually their ideal drop rate is around 35%. But that would cause problems, as players would run into streaks, both positive and negative. So they used a page from Warcraft III, where a unit that would crit would use a progressive percentage. So now, every creature drops the item, but the more a player has killed a creature type, the more likely it is that a player will get the item. So the first time you've got a 16% chance of getting the item, second time 32%, third time 42%, and then eventually up to 100% chance you'll get the item. In the beta, it got rid of the bad streaks, but it also got rid of the good streaks, so they eventually upped the "ideal" drop rate to around 42%.
Finally, in conclusion, every little decision we make as designers has an effect on the player. It's not good enough to just put some stuff into the game as an experiment -- it's up to the designers to make sure players have an experience, not just a chance to have one. Grand Theft Auto IV is an example -- it's an open world, but every time the player does need something to do, the developer is always right there to tell them where to go. You can only be on one mission at a time, so it's actually very constrained, but it's elegant game design, so it feels like an open world.
Q: Of any game I've ever played, WoW seems to have the most shared information -- Wowhead, other sites, etc. How do you design around that?
A: I talked about that when I talked about mystery -- we wanted to eliminate the mystery from what you do, and put it into the actual gameplay. Addons like Questhelper and Carbonite have forced us to do more than what we would have usually done. It's forced us to evolve with how players are playing the game.
Q: How many different quest designers do you have on the team?
A: I think the number is around five currently. It's gone up and down over time. There's a lot of collaboration -- no one works solely by themselves. Original game was shipped with a team of 60, and now they're up to 130-140. But we can't let that team get too big, or it becomes a question of quality.
Q: Since the game is so big, where did you start with testing and how did you do it?
A: We didn't start with a formal testing guide, but one rule we have is that you have to play your own content. Which seems obvious, but a lot of new employees will put things in that they've never done themselves. Everyone at Blizzard is a gamer and wants to game, so sometimes they have to kick them off of Starcraft II to do other things. Friends and family tests and conventions are playtests as well. We don't have a formal process -- playtesting starts the minute someone implements something.
Q: What do you think about time as you develop these quests? Is there though put into flying around the world to complete the quests?
A: In our earlier quests, compared to the expansions, Lich King is really the direction we wanted to be going. We have far fewer quests that have a ridiculous travel time, and we plan to go back and fix a lot of those old quests. Our philosophy was that we need to move players around the world so they can see everything, but what happened was some quests sent you way too far. So now we're careful when we send you to another zone. It was an early mistake on our behalf. We have talked heavily about player pacing, player fatigue, putting in peaks and valleys, but some of the flight paths are still too long.
Q: WoW feels different from Warcraft II and Warcraft III -- did you have an intentional shift away from that IP into the bigger game?
A: Rob Pardo was our lead game designer on WC3 and WoW, and he was working on the Warlock class. He said, we'll give the Warlock Death Coil. And everyone said, no, the Death Knight has that, not the Warlock, and he said it fits more on this class, it's a mechanic we need. So that was intentional. But there were other times where it was maybe a mistake -- Horde and Alliance were buddy-buddy at the end of III, and now they're trying to fight again. Night Elves have lost a lot of their edge in WoW as well, compared to the earlier games. We'd like to correct that, but it's going to take a long time. "Everyone knows Blizzard speed -- we'll get to it." Our core is Chris Metzen -- it's very much about listening to what he's thinking of and then carrying out that vision.
Q: Thanks for giving Locks Death Coil. I had a question about pacing -- do you have formal metrics for that, or is it just a feel thing?
A: We do both. For combat, we definitely have a metric -- we wanted combat to last one minute. Nowadays it's much quicker than that, but we used that as a starting point. But for other things, it's very gut. The pacing of a quest chain that wasn't broken back in 2004, we might need to readdress it now, based on the fact that the players have evolved, and over time, we find mistakes in our own game.
They're telling me that I'm out of time, thanks for coming.
Stay tuned to WoW Insider for more from GDC and Blizzard at conventions throughout the year.