More classes are not always better. Once you get enough different units or classes, players can only handle so much. When you see someone, you might not know what they can do, and this matters because when you want to form a group, you lose track of the strengths and weaknesses. In battlegrounds, you need to know instantly what the opponent can do to you. Even if you have 50 completely different ideas that are cool, it's still important not to use them all.

　　Our class ideas originally came from Warcraft 3. What we chose to do was to take the heroes and combine them. Warrior got aspects of mountain king, blademaster, and Tauren chieftain from War3. We chose to concentrate the coolness.

　　Tradeoffs. Every decision comes with tradeoffs. designers are greedy by nature ' we want everything, moms, dads, cats and dogs playing together. Nothing in game design is black and white, it's all shades of gray. Whenever we can, we try not to compromise. It usually results in both sides being dissatisfied. If we had solo dungeons, then he group dungeon fans would feel their achievements would be cheapened. So we chose specifically not to have solo instances.

　　An example of Tradeoffs: system requirements of Wow versus Crysis, for example. Crysis looks awesome. But we would rather have the broader market. So that forced us to the stylized art style that is resistant to looking dated. It did generate lots of negative press, and our graphics programmers always wanted to push farther too. You just have to be prepared. But every game we've released, we have gotten the comment that our screenshots were not up to par.

　　There are benefits to the cutting edge side too. It's easier to market, and developers want to make the best quality art. You're fighting against developer psychology if you choose the other route.

　　World size vs teleportation is another. WoW vs Diablo. We wanted to the scale of the world to feel epic. But you get players getting frustrated and calling it 'World of Walkcraft.' You use flight taxis to maintain integrity and having limited teleportation means you can have remote areas where you consciously do not provide a flight path to it.

　　But on the teleportation side, you get a lot more social connectivity, which is what MMOs are all about. There's a barrier there if people have to travel and coordinate. We consciously decided to have that tradeoff. Players do want the convenience.

　　Another tradeoff is prestige gear versus customizable gear. Players ask for dyeing armor, all that. When I played Ultima Online I loved that. It was a great feature. But there's only so much art time you have, and we chose instead to concentrate the coolness on armor from specific rewards instead. The whole point for a lot of hardcore players is to show off your advancement. So we chose the best gear to be from raids, so we can recognize someone's achievements based on their gear. The tradeoffs is that you lose everyone looking different and users expressing creativity. And if you try to have both, you'll end up muddled and somewhere int he between.

　　The Blizzard polish. Polish is the word associated with us in reviews. There's this big assumption that polish is something you do in the end. That we're successful because we spend 6-12 months at the end polishing. We do get more time, but we do the polish right from the beginning. It's a constant effort. You have to have a culture of polish. Everyone has to be bought into it and you have to constantly preach it. if you leave it to the end, it'll be more difficult.

You'll get a lot of 'why does it matter that this feature is polished' It's so small.' But people notice 1000s of polished features, not the single polished feature.

　　Polish starts in the design process. (pic of skeletons in a room, which he says is the designers in a room). We're kind of in a new era at Blizzard, when i started we had very few people with the title game designer. That's been changing over the last few years. It's interesting bringing in an experienced designer from outside, because they want to make a unit week, add a mechanic constantly, work 100 miles and hour. We have to get them to slow down. You need to talk through things with everyone else, and you have 100 features and they all have flaws and don't work with each other. So when we are in a design meeting, we try to consider everything. Will it work in this raid encounter, in PvP, as a newbie, for the art, solid mechanics, etc. Contrary to popular belie, we do consider production. Mounted combat is an example of something killed by production time. Bounce ideas off everyone. Let the beer goggles wear off.

　　When we develop maps, we do it on the whiteboard, so we can iterate, and there's no cost to changing things.

　　Phase 2 is when we actually make something. The first thing we try to do is make it fun. Northshire valley, for example ' we spent an inordinate amount of time on it. Where do we put the trainers, how does the combat feel, etc. We probably spent more time on it than any other area, by an order of magnitude. After we made it fun, then we made it big. We didn't go out and build the entire world of WoW until we knew what we were building. It didn't make sense to do that until we had figured out all the details of the fun. If you have to retrofit the fun into the content, you're gonna be screwed. When we went into the friends and family alpha test, people were surprised that it was fun. It was a lot easier, once we knew what was fun, to do levels 10-20, and 20-30 and so on. The design at that point was creative design, not mechanics.

　　Control is king. Game control is taken for granted a lot of times. I remember on Warcraft 3 I could feel a little bit of lag on the mouse cursor, and I kept saying it to the programmer, but he kept saying he couldn't see anything wrong. Finally he coded in a hardware cursor so we could run both cursors at the same time, and lo and behold there were three frames of lag. And that matters, it's important. People will leave over that, but you'll never know that is the reason.

　　'Beware of the Grand Reveal.' This is a pic of a dungeon that was supposed to be in the original release but is in the expansion, because the subteam went off to work on it in a vacuum, disconnected from the rest of the team. The grand reveal was when they came back and showed it. It was supposed to be a raid dungeon but the doors were too narrow. So back to the drawing board it went, three months of redo because we didn't redo along the way.

　　Lastly, have fun with the game. Put in the little in-jokes. If developers are having fun making the game, chances are the players will have fun with it too.

　　Phase 3: the finish line. Feedback strike teams is something that we have used for a long time. We pull devs from all the teams and put together a diverse group with a mix of play styles ' RTS guys who don't like MMORPGs, etc.

　　Don't take small decisions for granted, especially in that newbie experience. We had cases early on where people grouped up with 1 other person that they would get into the next area at 4th level, and that meant they had a bad experience. So we try to ask a lot of questions and don't let things die on the feedback and striketeam list.

　　The beta test for us is not about finding bugs. It's not really about getting a lot of game feedback. it's about stress testing from a technological and gameplay level. We encourage our testers to exploit the hell out of the game. In our RTS beta tests, people always get upset that we run a ladder in the beta test, because the guys on top are exploiters. But that's the point ' we want to see who the top ten exploiters are so we can look at their games!

　　Don't ship until it's ready. This matters even more with MMOs. You might hear that it's improved later, but no one actually goes back to try it. You will really cripple yourself, you put at risk the next five years of your product. So hopefully all you publishers will give the developers more time.

　　I hope we turn this genre into something special. The thing I think is really unique about MMO games ' you look all the other genres, and the genre depicts a very specific type of gameplay. But massively multiplayer, this genre has the biggest frontier, it has the most we can achieve, and we should be pushing at all kinds of different directions.