The designers of Dishonored, Bioshock 2 and Deus Ex swap stories about making PC's most complex games

PC Gamer: I'd like to dig in a little bit to the process of creating and designing an immersive sim. I think most people who pay attention to videogames know that it's not a linear creative process where you build from the beginning of the story to the end. A game is usually not really playable or fun or 'complete' until very late in development. In the case of immersive sims you're potentially talking about a dozen systems, from AI to weather to hacking, or combat, all these incredibly complex things. How do you go about building these and testing them when it's not fun, or when certain things aren't online, when it's only half complete? What's the process of choosing those elements and testing how they work together?

Steve: That is a really interesting question for the guys here who have worked on the big titles. I've worked on, basically, sequels to immersive sims. I worked on Bioshock 2 and Infinite, but I didn't work on the original Bioshock, and obviously that's kind of a continuation of System Shock 2. So I'm interested to know, when you are building a game that is based on this bedrock of multiple strata of systems, do you try to block in as much of the different player abilities and AI systems as you can as early as possible, or is it an ongoing glazing of 'what if we added this, what if we added this' over a long period of time.

Ricardo: There's a lot we could say there. Some of it is what you're saying, Steve. I think we try to get a 60 percent version of as much as possible in, as quickly as possible. Because part of the fun, of course, is not just the thing existing in isolation, but when it interacts with all the other systems. People have likened it, a little bit, to making a stew. Individual elements aren't that great, together they're okay, but they kind of have to live together in the pot for awhile so that you can begin identifying, like 'this one mechanic doesn't contribute very much. This other one, though, we should double down on.'

By the end, maybe we make 25 percent more than necessary, mechanics, that end up getting stripped out, and we focus on the ones that end up being really successful in the whole mix altogether.

Steve: It feels like it's inherent to this kind of game that for it to actually be the game at all, there's this critical mass that's required. You can't work on Dishonored for six months and only have two player powers, because it's just not relevant to what it's going to end up being. But also you obviously don't just write your perfect design bible and you're like, 'here's the dozen powers and exactly what all the enemies can do' and just make it. Finding that balance must be really challenging.

Warren: Design documents are always right. [laughs] There was one point on Deus Ex where the documentation was 500 pages, but we're not going to talk about that. It was ridiculous. The final version was 270 pages that nobody read. Anyway, the interesting thing about making this kind of game is that you guys are all right, until those systems are online you don't even know what you have. Alpha is the point on a game like this where the game is complete and finishable and playable and sucks. 

To make this all work—the money guys love this, while you're working on it the first two years or whatever it is, the game is not there, it's not there, it's not there, and everybody's going 'oh my god,' biting their fingernails down to the nub, because they're giving you all this money and they can't see the game yet. You have to go and say 'relax, it'll be okay, everything will come together.' And then you hope they'll give you enough time in alpha, at least this is my take on it, they give you enough time in alpha to make it right.

On Deus Ex we implemented the skill system that a couple of us came up with pretty early on. I think we actually got to alpha with that in there. And then we invited guys like Doug Church and Mark LeBlanc and Rob Fermier, and even I think Gabe Newell came down, and they played it, and they said 'wow, this skill system really sucks.' And I think it was like 24 hours later, Harvey, you had a completely redesigned skill system. Thank god! Because it wasn't at all what we thought we were going to make. Until we played it, and saw it in context with all the other systems, you're just taking your best guess.

Harvey: I wish Raphael Colantonio was on the line with us today, he's traveling to GDC. But one of the things he talks about a lot is how much iteration games like this actually require, and how flexible a studio has to be, and how you have to train the team not to think like traditional developers. You have to be willing to react very quickly. So much of this kind of game is a synergy, and the magic only happens very late. And I can tell you from experience, sometimes they don't give you the money to finish it, to get that final three months or whatever. But in almost all the best cases of these types of games, the ones we've worked on here and the ones that friends have worked on, Deus Ex, Dishonored, Bioshock, you hear these stories about how things almost came together at the end but then we got three more months or six more months and then we just started hitting it with the magic in place.

By contrast, you have developers who say 'on day one, you need a loop, and if that loop is fun, you just iterate it and your game will be fun. If your game is not fun on day one, your day will never be fun.'

Warren: I FIGHT THAT EVERY DAY!

Harvey: Like it's some sort of dogmatic blueprint approach. Well, for some games that is true if they entirely depend on one arcade game loop, then yeah, probably. But these games are something different. They're a sense of presence, they're exploration, they're player pacing, they're toying with systems, and they really rely on this gestalt.

Steve: I think there's something interesting you're saying about that last three or six months that can also be extended to things like DLC and direct sequels. I worked on Bioshock 2 which was a direct sequel to Bioshock 1, and I was the lead designer on the story DLC for Bioshock 2, and at that point as a developer you're kind of in that space of saying this has been developed to such a degree that you have the familiarity with it, you have the stable base to say 'now our job is to know this stuff well enough to do something really good and really interesting with it' that you don't find in that initial build-out. And that's kind of what you're doing in those extra three or six months. You spent all that time making the game. Now we know what it is. And we can actually use this time to express what we've learned that we might not have been able to otherwise.

Ricardo, you worked on the Dishonored 1 DLC, right?

Ricardo: Absolutely, the Knife of Dunwall and Brigmore Witches stuff.

Steve: I think that's sort of an extension of that idea that these games are such a, the end product is greater than the sum of its parts. Having that ability as a designer to work within that established space and do things with it that I wouldn't have thought of or known how to do earlier in the process is especially relevant when you're making games of this complexity and this relation to the player's role.

Tom: There's also a huge technical benefit of being that late in the project. Particularly with these systems-driven games, you have to build the systems, and then once you've done that, making a new ability or a new item is actually almost trivial. You hook it into the systems that already exist. And the whole point of these games is those systems have to be consistent, have to be universal, so you have to get that right first, anyway. And once you've done that, making an ability that uses those systems is super easy.

I've just hit this point in Heat Signature, so I'm really excited about it. Now if I want to make a gun that hacks things when you shoot it, it's literally create a gun, add the hack damage type to it, and it's just done, it just works.

Steve: That even happens in a story game like ours that is not about these deep, dynamic, interactive systems. As a content creator, making these kinds of games the arc is really on a logarithmic scale. It's ramping and then you hit that tipping point where now, working on Tacoma, I have enough of the tools to say 'oh my god this room is so empty, it needs stuff in it,' and then you work on it for a day, and you're like 'oh, we've built enough that we can make this feel very populated and unique and like a real place very quickly' in a way you couldn't have earlier.

Or you can say 'I know how our AR character system works, I can extend this scene that we already have to do something else' because we've been building up those tools over time. Once you have the toolbox, which takes a long time to get to, and you have the familiarity with what all of those tools can actually do, that ability to quickly and very creatively extend what you already have into things that feel very unique and memorable to the player finally appears.

Ricardo: What you're citing is one of the reasons I actually love being finished with the main game and getting the chance to work on DLC. You have that baseline there that you can build on top of, and it's so easy to add things. In the Dishonored DLC it was really fun to get to experiment with Corvo's base powers to make new powers for Dowd, the main character in the DLC. Like adding the ability for when Dowd targets his Blink power, the whole game is frozen. So it's more like a tactical, thoughtful consideration where you're going to Blink. That was only possible because the main game, all that stuff was already executed and established, and we could play in that sandbox.

Warren: You do need to be thinking about player improvisation early, though. We did build those proto-missions [on Deus Ex]. That's what I called them, I can't remember if anybody else did. We built that White House mission where everything was sort of hacked together, which didn't show how the game was going to play but showed the potential of it.

I can't really talk much about System Shock 3, but I will say that we're just beginning to prototype a bunch of stuff, and if you think about giving players the ability to improv early, you can start to see the fruits of that early. 

We built one thing out—I should not be talking about this—where there are a couple ways to get past a problem. But I found one that no one knew was going to work. Instead of taking five, six minutes to play through this space, I did it in 10. Ten seconds. It was pretty magical when I figured out something that no one on the team knew was going to work, even early. And we've got another system that I'm not going to talk about that we've started prototyping. And already we're starting to see people use it, family and friends testers, they're starting to do things with it that we had no idea would work. When you start seeing that, even early on, that's the magic of these games. It's what makes them different.

If everybody on the development team knows what every player is going to do, my advice to them is just go make a movie.

PC Gamer: We've got a few minutes left before we need to head off to other GDC events. Does anyone have a question they'd like to ask anyone here, before we have one closing question?

Warren: Yeah, could you guys stop working on Prey? [laughs]

Ricardo: We are about to start working on it! [laughs]

Warren: I can't wait to play it.

Steve: Yeah, I'm super excited about it. I guess there's three of us in the room that are making space station games, but you guys get to ship yours first, congrats. I assume you guys haven't announced a release date? God dammit, I hope you guys don't ship the same time as us.

Tom: Is that your question? When's your release date?

Steve: Yeah, can you announce your release date, please?

Ricardo: Oh, it's May 5.

Warren: It's on the trailer, it's in there.

Steve: Oh, that has been announced? Fine, you guys get to be first!

Harvey: Honestly, I say this with all humility because I didn't work on Prey, it's a game made by the Austin studio with Ricardo and Raph and Seth and Susan running the show, it is one of the best games I've ever played. I tweeted something recently about having finished Dishonored 2, and now looking at Prey back to back, Brian Eno had this write-up, and it really made me stop and think about me as creator vs. me as player, what I like to play, and it's really an interesting contrast. In my career I don't think I've ever had the opportunity to do this back-to-back within the same studio. We finished one game and are about to finish another game with a different team, and just to look at the two and decisions they made differently than decisions I would've made, yet I love both games. I think Dishonored 2 is the best game I've ever done and Prey is one of the best games I've ever played, but they're very different in their decisions and how they arrive at certain decisions is fascinating to me. So May 5, yeah.

PC Gamer: So I'd like to close with a question, looking beyond Prey, where immersive sims are going to be going in the next few years, the far-off future of 2020, what do we still have to improve in immersive sims? What have they not quite cracked yet? Is it AI? Is it elements of level design, maybe moving beyond the conveniently human-sized vents placed on the backs of buildings?

Tom: Never! [laughs]

PC Gamer: Where are we going next?

Warren: Non-combat AI is an area where games in general really have some work to do. In the more linear cinematic games that we're not talking about today, I think there are some pretty amazing things going on. But in terms of characters who can react to you, whether they hate you or love you or are neutral towards you, we still have a lot of work to do on that front. I would say non-combat AI is one, and accessibility is another.

We talked about that a bunch, but making this so normal humans, non-gamers, can actually get to this, so we're not just making cult classics. We're making mainstream games that show the world what games can and should be. Accessibility is a big problem for us.

Tom: Coming at this from the indie side of things, I'm excited about stuff to do with the structure and format around the actual missions that you do. In a triple-A immersive sim it's almost a story that is told from beginning to end and there's maybe some branching, but you're playing as one character throughout the whole thing. I think that problem I was talking about earlier with people feeling obliged to stick to one playstyle even when they're not enjoying it stems from that.

I'm trying a game where each character you play as is a new life. Every time you die it's permadeath, but each time you restart you're a new person. I'm suddenly finding loads of immersive sim problems just go away if you just change the structure completely. It's a completely different format of game. This is just a baby step towards it with Heat Signature, but I'm excited to see what other people do with that, as roguelikes are a big trend with indie games. I just want to see that mashed into immersive sims in as many ways as possible because I think there are some really interesting things that happen there.

The thing about playstyles, I just have a missions listing board, and I've just realized recently, I've added the ability to have missions that you have to do stealthily. You'll fail this mission if you get spotted. And so you can just work playstyles into the mission listing board, then let people pick which one they want to do, and it's just natural that they would vary if they wanted.

Steve: From the other side of indie development, something that I think is really fascinating and valuable is immersive sims as a lineage is it's a very long lineage that has kind of continued to accrete properties over time. Warren, you've been there for the entire run—

Warren: Thanks for reminding me of my age, I appreciate it. [laughs]

Steve: Well but you were working on Ultima Underworld and System Shock 1, and our approach to exploring that, as a small independent developer, is rewinding the timeline and removing factors and thinking of it in terms of, if we went back to an earlier point in what these games are and explored a branch from there, and tried to find aspects of that experience that are inherent to it, but have not been the focus in a lot of ways...

I guess what I'm saying, to your question Wes, that there's one way of looking at this as: 'Where do what immersive sims have become go next, and how do we solve more problems and add more on?' And I think there's this incredible potential to saying: 'Well, but what have they been, and what was not on the main trunk of where they've gotten to, and what else is there?'

You think of an immersive sim now and you think of things like upgradeable player powers, AIs that have emergent abilities when they interact with each other, and having an economy so you can buy equipment and all that stuff. When I was working on Minerva's Den, and it was a reference forward design wise, and then I was working on Gone Home, I replayed System Shock 1, and it was sort of a surprise to me to realize there is no skill tree. There is no economy. This is about a place, and you as a character with a role in it. It has enemies in it, different ways you can address problems, but there's so much that we think of as being part of what an immersive sim is that is really just the version of it that we've arrived at.

Being able to say 'System Shock 1 is an immersive sim because it has a sense of place and it has you being able to fulfill a role within that space,' and so a game like Gone Home is kind of an exploration of how we apply that to a mundane setting. How we apply that to a space that's more familiar to you. How do we apply that to something where finding the audio diaries is the actual game, not just a thing you do while playing the game. Continuing to explore what else is already inside immersive sims is a really exciting thing to be able to do.

Warren: In some sense, actually, System Shock is actually the purest expression of what an immersive sim can and should be. All the character stats, upgradeable this and economy that, all that stuff you were talking about, it kind of turns things into a hybrid RPG-immersive sim thing that I love, I absolutely adore that kind of game, but in a sense if you're talking about the absolute purest form of the genre, for me it's going to be System Shock.

Ricardo: I don't have anything specific to say, other than the thing that's exciting to me is to see, Steve was talking about different ways it's affected other games, and I'm really interested in the family tree, or the lineage, of immersive sims. Seeing how that bleeds into other games. There's sort of a core essence to immersive sims. But I love it when people experiment with that. I think a game that draws from immersive sims doesn't have to be first person, for instance. There are some 2D games that sort of have that same fundamental philosophy of strong sense of place, plus very expressive interconnected game mechanics, that have come from the developers being fans of immersive sims, and that's why they made the game that way.

I love seeing more expressions of that sort of development philosophy in other genres and independent games. I don't have a super recent example, but I really loved Mark of the Ninja.

PC Gamer: Great game, yeah, from Klei.

Ricardo: That's a 2D game, or sidescroller, but just the way that you play that game, it's clearly founded on similar principles. Just the open-ended nature of the game mechanics. They're super fun. Like Steve was saying, the Gone Home and Tacoma-like games, they're more stripped down than the giant triple-A action immersive sim. But they're an interesting offshoot. I look forward to seeing more things like that. Offshoots that come from that lineage.

PC Gamer: Anything from Harvey?

Harvey: Yeah, in part, I would echo what other people have said, but I've been thinking about it a lot lately. And the thought that comes up over and over is purity. Right now we have a lot of stuff that we've accreted that we put in through legacy or because commercial audiences deserve a game of a certain size in order to pay a certain amount for it. Just to do one of these games, with AI for instance, with physics interaction, with the scope of the game, the development budget is pretty large. There are all these forces that pressure you to go one way or the other. You either go full-on where it's got tons of stuff in it and layers and layers and lots of different ways for the player to switch playstyles as they're going, or you go the other way and strip it all down to the bare essence and find something interesting. Whether it's the setting or a particular form of interaction, a particular tool.

One of those works better in the commercial space and one works better in the indie space. For my money, I would love to have the opportunity to just play around with, what is the minimum, here? I appreciate the hell out of games like Gone Home, of course, because it was innovative and revolutionary in terms of subject matter and the feel as you play the game, it was one of my favorite games that year. But our games are so big in terms of economy and sheer scope that I would love to make a more stripped-down game, but I wouldn't want to sacrifice that magic moment that happens when you manage to get a turret up on the roof and hack it to your alliance, and then somebody you weren't expecting comes around the corner and your turret opens up on them, but you happen to be in the line of fire. A whole sequence of crazy improv events happen that you have to react to. 

Some of the stuff is not just accreted baggage. It's where the actual synergistic gameplay comes from. Where it lives. And so it's thinking about how much do you need, and which do you need, that's not just painting by numbers. You know, 'oh it's an immersive sim, let's make the first code 0451 and add a crafting system or whatever.' The future is bright for deeply interactive games with a sense of presence.

Tom: The other thing I'm excited for for the future of immersive sims: I hope we come up with a better name. [laughs]

PC Gamer: I was actually going to ask, I don't know if it's common knowledge where the term came from. Was it a Kieron Gillen-coined term, or if it predated his writing on the genre.

Warren: I think Doug Church was the one who came up with that, isn't he? He's the first person I ever heard use it.

Harvey: I don't know, I remember a conversation with Rob Fermier, I think on Twitter, where we were trying to figure out where that term had come from. I think Rob's conclusion was that he first heard it from Doug, as well. 

Warren: Yeah, and we all hated it! It fell out of favor for awhile and recently it seems like it's come back. It's really odd.

PC Gamer: Has anyone come up with a description they like better?

Warren: No.

Ricardo: I don't know if it's better, but when we're talking generally with the press or gamers we avoid the term, because it sounds very inside baseball. We say first-person games with depth, instead, and then elaborate from there. But it is a bit technical.

Harvey: I like FPS-RPG hybrid.

Ricardo: Yeah, that works too.

Warren: Genre mash-up, yeah!

PC Gamer: So maybe by 2020 we'll have decided on a new name for the immersive sim. Well, I wish we could keep doing this. I could literally do this all day, makes my job easy. It's been a pleasure. Thanks for joining me! We should do it again. Maybe we can make it an annual immersive sim roundtable.

Warren: Sounds great. I'll have something to talk about next year.