Sections

In Their Own Words: An Oral History of Diablo II With David Brevik, Max Schaefer, and Erich Schaefer

Fifteen years after its original release, the creators of Diablo II recount the battles with Blizzard, the year of crunch, and the shower ideas that became some of roleplaying's most enduring pillars.

Retrospective by Kat Bailey, .

Planning and Production: New Worlds

Blizzard North had big dreams for Diablo II. After taking a few months to recover from the grind of Diablo, they returned to begin work on the sequel.

Erich Schaefer

The first six months were very slow. Through the middle of Diablo 1, the company had a very relaxed atmosphere. We would spend the whole day playing other games. We would all take trips to go watch movies, and go to Fry's and pick up the latest games. In a way, we wasted a ton of time. I think it was a good waste of time, and fun, but finishing Diablo 1 was such a grueling experience for us, especially, having not really done something that big before that, we took it real easy the first six months of '97. We got back and just goofed around a lot. So, imagine, once you get into that mode, it takes a while to ramp back up. So I think it was a bit of a slow sluggish start. I wouldn't say it was bad, though. I think we needed it.

Max Schaefer

Yeah, we needed it. And there's always this time when you're working with new tech and new systems, that it takes a while for the programming staff and the engineering staff to have things in place for the designers and artists to actually work on stuff. So it's kind of natural between projects that there's a little bit of downtime. But it's always a bit of a challenge to get it ramped back up.

Erich Schaefer

I think by the second half, and again, this is a long time ago, but I think by the second half, we had a solid Act 1, we had the town, the surrounding landscapes, and I think the initial dungeons. And I think we had all that stuff by the end of '97, so we really had our feel and our look and knew what we were doing by then.

We went back and forth a lot on whether we should wait to announce until we were almost done, or whether we should try to build hype early on. I do remember, there was concern that we were announcing it too early, but at the same time, I remember we announced StarCraft so early they didn't even have a game. They literally had some placeholder art on top of the Warcraft engine. And it looked completely different.

And so, at the time, we were like, "Hey, let's just announce these things, and build hype from the very beginning. Maybe it even gives us a little push to get the game done ourselves."' And then, after those two examples, Diablo 2, which was announced way too early, and StarCraft, which was announced before they even started, I think we then changed tactics and kind of said, "Hey, let's just wait four months before we know the game's done." But I know they've gone back and forth since. Most of those we do consider were announced too early, though.

Max Schaefer

Yeah, although, we did have some extra considerations with the Diablo 1, because it wasn't too long after it was released. It was gaining popularity, and we realized that we were dealing with a pretty significant cheating problem that was impacting people's online experience, and we wanted to have an answer for that. We wanted to say, "Hey, yeah, for Diablo 2, we're going to a client-server system instead of a peer-to-peer system, that should take care of it." We thought it would be easy to take care of the cheats then. Not so much. But, we did want to have a public message to address it, and part of that was, "Hey, we're making Diablo 2 now, and it's not going to have this problem."

Erich Schaefer

We had this wish list where we wanted to go [with Diablo II], and one of the obvious things was that we wanted it to not just be one town with one dungeon. We were going to make it throughout a series of lands. We ended up with three, plus the finale in Hell, but, we didn't really know. So, at first we were sort of just compiling a lot of visual ideas and where we wanted to go.

David Brevik

One of the things that we wanted to do that never got done in Diablo II was that we were trying to design this thing called Battle.net Town. You know, my idea was to get into a world, and you don't leave the world. I wanted to try to break that reality, break that illusion that we're not in a contiguous world. I wanted you to start in that world, not start in the Battle.net chat lobby, but start in that world. And Battle.net Town was going to be a glorified chat room where you could move around, but you could still go to the pub and chat with people, or go to a vendor, and then you would go to a specific person and you would travel to Act 1.

So, you never left the world, and that was kind of one of the things that didn't make it, but that was one of the things that we wanted the most, that I personally wanted the most, which was, I really wanted this kind of, more of an immersive world feeling to the game than we had, and make it more expansive. We're not just in the dungeon beneath this town, that there's a big world out there.

"There was no definitive design document [for Diablo II]. There was a loose structure. "Hey, I know we wanted to go here for Diablo 1, this is kind of the first act. This is kind of the story we're telling there." And then we get to the second act, and "Hey, where do we want to go," and then, "Oh, I got this idea for this level, it's going to be in space, it's going to be awesome." And we'd mock it up. "Yeah, that's great. Doesn't make any sense in the story, but what the hell, it's kind of cool." - David Brevik

We had been playing Ultima Online, and that had definitely some influence on this design, where you could walk from one part of the continent to the other. I was thinking, "Wow could we do this and mix it with Diablo?" And that was the way that we wanted to go, but we never really fully realized that dream. We got pretty close, in that we were able to stream a lot of the levels and things like that, but we never got to the point where you could go walk from the Sisters of the Sightless Eye Cathedral to the desert or whatever. That wasn't really fully realized for Blizzard until World of Warcraft.

Max Schaefer

The biggest thing was that we had the outdoors. We were going to expand from just levels of dungeon underneath a church and do real landscapes, real outdoor landscapes. So I was tasked with organizing the background team, and that was by far the biggest topic for us: "Now that we're going to be outdoors, what does that mean? Where should we go, where should we take this?" It really came down to, where would it be cool to be? It would be cool to start out in the Irish countryside, just like Diablo 1's town was in. But then, it's wide open, we should do desert levels, we could do rainforests or whatever. I think, again, we might have been one of the first, setting a long, now, long tradition, of having your second area in an RPG be the desert. You see that all over the place now. People start out in sort of an Irish countryside, then they go to a desert, then they go to something else. I remember, it was a whole lot of fun just pouring through books and looking at landscapes and looking at architecture and deciding where we wanted it to play. Where did we want to kill monsters?

I think doing a desert is always challenging, because it's mostly sand, and so you have to break it up with enough features that it's interesting. Sort of the same with a rainforest, in that it's so densely packed with foliage that you kind of have to find clear areas that you play in. It's sort of the opposite of the desert. Instead of trying to invent stuff to constrain you and keep you along the path, you have to clear out areas and make it plausible that you're in a rainforest, and yet you can see your character and there's room to fight monsters. To that end, we had the rainforest mostly take place along a riverbed.

Erich Schaefer

That sort of reminds me of another challenge we had. The gameplay was fine in the dungeon in Diablo 1, but we were bringing it to the outdoors with just wide open fields. We asked ourselves, "What are the boundaries? What are the obstacles?" If it's just a field, there's no real tactics where you move around. So we put a lot of work into what should the sizes of these outdoors be and giving them boundaries and borders that made sense so you didn't want to wander forever. So we came up with all kinds of weird low walls that surrounded a lot of our outdoors that don't really make any sense. I remember at the time a lot of people were very negative on it. It was like, why couldn't we just hop over the walls? Itt's going to drive people crazy that we're limiting their area. I thought it was a big concern. Turns out, it felt really natural, and I don't think it ever got brought up again. The idea of taking it to the outdoors was one of the big early challenges.

Max Schaefer

And it took a lot of iteration, too. I remember we had way too big an area for a while, and claustrophobic areas. It took a lot of testing and play to figure out, when do you need a choke point just to reorient everyone along the path, and how big an area was it before you miss things or you're lost or you don't know where you've been? It was just a lot of play back and forth and tweaking of the size of things to get it right, because it is really different being outdoors than being down in a dungeon.

Erich Schaefer

Also, we had to invent, as far as I know, we probably invented the idea of a quest log to keep track of what quests you were on, which ones you had finished. For ourselves, at least, we had to invent that. There was no such thing in Diablo 1. It was just, you could do whatever was there, basically. There were some quests, but they didn't really operate in parallel, and you couldn't track them and do them or not to them. So, that was all new territory: how people know what to do, where to go. The waypoint system for getting around, we had to invent all that stuff. So, none of those were in the design document. They just kind of... they were problems that came up as we started to play, started to make the big area, and we were walking out there in the field. Hey, how are we ever going to get back to town? How are we going to get back to here? Those sort of things came up as we developed, we didn't plan them out beforehand.

David Brevik

There was no definitive design document [for Diablo II]. There was a loose structure. "Hey, I know we wanted to go here for Diablo 1, this is kind of the first act. This is kind of the story we're telling there." And then we get to the second act, and "Hey, where do we want to go," and then, "Oh, I got this idea for this level, it's going to be in space, it's going to be awesome." And we'd mock it up. "Yeah, that's great. Doesn't make any sense in the story, but what the hell, it's kind of cool."

So, we'd put that in, and then, "Oh, let's go to the jungle. And we'll have little tiki men," or whatever the hell you want to call them there. Little goblins. And we'll have frogs and stuff. And then eventually, we knew we wanted to get to Hell, and what that would be like. We had sort of a loose structure to it, but there was no definitive design document, no specific things. Every day we would just iterate on it and change it how we please. And it's like, "Oh, I got this great idea, we're going to change it to be this way." And that was really more the way that we operated. It was like, as we played the game, we designed the game for features that we felt would be interesting additions to what we're doing, or make whatever it is that's broken better.

Erich Schaefer

Yeah, I think, actually, that's not that strange. We've never had, in any of the games since, the Torchlight or Hellgate or Rebel Galaxy here, I've never done a design document. Maybe that's my failure as a designer, in a way. At times we sort of fake one almost. We kind of said, "Let's just make sure that new hires have some idea of what's going on, or whoever our partners are have some confidence that we have anything." But, all our games, we've just sort of made as we go along.

Max Schaefer

It's part of our style that we try to have a running build going all the time, so that we can proof our concepts just basically by playing what we have, and it becomes readily apparent when everyone's playing the build what it needs. Do we need more variety in backgrounds, do we need certain kinds of monsters for this area? Once you have the framework in there, the details kind of fill themselves in.

Erich Schaefer

We sort of just plan for the month ahead, or the week ahead, even, at all times. "Hey, we need more monsters in this area, let's get a batch of monsters going." And again, I've said it a bunch of times, the iterative style where you, we constantly just evaluate the game. What's fun about the game, what's not? If this isn't fun, let's get rid of it, let's replace it with a different thing. And if we had a design document, we might be working on that thing that wasn't fun forever. Without the design document, we don't have to worry about that. At the same time, it makes bigger teams much harder, of course. It's hard to manage leads that then have to talk to their group and schedule out months ahead. That's pretty hard to do in our style. That's why, I think, that's why right now I'm working with two guys. It's just me and [Double Damage Games co-founder Travis Baldree], and it works fabulously with this iterative kind of style.

But it really takes some key people on the team, again, Dave Brevik is the great example with Diablo, to be able to just adapt on the fly. He will fix things that morning and get us going for the rest of the week on this task. So, it's very designer developer heavy, and not producer heavy, style. We did run into trouble and got too big, and Blizzard North, even after Diablo 2, got even bigger and ran into some real troubles with this style, but I think the style is key for my enjoyment and it worked out for our best game, so I think it's good, even though I wouldn't exactly advise it.

The Birth of the Skill Tree and the Development of Diablo II's Classes

The skill tree is now commonly known as an "RPG element," but it didn't become a genre fixture until Diablo II. As with many of Diablo's best ideas, it came to David Brevik in the shower.

David Brevik

One of the things I became infamous for is that I would often come in, and, I don't know about often, but every now and then I would come in and say, "So, I had this idea in the shower this morning." It was like, "Oh, Dave's shower idea has hit the game hard, because it's some major change to the game."

Erich Schaefer

At some point early on we went with the skill tree idea. We didn't start with that. That was a brainstorm by Brevik again. He was like, "Hey, let's make skill trees that are similar to tech trees in [Civilization II]," which I believe we were playing at the time. So that sort of set the pace. Then we started to think, "OK, what would be on the trees for these various characters? Should they have shared abilities like they did in Diablo 1, or should they have their own skills entirely?" I think one of the cool things, before I get to the specific classes, again, I think we kind of came up with this. I'm sure there's examples, but at least for ourselves, of, the warrior classes use spells just like the mages. So, before that, warrior classes in RPGs would just come in and hack on guys. Maybe they had some ability or something, but they didn't have a raft of skills they would commonly use like we ended up doing in Diablo 2.

David Brevik

I have no idea where I thought of that idea. I mean, again, I thought of it in the shower. But I don't know where it came from. It came from, I didn't like the way that we were, how we were going up levels and getting these skills, and it didn't feel like there was enough creativity or choice or things like that. We wanted to give people this sense of, "How do I choose how to play my character?" We had all of these skills, and it was kind of a mess, and there was all of these potential builds, how do we organize it in such a way that allows the people to easily identify?

One of the great things about Diablo 1, but one of the problems with Diablo 1, is that, from my hardcore nerdy perspective, you could make so many different builds, because everybody could do everything. But from kind of a general audience view, they were overwhelmed with all the possibilities. So we wanted to narrow that down into classes, yet still give a lot of flexibility within that class to kind of customize yourself and make you different from everybody else. And that's really the concept behind it. It was, "I'm going to make my character very different, even though I'm a Paladin and you're a Paladin, my Paladin plays very different from your Paladin, because of the choices I've made." That was really where the idea came from, and this was just a way to organize that idea.

Max Schaefer

I remember even on a fundamental level, kind of one of our goals with the character classes is that they would be slightly different than the stereotypical RPG character classes, but recognizable enough that you would know what they do. You should be able to look at the character and kind of get a sense of the way they play and what they do. That was a principle that another one of our key guys, Matt Householder, termed "familiar novelty," where you're seeing something new that you haven't seen in a game, but you understand what it is right away.

David Brevik

One of the things that he did that was really good for design for us was the way that he sort of set up the spreadsheets, because we used Excel to do all the data. And the way that we set up the balance, and the way that we did skills, I think [Stieg Hedlund] really brought that with us. So, we were able to put in more content with the way that he designed the way that we were going to do the data stuff for the levels and the difficulty, and things like that. That was definitely a big factor as well, because he helped to contribute in many other ways with design on all sorts of stuff, from monsters to whatever, all sorts of things. A lot of the radical ideas, more radical design ideas came from either myself or Erich Schaefer, I would say, the most.

Erich Schaefer

The classes themselves developed during play. I think it was largely Dave and I saying, "Hey, what would be fun to do with this guy?" and just cooking up skills on the fly; but a lot of times, most classes had advocates in the office, and people were big Paladin fans, or big Necromancer fans. They would just throw out ideas to do. The classes developed as we went based on the artists, Kelly Johnson making some of these characters. Just the way he moved, and the way that they sort of looked, kind of developed, "Hey, this is what this guy would do. Obviously he would have a shield slam." So, I think, again, a cool part of our iterative process is just like, "What would be fun to do with this character now? How can we go even more gonzo when he levels up, and gets even cooler things to do?"

I can't remember any sharp disagreements [over classes]. I remember at the end, when there was a lot of concern over balance, and I remember, I'm not going to name names, but, people would say, "This skill is way too good." And we would argue about the balance of skills. And I think we ended up patching the game many times, but there were some really bad balance problems, due to just kind of weird arguments at the end, and since there were only two or three or four of us who really had in depth knowledge of how to play these characters towards the end game, we made some weird decisions just based on personalities. I don't remember, I think everybody was pretty much on board with the looks and the feels of the characters as we went along, though.

Diablo II's Skill Tree

Max Schaefer

I remember, I think we did spend more time going back and forth, I think, on the look of the Sorceress. That was championed mostly by Mike Dashow. It ended up wonderful, but I do remember going back and forth on that. Paladin looked great from the get-go. The Barbarian too. Necromancer was such a weird class, nobody had any expectations or really strong feelings on what he should look like. And, what was the other one, the Amazon kind of did herself as well.

You start with concept sketches, and then first in-game models, and, there's always a little bit of, "Hey, that didn't look like it did in the sketch. Why is that?" "Well, our camera angle and the scale makes it such that… That type of clothing doesn't look right anymore… Or, the way they move or walk looks a little bit weird." And normally, you put it up on screen, and everyone agrees, "Hey, that doesn't look right." "OK, we'll go back and fix it." And that's just the standard way that things work. I know we worked on the Paladin's animations for a while, because his look didn't look quite right, but his walk was always spot on. I maybe even have an incorrect memory about this, but I remember we went back and forth about the Sorceress quite a bit before it was nailed down.

Erich Schaefer

Yeah, and I kind of remember a lot of joking that it looked like the Necromancer was wearing a skirt. We ended up leaving it, because enough of us liked it, but that was, not necessarily a contentious point. But it was pretty funny.

David Brevik

It's amazing how many new standard things came out of Diablo and Diablo II. It surprises me all the time. The rarity thing [for loot], for example, just kind of made sense. In some roguelikes, they would have your common item and your magic item, so they were different colors of text or whatever. If it was a magic item it was blue, if it was a normal item it was white, that kind of thing. And, so, we took a step further and went with the rarity levels. The rarer something is then it has a different color. That really has stuck with games, and then they took it to a whole new level with World of Warcraft, and it really became standardized roleplaying stuff ever since then.

This article may contain links to online retail stores. If you click on one and buy the product we may receive a small commission. For more information, go here.

Comments 38

Comments

Close