Gamasutra staffers recently had the pleasure of talking to Harvey Smith of Arkane Studios while livestreaming Dishonored: Death of the Outsider, the well-received standalone followup to Dishonored 2. Smith's comments about the design of various features in the new game, and in the franchise more generally, were fascinating. So fascinating that we decided to transcribe portions of the stream.

Read on to learn about fascinating features that were cut for the game, as well as Arkane's approach to playtesting, whiteboarding, and designing powers.

And for more developer insights, editor roundtables and gameplay commentary, be sure to follow the Gamasutra Twitch channel.

STREAM PARTICIPANTS:

Harvey Smith, creative director of Arkane Studios

Bryant Francis, contributing editor at Gamasutra

Alex Wawro, editor at Gamasutra

On designing Semblance and the other powers in the Dishonored series

Wawro: There's a power in there called Semblance that allows a player to adopt the visage of an enemy, and that seems like a huge deal for me, as someone who's played a lot of these games, and is constantly trying to stay out of sight, that you can do social camouflage. It seems like it has a lot of interesting ramifications for how you design levels. Is that on point? Can you talk for a bit about that came to be implemented and what challenges it brought?

Smith: Any time someone asks about Semblance, I have to say a couple of things. One, the way it came together was amazing. Earlier Jerome Brown, one of our senior games systems designers, thought about how many pieces had to come together for Semblance to be as cool as it is, from the effects, to animation, object design, narrative design, level design, the programming overall of it, the sound as you take someone's face, it's all in symphony. And the other two things, I have to say it's the power I was most dubious about.

Early on the game design team pitched a couple of ideas, and they always come from the team in different pieces, and even if you pitch something, someone else makes it better by suggesting something else, some constraint or whatever. We were kicking around the powers, we were about to start prototyping, and people really wanted to do this identity-stealing power. And I was so dubious.

We've seen Hitman, you can take hostages, Corvo possesses people and you can walk past their allies. We've seen variations of this off and on, and it has these problems. Invariably there's some situation where you take someone's face and you walk in and talk to their husband or wife or best friend, and they don't quite respond right, and it lets you peek behind the curtain. The illusion is broken.

But the team was so convinced, they were so passionate about it, kept working on it and adding constraints and rules. The level and narrative designers signed up for the challenge, and they just did a really good job of supporting it, to such a degree that it surprises me.

I was telling this story the other day, there's a level later where you revisit the Royal Conservatory. Months have passed since Dishonored 2, the lighting is different, a different faction is living there, it's been updated in architecture. But you go back to this known location in Karnaca, and a major scene plays out at some point between two of the major characters, and one guy is working at the vivisection table, and he keeps working and keeps muttering and talking, and she comes back to the door, and she's not going to wait much longer he'd better get a move on, and he says yes yes, I'll be there in a moment. I thought, what if I took him out, take his face and went to the intended meeting with her, if she'll respond?

And so he did, he cleaned up and started to head out, and so I choked him out and took his face and walked in and talked to Sister Rosalind, and it all worked. And I was like, wow, people went further on this feature than I thought we were going to. That's the first thing that comes to mind, this is a feature that people on the team really believed in and went the extra kilometer to make it work. We keep getting feedback that it's one of the features that people are really having fun with.

What we tried to do with Billie, early conversations we had with Dinka, we really tried to make this less of a game where you could stab people or throw fireballs. Instead you have to get in someone's face. You have to physically cross the space, or get behind them, or get close enough to eavesdrop on them or listen to them, it's very visceral.

And so powers like foresight, it's my favorite vision power that we've ever done in a Dishonored game, because it's so much more active, in a way? It doesn't let you cheat. You stop time, you kind of astrally project, you see Billie leaving her body behind, you ghost around and scout where the objects are and where the guards are at, then you snap back to your body and it picks up from where it stops. Anything could happen at that point. The guards could actually turn from where he was going, but you have a glimpse of what he thinks is going to happen.

That's why it's called Foresight. It's just an example of a very active, visceral power. The same is true of Semblance, we just tried to err on the side of making you play the game. Powers aren't so much a "smart bomb" in Billie's case, they're alternate ways to play the game.

How Arkane handles playtesting

Smith: Arkane believes very much in iteration and playtests, so, constantly, the producers are organizing playtests where we're bringing in people, we're watching them play, very often they're streaming to every single developer in the building's monitors. So they'll be working through the day while they have this going on on the right. We have guys like Christophe Carrier, level design director, who will later go get all the videos from the playtests, in our labs, and watch every single video and make notes.

That's above and beyond, I don't do that. I used to do that, I mostly get an aggregate sense of what everyone is feeling now. But Christophe still does that, and it's really useful. Where people are stuck, where they're lost. When players are lost let's not do the most obvious thing, let's don't just put a marker to whatever. Let's try to find a little thing that guides them in the right direction, or that draws their eye. Then let's do another round of testing, and if people are still not getting it, then let's go stronger. It's part of our values to try to give you the experience, to let you discover it, to let you explore it.

When Arkane commits to ambitious features, and when it decides to kill them

Francis: We've talked about making these kinds of games, about making these intricate features that take a lot of resources that wind up defining the games. You've talked about being scared of these features, how there's something really intense that you know is going to take up a lot of resources but could be totally awesome and really hard. Can you walk us through, where's the point where you realize it's worth over-committing to this path, if you will, sand out those edge cases, to build that extra logic, to create the scene like the one you described earlier regarding Semblance. Walk us through that process through which a feature becomes really killer, or you have to walk away from it and kill your baby.

Smith: So this is a good example, this guy's painting here in the stream, and you get the sense that he works around the corner. If you just walk around the corner there are some dogs that are hostile to you. But as long as you're him, while you're using his identity, the dogs shouldn't react to you. But that's a case where Foresight would have helped you, could have zipped around the corner. Billie also has a talisman called Rat Whispers, where you can listen to what the rats think, they've seen things.

It's just part of our DNA, based on the kinds of games that we love, that we have faith that if we put all the detail, all of these alternate ways to do things, all these little improvisational tools into the game, that some percentage of your players will find them, that they'll have a "magic moment," and that's constantly what we're going for.

By "magic moment," I don't mean that we scripted the perfect moment here whatever, I mean like we get some unpredictable point for us along the way, and different for every player, there will be some moment at where you use the power at the right way that you narrowly avoided some situation that felt very dynamic, that felt systems-driven. A guard happened not to notice you because of a sound you made earlier, he turned in a different direction. That's really why we play games, looking for those moments, and so it's that faith, I think, that carries the whole team at Arkane along, knowing that that's the kind of experience that we're providing people.

The downside to it is, it really really works on certain types of players, and it really doesn't work on other types of players. And so there are people who are just like, sprint through the game, see a tenth of it, kinda don't get the best parts of it, but maybe they still have fun by blasting things, drop-attacking people and all that. We try to widen the game as much as possible, but clearly it's made for people that like a particular kind of pace, and a particular kind of exploration, and a particular kind of narrative.

Francis: Are you able to tell us when in development you think a feature like that isn't working out? We're lucky enough to see these features make it all the way through, but I'm sure throughout Dishonored's design history there's been powers that just never quite made the cut.

Smith: Yeah. And you feel guilty, as team lead, when you push an idea too far, you feel really guilty for making people work on it. We have examples of that, we have examples that we killed very very early in conception, what (Arkane founder Raphaël Colantonio) used to call, I don't know if it was an Italian colloquialism, "Killing it in the egg?" Where we'd kick some idea around, and somebody, like our Tech Director, would go "Eh, I don't think we should do that, because of X, Y and Z."

In Dishonored 2 there was a power in which we were going to let you walk on the walls and ceiling, like invert gravity? We were warned away from that very early, so we modified it heavily. It eventually became the power of Shadow Walk, which we were very happy with.

There was another example in Dishonored 2, Emily had a power called Void House, which just meant that across all the levels, in its initial conception, across all the levels we were going to hide two or three portals to the Void, that would take her to a special place where there was a house with a giant tree growing in it in the Void. The whole place was reflective of the current chaos state, and lots of little things she had done along the way would be, like you know, maybe somewhere there's a giant mound of bodies floating on a little island floating out nearby, for all the people that she had killed. There are all these ideas that we kicked around. Like a little library, a table full of food, and we carried that idea along, but finally just cut it. It was very painful to cut, but it was also the right call.

You just develop this sense of, we're running out of time, everything's too compressed, this idea is further along than that idea, this one we don't know exactly how we're going to solve it, whether it's going to be cool or not.

Deciding to cut features, and Harvey Smith and Raphaël Colantonio's whiteboarding process

Wawro: I hate to keep asking for examples of failure, but we have a lot of students and game developers who watch streams like this and watch the videos, and it's fantastic to hear about things that are risky, like the Clockwork Mansion, that pay off. Were there any things in the production of Dishonored 2 or on Death of the Outsider that you guys really wanted to stick to, but had to end up cutting at the last second?

Smith: Oh yeah, totally. We had a Dishonored 2 level called the Wind Corridor where, at the peak of Shindaerey Peak, the cleft in it, powerful winds would blow through, and we had turbines, more windmills and all that. There was a prototype where, once the alarms started sounding, these huge alarms, the winds would start building up, and all the workers would go inside and shut the doors, and anybody caught outside would start to struggle against the wind and eventually be blown away. And it was pretty cool. It was big, epic and over-the-top, but we ended up cutting it.

But we preserved things like the windmill technology driving things like the Wall of Light, and powering security devices. In Death of the Outsider there's a new security device called the Safeguard Floor, that we pitched at least in Dishonored 2 but we didn't have time for, so we really wanted to give it its due, so we put it in here, and spent a lot more time prototyping it. But that's one example of a level that we cut fairly late, that was going to be another of our big, signature levels, where you'd move through this space, and periodically these alarms would sound, and the wind would kick up. And you could do things with the wind. Like, jumps would carry you a much greater distance, because of wind power. That's an example of where it just didn't work out.

Another example is, initially we planned many different physics puzzles, where we had, like, these sharks hanging on meathooks. You could put a spring razor on the shark and push the shark down the line and, as it got closer to people, it would trigger. It would cut up the shark's body and anybody that it bumped into. And we still do have some physics interactions in Dishonored 2 and Death of the Outsider, but not as many as we initially planned, just because, sometimes the way you play, observing, strategizing, pouncing like a lion, sneaking up on a person and pouncing, is a bit at-odds with some of the physics interactions.

If your game is really about physics interactions, like Garry's Mod or whatever, then it fits in better. You're constantly massaging the whole, and even things that you love, that have a high degree of excitement for, sometimes you have to be disciplined, and let it go.

You should constantly be reevaluating the aggregate of your game. One of the things Raph and I did, that we tell students when we talk to them, is that Raph and I had this exercise that we did, I would say bi-weekly, like every other week we did it, where we'd just say, okay, here's the game. And we'd go to a whiteboard (we shared an office at the time), we'd go to a whiteboard, the office had a window looking out on the rest of the open floor space, and we'd use the glass and some whiteboard markers, or we'd use the whiteboard, and we'd just go, over here we have the A.I. stuff, like sneaking, like "What was that? It was just a rat."

Here we'd have the combat suite, first-person melee, here we've have the inventory items, the gadgets, the spring razor, the crossbow, the telescope, here we have the bonecharms, later in Dishonored 2 that became bonecharm crafting, and we'd think about, is there enough stuff in this area? How often am I going to switch from this activity to this activity and back? And then something will pull me over to this activity? How important is the narrative layer?

You constantly look at the whole game as an experience, you try to write out what it would be like as an emotional experience, how tedious it was versus how dynamic it was, is it too overwhelming by contrast? On a whiteboard. And we just do it over and over. So sometimes we'd kill entire areas. And actually I think that's where bonecharms came from, in that we, from doing that exercise, at some point the game was a little too pure? It was missing that one little extra detail where you could customize your character, or find some other little trinket in the world that would modify your mechanics a little bit, that would give you that one extra little thing to play with.

Every time we'd look at it and look at it and look at it, we'd do that with the creatures, we'd do that with the powers, etc., and I think based on that, that's probably how the bonecharms were added, I'm pretty sure. And probably in an ancillary conversation, or a corollary conversation of that, was probably where we added all the non-lethals for all the major targets, it sprang from that as well. Just looking at your game as an overall experience, what all it includes and what it doesn't include.

Wawro: What I think is so interesting about Dishonored, as a series, is that it leaves a lot of room for the player to express themselves through play. And one of the key ways you can do that is by choosing not to kill the targets that you are sent to kill. The notion that that just came out of a whiteboard sketch one week is striking.

Smith: Many things along the way I don't recall how they happened, but that one is very specific, because I remember Raph and I were in this meeting, and we shared an office at that point, and we were just sitting there shooting the shit back and forth about the game, anxious as usual, wanting people to love it.

But it was pretty early on in development. So, at this point, programmers were working with game designers on features that we're playtesting, level designers are all working on their levels that we're playtesting, and we're going back and forth on what our reactions to those things are, and how we should fight for cuts around, or additions to.

And at that moment, I remember us talking: you know it's true, that because of Bend Time, and Blink, and also just the stealth model, wait until the guy looks the other way, you can literally not kill any of the guards in the game, the civilians run from you, so it's literally true that, currently, you only have to kill the main targets. And because we only have like nine main targets, I said, "I bet we could come up with an alternate resolution for each of those." And I remember Raph saying, "That would be cool, but it would be really hard."

I was like, "Let's take an hour, let's bull it out a bunch, we'll list the main targets, we'll bull out some kind of idea for an alternate resolution, we'll pitch it to the team." And so we went away, Lady Boyle, Lord Regent, Overseer Campbell, etc., and out to the side of each one we wrote something. Like, "Find evidence that gets them convicted," "Put him in a create with a bunch of food and ship him off to a far island," we wrote down a bunch of things like that. And then we presented it to the team, and of course the team was immediately able to come up with better ideas for about half of them, like someone pitched the Heretic's Brand for the Overseers, like, if anyone is ever branded with this, you're not even allowed to talk to them, this symbol. That was a brilliant idea from one of the level designers, I think.

So pretty quickly we realized, if we just did the extra level design work and artwork and everything else to support things like the Heretic's Brand, or reading out the confession of the Lord Regent over the P.A., so everybody hears it and he gets arrested, that is some work of course. But if you just did nine moments across the levels like that, then suddenly you don't have to kill anyone in the game, and that's an interesting talking point, and it's interesting for players who want to go one way or the other, to be able to experience those, and so pretty quickly, that was one of those flashes of inspiration where, and hour of work, the team made it better, then, of course, the team had to go do, like, how many man-months of work that was to implement all of that, and we had to constantly playtest and iterate on it and comment on it along the way. But it's a good example, it's like, it's such an iterative process here.

For more developer insights, editor roundtables and gameplay commentary, be sure to follow the Gamasutra Twitch channel.