GDC08: Live from Rob Pardo talks about Blizzard's approach to MMOs

Rob Pardo, Blizzard's Senior Vice President of Game Design, is speaking at the Game Developers Conference in San Francisco this afternoon and WoW Insider is here live to catch it all. Though the auditorium is just filling up, Pardo should be out in a few minutes to tell us all about Blizzard's approach to multiplayer game design. Catch the full details after the break.


12:01 PM PST: Pardo takes center stage!

Blizzard makes the multiplayer game first and fills in the single player game next. A good example of this is Starcraft 2: development had very little done on single player except for basic story development. Once you have a great multiplayer game you can build the single player game, but it's hard to go the other way around.

12:03: PvP is different than co-op. Important aspects:

  • Balance

  • Skill differentiation

  • Ladders/ratings

For co-op, important aspects:

  • Chat/communication methods: this goes beyond just chat. Tools for raid leaders so they can organize raid members and communicate strategy.

  • Complementary classes, races, etc. Players need to be able to act as a team.

12:06: There are also non-combat styles of player versus player. For example:

  • Economy. Players compete in the auction house in WoW, bid, outbid, etc. As an example, EVE Online has a strong economic system along these lines.

  • Politics

  • Racing, i.e. competition to race to max level

  • Dating (yes, dating)

12:09: Skill differentiation is important -- skill should be part of advancement and competition. Examples of skills:

  • Twitch/Micro (FPS/RTS)

  • Multitasking (RTS)

  • Strategic thinking (RTS)

  • Economy dominance (RTS)

  • Knowledge of opposing classes and races (All)

  • Map Knowledge (All)

  • Avatar improvement (MMO) -- questionable, as this can be a time investment rather than a skill

12:12: Game balance. You can't have everything, so you have to pick criteria to focus on. For World of Warcraft, Blizzard wanted:

  • Solo to max level

  • Have an important role in a group

  • Have an important role in a raid

  • Competitive in group PvP

  • Be fun!

For StarCraft II, Blizzard wanted:

  • Each race totally different

  • Better players can win games fast -- this is also a skill differentiation issue

  • Offense over defense -- you shouldn't play SimCity in an RTS!

  • Creative strategies

  • Every unit has a counter

  • Be fun!

12:15: Math vs. fun

  • Math is the foundation of your game system.

  • Everything should feel overpowered -- don't use math to balance the game to mediocrity. Every class should feel powerful.

  • Must understand the game's nuances -- you need to play the game. You can't balance the game by watching replays and you can't understand the game through spreadsheets.

  • Balance for all skill levels -- though balance is most important at the upper skill levels, you need to consider it for players of all levels. Also balance for groups of all sizes.

  • No super weapons! No iwin button. It may be fun for the player who has it but it's not fun for everyone else if one player is completely unstoppable. Everything should have potential counters to keep players from feeling helpless. As an example of how Blizzard approaches super weapons, look at the nuke in StarCraft.

  • Use your beta! (No, StarCraft II is not in beta.) You might buff things in beta to encourage people to use them and try them out, and then drop them down to a balanced level later.

12:21: UI affects balance

  • Unlimited selection -- imagine if you could select a hundred Zerglings (etc). This makes them much easier to control and concentrate fire or coordinate attacks.

  • Multiple building selection -- similar to above.

  • Sub-groups -- if you had a number different units selected together, do you only get the basic abilities of all or do you get access to abilities usable only by one or two of the units? The latter makes control easier.

  • Click-to-move -- this was experimented with in World of Warcraft briefly and made it more Diablo-like. You could click on a player and auto-path to them to attack -- this was considered too much automation and was quickly taken out.

  • UI Mods -- these are great, but they can be used to make the game much easier for some. Do you balance the game for the modders or the nod-modders? Sometimes we make the choice to break the mods and make changes to prevent them from working.

12:27: Balancing never ends

  • For players, change is almost always bad.

  • But you need to maintain your game to keep it relevant. If you want your game to last for years you need to support it for years.

  • Ban the cheaters! It's important to keep them out or they'll drive players away. You need to have technology to find them and get rid of them.

  • Plan your patches, but leave time to be reactive. Blizzard plans World of Warcraft patches far in advance, but they also need the ability to respond to game strategies or events immediately.

  • Don't panic! Though a strategy may develop that seems overpowered, players may develop strategies to overcome it. If you make game changes too often, players will use you as a crutch instead of figuring things out on their own.

12:31: Player psychology -- not always predictable!

  • Perception of fairness. Your game may be completely balanced on a spreadsheet, but if the players perceive it differently that doesn't matter. You need to be aware of these perceptions when you make changes -- you can't argue with a million people on your forums and player opinions are going to spread.

  • Players hate losing. For example, in Counter Strike you could win ten games and lose ten games in an hour -- it's so little time commitment that people don't mind losing. It's a problem with WoW's battlegrounds because they require a larger time commitment. Blizzard has tried to balance this by giving losing players something too. But if players lose consistently they might stop playing your game.

  • How do they climb the ladder? For example, if the loser gets nothing and/or loses points when they lose, your top players have no motivation to play. They have more to lose than win if they bother playing, which isn't any fun.

  • Right amount of complexity. For StarCraft II, Pardo says he feels 15 different units is the magic number. It's enough to have complexity but you can learn it and, as you become experienced, you can know the ins and outs of each unit. As a counter-example, look at Magic: The Gathering, which has thousands of different types of cards/skills/monsters. It works for MtG because MtG is turn-based -- everyone has time to think about whatever comes into play that they might not have seen before.

12:38: Incentives drive behavior. As an example, look at Alterac Valley...

When Blizzard designed it, they wanted it to be an epic battle where each team would rush into the middle of the Valley and meet in an epic battle between Horde and Alliance.

However it turned into an issue where both sides would run right past each other in order to storm the opposing faction's base. Pardo notes that if there were a /wave emote usable while mounted, they would wave at each other as they passed. "Have fun storming my castle!"

12:40: More examples:

  • AFKers in AV. They get honor whether they participate or not, so they game the system.

  • Inspect -- players felt like they were being stalked by people following them around inspecting them. Blizzard removed the expect message.

  • The rest system -- it was initially perceived as a penalty. By simply presenting it in a different way, it's now seen as a bonus.

  • Warcraft 3 "thumbs" -- to prevent people from using the same maps all the time, Blizzard allowed players to give each map a thumbs up/thumbs down. Then, when put into a match, it would attempt to match favorite maps. But sometimes you'd wind up in a thumbs down'd map anyway... players objected. Blizzard added in the ability to completely veto a certain number of maps.

12:44: Visual clarity

  • Does a unit, gun, weapon suggest its function and power?

  • Special effects can cause confusion in battle -- designers may love all of the possible visual effects, you have to make it easy to determine what's going on

  • Team differentiation -- Alliance vs. Horde, team colors, etc

Who runs faster? Who can take more hits? Make it visually obvious. It's more difficult in an MMO because players can customize their characters, but Blizzard has tried to suggest these types of things with class armor sets. (Just look at the same character model in two different armor sets.)

12:47: Maps

  • Random vs. pre-made -- there are advantages and disadvantages to both. (And no matter how good your generator is, player perception is always going to be that the map generator ruined their chances. You have more control over this with pre-made maps.)

  • Number of maps -- more is not better! More maps means players will never get the feel for any map.

  • Black shroud vs. dark fog of war -- players might not want to play new maps because it starts out completely blank map that you need to explore to see. By changing this to a "fog of war" that allows you to see basic elements, players aren't scared of new maps.

  • Larger maps aren't necessarily better.

12:50: Matchmaking

  • Fewer buckets are better. If you give players too many options (you select your maps, you select the race of your opponents, etcetc) then eventually you won't have anyone else in the same "bucket" as you are -- no matches!

  • Feedback on the system is critical! Players want to know whether it is working or not. For example, Blizzard's LFG feature: no feedback! No one uses it. As an alternate example, the battleground queue system provides information on queue waiting time, etc.

12:52: E-Sports

  • You need to decide up-front if you want an E-Sport viable game.

  • Remember that part of the fun for players is watching these matches! So the ability to do replays and a spectator mode is important.

  • You need tools for referees to intervene whenever possible.

  • Web support -- providing data and information about matches.

12:55: Q&A

Q: What about spectator mode for WoW PvP?
A: We didn't design WoW up-front to be an E-Sport game. We're now trying to retrofit this in -- I hope one day to implement some sort of spectator mode, but it's not trivial.

Q: Asymmetry seems to be wanted by many players...
A: You can't balance a Zergling and a Zealot one-on-one. You have to take a bigger approach. At first we're balancing for fun.

Q: On balance and psychology... why don't you have specific statistic reports on the web showing informational breakdowns on who wins which battleground/etc?
A: This is an example of where information can work against you. We consciously don't present which race wins most in Warcraft 3. The numbers could be skewed by a few really skilled players who favor Orcs, and then everyone wants to play Orcs and it's unbalanced.

Q: Do you think there's a sweet spot you can hit where your game is balanced or do you have to keep buffing things to keep players happy?
A: For example in StarCraft we haven't made a balance change in a long time. For example, in StarCraft a couple of years ago we implemented a bug fix patch which introduced a balance bug. If we weren't actively watching it and ready to fix it, the game could have died.

Q: What's your opinion on cross-server interaction?
A: When we first developed WoW we wanted the realms to be communities in and of themselves. But it's something we've slowly been softening our stance for -- you can see this already in cross-realm battlegrounds. In the future we might consider doing cross-server chat.

Q: You were saying about not letting super-weapons into the game...
A: In a competitive player versus player game you want to avoid weapons, skills, tactics, that one player can get that the opposing player doesn't feel they can realistically stop. As an example, in Descent 2, they introduced powerful weapons -- a new player could pick up a power-weapon and one-shot you. You don't want that.

Q: You stated that you developed the PvP side first... is that also true for WoW?
A: For WoW we developed the PvE side first, it was focused on PvE. This goes back to the fact that we didn't design it for E-Sports.

Q: What's become more apparent in Alterac Valley is the chokepoints that seem unbalanced. Horde has a chokepoint in the center of the map while Alliance has a chokepoint right at their base. If the Alliance holds the chokepoint, both sides get honor, but if the Horde holds the chokepoint, the Alliance gets nothing.
A: This is kind of off-topic... AV is the ultimate post-mortum in PvP game design. We talk about AV about once a month and implementing changes.

Q: Are there any plans in the future to integrate any kind of video capture?
A: It's on the wish list, but I don't know if it's realistic to get it any time in the near future. We didn't design the technology up-front to allow for it.

Q: Have you considered micro-transacrtions in WoW?
A: We chose to go with the subscription-based model instead of that approach. We've taken the approach that we want players to feel like it's a level playing field once they're in WoW. Outside resources don't play into it -- no gold buying, etc. We take a hard line stance against it. What you get out of microtransactions is kind of the same thing and I think our player base would feel betrayed by it. I think that's something else you have to decide on up-front instead of implementing later.
Q: But it might make it easier for the casuals to catch up....
A: They aren't going to be the ones spending the money.

Q: You mentioned that in Alterac, people weren't quite doing what you thought they would. How often do people do unexpected things?
A: People do unexpected things all the time. The coolest thing about doing multiplayer game design is that you do get better at predicting player reactions, but they always surprise you.
Q: How often do they do exactly what you think they'll do?
A: I don't know -- when we do our job right, I guess.

Q: I know WoW wasn't designed for PvP necessarily... but what do you think of players who just stand near people for killing and get honor, while people doing defense and not killing anyone would get less honor, even though they're helping the team more.
A: It's very hard -- whatever you do, players will find a way to exploit it. If I had the ultimate solution, we wouldn't have Alterac Valley issues anymore. If you can think of any way to exploit something you'd better not put it in.

Q: About end-game balance... you reach max level, you don't get any experience, but there are all of these raids you can run. Why do you have a content cap as opposed to a level cap? (I.e. you could keep leveling, but there's only content up to level 70.)
A: The reason we go with a level cap approach is that ultimately, we need to know what we're balancing our game around. If we had a soft cap and players could keep leveling, players will keep leveling and then there won't be any quests, any content, and you run into the South Park issue where you're killing thousands of boars... we're devoting our time and energy to developing new content so players at the bleeding edge hopefully won't have to wait more than a month or two before there's new content for them to explore.

Q: What about balancing solo play? Does it bother you that a warrior levels more slowly than a hunter, for example?
A: We wanted to make sure that all players could have a fun play experience by themselves. From a balance perspective, it's not so hard to do that. It's more complicated when you try to give everyone a role in the group and make everyone viable in PvP and you throw in talent specs...

1:19: We're out of time and Pardo has left the building -- but stay tuned for more updates from GDC.