Chapter 23
Chapter Select

Pause Screen: In the Guts of Game Design with Josh Sawyer

Obsidian's game director talks his background in tabletop RPGs, the computer game that opened his eyes to electronic entertainment, and game-design conceits that often confuse players.

8

Like journals sprinkled throughout RPGs, Pause Screens go into more detail on the people, companies, and cultures that contributed to the success of the Pillars of Eternity franchise and the Infinity Engine line of roleplaying games.

Josh Sawyer's roots go deep.

Entering the industry as a web designer, Sawyer got his feet wet—and cold—designing snowy settings and dungeons for Icewind Dale before captaining the team assigned to Icewind Dale II. As game director at Obsidian Entertainment, Sawyer has spearheaded the design of beloved computer RPGs from Fallout: New Vegas to Pillars of Eternity II: Deadfire.

Sawyer’s tool of choice is not Photoshop or a C++ compiler, but a set of dice. Tabletop RPGs were his first love, and the trappings of games like Dungeons & Dragons can be seen throughout his work. There's a good reason for that. The Dungeons & Dragons license made possible the likes of Baldur's Gate and Icewind Dale, two poster games of the Infinity Engine era.

I spoke with Sawyer about his foundation in pen-and-paper gaming, how he hopes the RPG genre moves forward, game systems that tend to confuse players, and other design topics.


David Craddock: I think one of the reasons you've been so successful at making the Pillars-style of RPG is your familiarity with tabletop RPGs. How did you develop that interest?

Sawyer frequently plays in and DMs tabletop games at Obsidian. (Photo courtesy of Sawyer's Twitter page.)
Sawyer frequently plays in and DMs tabletop games at Obsidian. (Photo courtesy of Sawyer's Twitter page.)

Josh Sawyer: I've loved playing games since I was a kid. I first got into tabletop roleplaying games with the red-book Basic Dungeons & Dragons with a couple of friends. Actually, at first it was just one friend; it was kind of hard to play Dungeons & Dragons with one person being the dungeon master and the other being the player, but we did it. After that, we played blue book, or Expert D&D. We had a few friends we played with, so we had a proper group, although we were all pretty young.

I met a kid at a public library who was playing a game on a Commodore 64, and I was just blown away by it. The graphics looked so cool, and the sound effects were so awesome. I was used to playing games on an Apple IIe, so the Commodore 64 looked amazing to me. The game he was playing was Bard's Tale I, and I just thought it was so cool. This older guy—he was older than me; he was in high school—his name was Tony. Tony introduced me to his friends, and they would get together at the library and would play Advanced Dungeons & Dragons, and I got into that. It went on from there. I played a lot of computer roleplaying games and tabletop roleplaying games.

David Craddock: When did you start experimenting with game design?

Josh Sawyer: I think it was in high school when I started modifying games, modifying the rules. I was playing D&D so much that I was running into what I felt were friction points: "Well, these rules are kind of obnoxious and none of us seem to enjoy using them, so I'm going to start changing them." Once I started changing them, I realized I could do whatever I wanted to do. I felt liberated. I would look at a rule, and whenever I thought there was something that I, or more importantly, the groups I played with, thought would be enjoyable, I'd just modify it. In college, I started designing my own tabletop roleplaying systems. Luckily I had a pretty big group of people to play with and get feedback from.

I didn't really expect that I would get into making video games, even though I'd been playing them throughout this whole period. This might sound kind of strange, but because I grew up in Wisconsin, and at the time, TSR was there, I had some idea of what it was like to get into the tabletop industry. But I really had no idea about getting into video games. I had this understanding that there were companies in California that did that, but otherwise had no idea how one got into that stuff.

David Craddock: Since you were just starting out, did you have any goals as a designer? Or were you just tinkering?

Josh Sawyer: I wasn't trying to recreate D&D. I did continue playing D&D, but there were other things I wanted to do based on ideas I'd seen in other games. I looked at a lot more tabletop roleplaying games, whether it was things like GURPS, or Stormbringer, or all sorts of cool, more niche RPGs from the '80s and the '90s. A lot of my focus—and I don't think this is an uncommon tendency among young designers—was, "I'm going to make a more realistic RPG. I'm going to make something that's gritty." That was a lot of my focus: Making things that I thought were realistic or natural.

It was a lot of fun. I got a ton of feedback, and I iterated on it a bunch. I also had a classmate, Jeremy Strandberg, who was working on his own roleplaying game that was kind of like Shadowrun in terms of setting, but mechanically different in a bunch of ways. It was cool. We were trying out all sorts of things, and we were playing all sorts of TRPGs. It was exciting: We'd play GURPS, we'd play Legend of the Five Rings, we'd play Shadowrun, and then we'd run our own stuff.

It was a pretty cool time. I wasn't really going to class that much. There were times in college when I was in four tabletop campaigns a week. It was a little extreme. I wouldn't necessarily recommend that to people, but I did get a ton of exposure. Not only exposure to different roleplaying games, but in testing my own games, I got a ton of great feedback, and I got used to that feedback loop, and tried to take criticism to heart, in a good way and a productive way to work toward a goal that made players excited about playing with the mechanics.

David Craddock: In your experience, what are some elements of tabletop games that have not, or cannot, carry over to computer and video games?

Josh Sawyer: My imagination. It seems so naive, but one of the first things I did was start designing unique items. I've always loved unique, magic items, and there's tons of cool Forgotten Realms modules and source books that had these really neat items with these long histories. If you look in The Magister's Element in the Forgotten Realms, it's full of all these really crazy, unique items that have these page-long histories with all these weird powers. [During development of Icewind Dale] I was like, "I want to do this, and I want to do this!" And the programmers were like, "Uh... how about none of that? Or how about a quarter of that?"

It took me a while to realize, oh, there are things the engine does, and there are things the engine doesn't do, and then there are things the engine sort of does, and maybe those could be extended. I had to very quickly understand that an engine contains a set of rules and capabilities, and you have to work within those as much as you can. Or, if you want to extend them, that's a cost. That's a burden on other people that you have to consider.

Another thing was I very quickly realized that the environment of making and playing a CRPG is not quite the same as tabletop, which I'd already intuited, but there are things in tabletop that don't work in CRPG environments, just because of the nature of the medium.

David Craddock: Icewind Dale was the first game you worked on as a designer. What was it like working at Black Isle Studios during that time?

Josh Sawyer: I don't remember there being a formal process. I think we would just talk to each other. Steve Bokkes, who had also worked on Torment, was in my office. He would give me feedback on my areas, and I would give him feedback on his areas. We would all give each other feedback on stuff, and Chris Avellone would give us feedback on writing. It was less formal and more ad hoc.

I feel like we all played the game a pretty healthy amount. Because there were only a handful of people really working on the game--half a dozen designers, maybe a few more—we were all in the guts of the project. It was pretty easy for us to talk with each other, give feedback, and then turn that around.

David Craddock: You were on a panel at this year's Develop conference with a few other developers, and one of your quotes made headlines. You said something to the effect of fans are resistant to change in RPGs, and that you'd like to see developers be more "radical" in their approach to RPG design. Could you contextualize that? I found your remarks interesting because, of course, over the past few years you and the Pillars of Eternity team have been heralded for, essentially, making games that hearken back to old-school concepts.

Josh Sawyer: People should understand that that quote was in the context of a panel specifically about advancing the genres we're working in. It's funny because I was one of four people, and we were all talking about ways in which to advance our genre, and I was a little annoyed that I was singled out. [laughs] I know, I know. Certainly much more prominent people than me were talking about it, too. But I think in part it's because roleplaying games tend to be pretty traditional, and I make pretty traditional roleplaying games, so I'm part of that.

I wasn't trying to say that traditional-style roleplaying games shouldn't be made. Obviously, if I thought that, I wouldn't have made Pillars 1 and II. But I believe there are a lot of cool and interesting tabletop and computer roleplaying games that aren't necessarily fantasy or sci fi; they don't necessarily have classes; they don't necessarily have ability scores or things like that; or they might be different in a bunch of other ways. All I was trying to say was I think there are a lot of interesting ideas, especially coming out of indie developers, and I think it would probably be good for the RPG community to give them more of a chance, and to not be quick to gate-keep what is or is not a roleplaying game based on preconceived notions.

I ran into that when I started working on the Aliens roleplaying game [at Obsidian]. I remember lots of threads where people were saying, "I don't even see how that's possible," just the very idea that you could have a roleplaying game in the Aliens universe was just, no, you can't do that. I think some players are more accepting than others, and I don't want to paint the community with too broad of a brush, but it is a genre that has a lot of very honored traditions that have been carried forward for a long time. I don't think we need to get rid of them, but I do think we should welcome people who want to experiment with formulas.

David Craddock: Experimentation and innovation seem like hard-and-a-hard-place spots for developers. Take Nintendo, for instance: Fans complain that all Mario or Zelda games are the same, but when Nintendo dares to shake up the formula in games like Super Mario Sunshine and Majora's Mask, fans complain that those games don't follow the formula.

Josh Sawyer: Yeah, I think that's true, and it's something that came up when Julian Gollop, who made the original X-COM—he was on the same panel, right beside me, and he was talking about the recent Mario/Rabbids crossover. It's sort of X-COM-like, and he said, "That's such a strange game, but it's a very cool game that did a lot of really interesting things." That's the sort of stuff where I think that, if people are open-minded about games in general, you can see a lot of really interesting things even from big studios.

Craddock: What have you found to be one thing that most players fail to grasp about game mechanics? Something that you find needs to be explained, or should be explained, regardless of genre?

Josh Sawyer: I don't know. It's weird. I feel like roleplaying games are in a weird space because there really is so much about them that's built around tradition. I guess a thing that needs to be taught, really, are basic dice-rolling mechanics. When people come to a roleplaying game like Pillars of Eternity, or like the Baldur's Gate and Icewind Dale games, if they haven't played a tabletop roleplaying game, the role of dice and randomness is something that I don't think comes easily to people, especially for players who play more contemporary games such as action-RPGs and first-person RPGs.

Those mechanics are things that people often either don't get at all, or they misunderstand. Because these games are so complex, and a lot of those dice mechanics are really tricky, it's a very difficult balance: We want people to be able to just jump into a game and play without inundating them with tutorials, but if we don't try our best to educate them, it can be very difficult for players to understand why they're succeeding or failing.

That's one of the reasons why, in Pillars of Eternity, for example, we [show] percentages to hit over the heads of enemies when you target them to attack. At the very least, players will say, "Oh, I have this much of a chance to hit," but that might be another step for them to understand the reason why that percentage is what it is: It has to do with their accuracy versus a specific defense. Some players don't get that, like, "Oh, I only a twenty-six percent chance to hit this guy. Why even attack him?" Well, you're attacking his Fortitude, and his Fortitude is 120 and his Reflexes are 70, so you should attack him with Reflex-based attacks.

Actually, I would say probability in general [is difficult to teach]. It's something the XCOM devs deal with a lot: People have a very difficult time understanding how probability works. This is true in general; it's not specific to games. There are people who will see, for example, a ninety to ninety-five percent chance of success, and if they [fail], it blows their mind because in their mind, ninety to ninety-five means one hundred.

Probability is very difficult to communicate. There have been so many instances of that. I remember one while making Icewind Dale II. Third-Edition D&D switched over Rogue characters: They no longer have backstab damage, they have sneak attack. Every few levels, a Rogue got another die. So, you'd start with 2d6, and then three [d6], four, and five, all the way up to fifteen d6. That's fifteen, six-sided dice. The actual maximum you could get is ninety points of damage, and I got a bug from QA at Black Isle saying, "Why do I never see ninety points of damage?" I was like, "Well, it's fifteen, six-sided dice being rolled." And they said, "Yeah, so I should see it sometimes." I'm like, "No, dude, that's not how probability works. There's a really, really steep bell curve in operation, here."

They wouldn't believe me. They refused to believe me. I showed them a chart of distribution, and they still wouldn't believe me until I took fifteen dice into a room with them, and I rolled, in front of them, fifteen d6 twenty times, and I added them up and charted them on a piece of graph paper. I said, "Do you see how this falls inside the bell curve?" Finally they kind of grumbled and accept it.

Probability is a thing that's very important in tabletop games because many RPGs try to emulate a tabletop experience, or borrow heavily from that, and it can be really, really hard for people who don't understand what's going on underneath the hood.

Craddock: I'm working on a book with Julian Gollop, and he said the same thing: Any time players see a probability of around eighty or above, and they fail, they assume the game cheated them, or there was a glitch or something.

Sawyer: I saw Julian demo Phoenix Point at Reboot in Dubrovnik, and really thought it was very cool, the way it shows you chance to hit. It doesn't show a percentage, it just shows a circle around the target that is being fired at. So you can see, oh, about this much of the enemy is inside the circle. It's a visualization, so people can see there's a little spot off to the side where a bullet could [go astray]. Because there's no percentage, when you miss, it's way easier to understand because you saw that your bullet spread included gaps where you could miss. I actually think that Julian's way of approaching [probability] is much more intuitive for more people.

Craddock: Do you think that visual systems would benefit orthographic RPGs like Pillars of Eternity, or do you think dispensing with traditions like showing percentages and dice rolls would alienate long-time players who want to simulate that D&D dice-rolling experience?

Sawyer: I do think there's a risk of alienation, there. One thing that I think solved the problem somewhat is we included a graze range, which I think did help a lot. Graze was a mechanic we introduced in the first Pillars. We initially got rid of it in Pillars II because some traditional players didn't like it. What graze did was, let's say you have a fifty percent chance to hit. That meant there was a fifty percent chance of hitting, straight up, but beneath that fifty percent, there was a small margin—I think in Pillars 1 it was thirty-five percent, which was pretty big—that you would graze. A graze only did half damage, and status effects [as a result of hitting an enemy] would only have half their duration, but you still made contact. You still did something.

For example, if there was a thirty-five percent chance [of grazing], if you saw sixty-five percent chance to hit, what that really meant was that there was a one hundred percent chance that you were going to land something: thirty-five percent chance it would be a graze, and a sixty-five percent chance to land a hit or better. That really helped because when you got into those high percentages, even if you didn't land a capital-H hit, you landed something. That, I think, avoided that feeling of, This is fucking bullshit.

We got rid of it in our Pillars II backer beta, and people went, "Oh my god, I hate this. It feels so [loose]" because it had gone back to feeling exactly how it was in D&D. So, we added the graze range back in, although it's only twenty-five percent. It does make a big difference, though, because that means if you see a seventy-five percent chance to hit, that really means that you have a twenty-five percent to graze, a fifty percent chance to hit, and a twenty-five percent chance to crit [land a critical attack], meaning you're never going to miss in that circumstance. That really helped a lot.

Craddock: How do you go about designing a tutorial that doesn't feel like a tutorial? No walls of text, no constant cutscenes, that sort of thing.

Sawyer: It's hard. One thing that we did in Deadfire is... Well, there are a couple of things. One is that we talked about game concepts and how we introduced them. We tried to be very careful in our content to not introduce mechanics before they're absolutely necessary. For example, when you're on the first island in the game, there are certain things that made me say, "Don't put afflictions on characters early, because then we'll have to talk about afflictions, and right now that's an advanced concept that players don't need to know. Right now players need to understand movement, and accuracy, and armor, and things like that."

There were certain things like afflictions, inspirations—all these other things, including the ship system, [that we made a point of explaining later]. I said, "The ship system is really complicated, so push that until way later until players have the basics of combat down. So, pacing is a big part of it.

Another thing that we do is now, when our tutorials come up, they come up on the side [of the screen], like, "Hey, do you want to look at this?" If not, it'll stay in a stack if you need it, or you don't have to look at it if you don't want to. For tutorials themselves, we tried to make sure that each tutorial has a lot of text in it, but we split that text up over multiple pages, and that the dominant feature of each tutorial page is either, say, a skill image that illustrates exactly what we're talking about, or a movie that shows exactly what we're talking about.

For example, Empowered is a new system in Deadfire that a lot of people weren't picking up that it was there. We kept increasing the size of the icon, and we animated it, making it all swirly and glowing, but people still missed it. Now, when we [wrote a tutorial] for Empowered, we actually show the mouse cursor going over to it, clicking Empower, clicking on an ability, clicking on an enemy, just showing what it does. We show that over three pages, each with its own movie, and a short paragraph of text, so they were much easier for players to digest.

Finally, we have a tutorial section in our journal. So if you say, "I shouldn't have skipped that one tutorial," or, "I remember that tutorial, but there was something in there about vision cones that I missed," you can just go back in and look at them whenever you want.

Josh Sawyer.
Josh Sawyer.

Craddock: How do you design a complex system, such as the ship system in Pillars II, that needs to be easy to use?

Sawyer: It's hard, especially for the ship system. Our stretch goal for the ship system was actually for crews. It was difficult to conceive of mechanics that would make use of the full ship through outside of ship-to-ship combat without creating a whole bunch of new content. It was something we experimented with. We actually had two versions of it that we went through, getting the level of detail necessary for it to be interesting.

I had to go back and say, "Okay, what are the essential things that, if we have ship-to-ship combat," what are crew positions that are important, and which ones feel superfluous?" Because it's such a complex system, I knew we'd need to pare it down as much as we could while still keeping it rich. And then, "What are some roles on the ship, such as cook and navigator, that are important for a ship, but aren't necessary for combat?"

The important thing in designing any system is asking, "What are you trying to do?" With the ship, the goal was to make you feel like a captain. To make you feel like you're responsible for assigning these people to their jobs, for taking care of them, for keeping their morale up, and making sure that the ship succeeds. That's one of the reasons why, in ship-to-ship combat, all of the commands you give are shown in quotation marks. We want you to feel like your character is yelling to the crew. We want you to feel like you're the captain, and you're the one making those calls.

With that in mind, we said, "We have to have a system that has enough salient elements to it that really captures the feeling of you being on a ship and ordering people around." When it came to the nature of the mechanics of ship-to-ship combat, we said, "What feels important?" Distance feels important, so do the types of cannons you have. There can be crazy events that happen on your ship, like your sails catching on fire. You can get cargo that comes loose and knocks around, injuring people below deck. These are things you have to command your crew to deal with.

Accessibility was something we worked on after we had accomplished the main goal, which was to make you feel like a ship captain. There were a number of ways we tried to simplify that. We had talked about more complex mechanics to layer on top, but we decided not to pursue those because the system was already complicated enough. We had different cannon types that required different ammo, and found that confusing, so instead we changed cannons so you can choose what ammo you fire, which makes that more intuitive. It also tells you what each ammo type targets. For example, if you shoot cannonballs, you're targeting the hull and sails of a ship. If you fire a grapeshot, you're trying to injure their crew.

Those were things where we tried to, as much as possible, make systems more approachable. In the end, though, we did realize that there were some players who were not going to dig the [ship] system one way or another. Either they just don't want to do it, or they don't like our implementation of it, which is why we have close to board. Very late in development, we implemented a command called close to board, so at the beginning of [ship-to-ship] combat, so you can say, "You know what? I just want to do the part where I jump on their deck and start killing everybody with my crew." You can do that, you just automatically pay a penalty and take a bunch of damage, and some of your crew members get injured, but if you don't like that system you can skip straight to deck fighting. That's perfectly okay.

Craddock: When I talked to the guys who made FTL, they said, "X-WING was a big influence on us, but we had trouble designing a game around that, and once we zeroed in on what we wanted to do"—make the player feel like a captain—"we were able to filter all our systems through that metaphor." Do you find that as a designer, once you hit on a metaphor, does that simplify design for you?

Sawyer: It does help to simplify it, that's certainly true. I think one of the struggles is that the ship system sits on top of a whole other game, whereas I think FTL feels more focused because ship-to-ship play is the core of the game. Then there are these cool strategic choices you have to make, jumping from system to system. In Pillars, the greater difficulty is, we have a whole game about exploring a world, walking around cities with a party of adventurers, with a ton of equipment, going into dungeons to do quests. Also, when you get on your ship, you have a whole crew, you have cannons, sails, different ships you can upgrade to, and you can engage in ship-to-ship combat all over the sea.

I think that if our game was just focused on the ship part, it would be much easier to design. I can't say our ship-to-ship combat holds a candle to FTL's, because theirs is really good. But I do think that when we hit on that key idea of, "Yeah, we want you to feel like a captain," it does help refine design in general so you can at the very least refine it down to stuff that's essential for achieving a goal.

The first floor of Kresslack's Tomb, one of the dungeons Sawyer designed for Icewind Dale.
The first floor of Kresslack's Tomb, one of the dungeons Sawyer designed for Icewind Dale.

Craddock: Who are game designers you respect, and who have influenced your career?

Sawyer: I would say one big one is Arnold Hendrick. He designed the Darklands roleplaying game, which is from 1992. It's a MicroProse game. He was a historian, and I'll say that Darklands was one of the things—because it's historical fantasy—that got me really interested in history, and ultimately was one of the reasons why I ended up picking a history degree.

I thought that his approach—and Sandy Petersen's, who was also one of the designers of Call of Cthulhu—I felt that their approach to making an RPG was extremely fresh. It was the first roleplaying game I remember playing that didn't have classes. It didn't have levels, it didn't have alignment. Characters didn't really get more health as they advanced; in fact some of them lost health because they would age. Priests called upon saints for aid. They didn't cast spells. Alchemists weren't wizards. They made potions, and would throw potions to achieve magical effects. These were all things where I was like, "Wow, this is really different."

It had a life path system, and I had seen some games that had life path systems, like Traveller, but to see it in a computer roleplaying game, and presented in this really cool, historical-fantasy perspective, made me say, "Wow, this is so cool." I recognized that there were problems with it. Even though I had seen different tabletop games, [Darklands] made me say, "You know, computer roleplaying games can be lots of different things. They don't have to be like D&D." Which is, I know, I'm the guy making Pillars of Eternity, which is very much like D&D. [Hendrix] was very influential on me.

Tim Cain is another. A lot of what I do with freedom of player expression and choice, and how players can move through a story, I look to Fallout 1, especially. The structure of Fallout: New Vegas, that idea that you can go straight to Benny [an antagonist] as soon as you wake up in Goodsprings. You can just head to New Vegas and find Benny. You don't technically have to do anything else. That comes from Fallout 1. You can go straight to the Necropolis in that game. There are big inspirations that I took from Fallout and Darklands especially.

Craddock: A lot of creators talk about needing to be passionate about the project they're working on. You know, make the game you want to play, write the book you want to read. You've talked to me about making objective decisions because you're ultimately designing for an audience, and games have to make money. How do you balance designing for yourself with designing for an audience? Maybe there's a feature you really want, but you're not sure if it will take off.

Sawyer: It's difficult. When it comes to my personal tastes in roleplaying games, whether it's tabletop or CRPG, I generally don't play class-based games. I prefer classless games that are a little more organic. Designing a class-based game, especially with a lot of classes, playable races, sub classes, and all that stuff, is very hard. There's a lot of stuff to make, and a lot to track and balance. The goal is, no matter what anyone plays, every combination [of abilities and choices] must be viable. That's a challenge. It's also challenging to make sure combinations aren't so ridiculously overpowered that they make the game trivial.

For example, I knew that people would really dig sub classes and multi classing. I knew that in doing that, it would open up a lot of balance issues that I wasn't super comfortable with. But I also knew that ultimately, people would be very excited about it. That was a time when I said, "You know, I think this is something people will really enjoy."

On the other hand, there's a change we made that was very unpopular, at least at first. I know some people are still mad about it, but when went from six party members in Pillars 1, to five in Deadfire. That was based on a variety of concerns. The concern was not, "Six party members is too many," but there was a larger concern of keeping track of everyone. The clarity of combat could be very hard to follow. There are a number of things we did to make the pacing of combat better, to make it easier for you to use all of your party members. One of those decisions was to go from six to five party members.

There are quite a few players angry about this, but I'm seeing more people saying, "Wow, combat is so much easier to follow." The visual effects are less overwhelming. We dim the visual effects when we pause the game, so if there's a big fireball being cast on your party, you can see what's going on when you pause. We have combat speed sliders so you can take the combat speed way down if you want. And the overall pace of combat, the base speed, went down as well. So, when I see people realizing, "Oh, all of this does actually make combat easier to follow," I'm not doing these things to aggravate people who want six party members, but sometimes I need to take a risk because I think the game could be very great.

As another example, there was a [debate]. People hoard consumable items, like potions and scrolls, in these games. One of the things I wanted to do on this project was, I proposed converting all potions and scrolls to encounter or rest-use items. So, instead of being stacks of consumable items that some players would burn through and earn a massive bonus, and other players would never use because they're always afraid they're going to run out of them—which is a very common concern with consumable items—they would have these things where it's like, "You have a potion. If you use it, you can use it once per combat, and then next combat you can use it again." If you get a scroll, you can do the same thing.

So, it more becomes more about your options and opportunity cost than, do you use your consumables or not? If you're not using consumables, you're not saving anything. Bobby Null, who's our lead designer, he's much more traditional than I am when it comes to roleplaying games, especially D&D-like roleplaying games, and he said, "I think people aren't going to like that." It was an area where I was like, "You know what? If people don't use potions, it's not the end of the world. I would prefer people use potions more often, but we're already changing so many other things; why push this particular button?"

It was something I felt we could move forward with, but, again, with something like this, there are only so many things you can change before people say, "This really doesn't feel at all like it's in the spirit of what I wanted."

Craddock: I'm sure this varies from project to project, but do you have tools of the trade you use, in some form, no matter what game you're making?

Sawyer: I always use Word, I always use Photoshop. Those are some basic tools. I document things, but outside of that, we have our Obsidian tool set, and that, thankfully, is something I'm able to carry from project to project. The core features of that are really the dialogue and quest editors. At Obsidian, we have a way of writing quests and dialogue that has been developed over fifteen years, and then before that at Black Isle. That tool set is very particular to the way we work. It has tons of incredible features in it. I shudder to think of trying to write dialogue and quests without that tool.

I'm sure other companies that specialize in RPGs probably have similar tools. For example, Telltale has their own branching-narrative tool. That's a thing that, because Obsidian tends to bounce around technology wise, whether we're using Unity or Unreal or the Cry engine for Armored Warfare, being able to work in the Obsidian tool set is both comfortable and a constancy for our development environment.

Craddock: In our discussions, you've explained that the first Pillars of Eternity was pitched to backers on Kickstarter as a spiritual successor to Infinity Engine games, but that Pillars of Eternity II was more about designing within the parameters of Pillars 1 rather than Infinity Engine games. Do you feel that a spiritual successor would be better served evoking rather than replicating?

Josh Sawyer: That's a hard question to answer. Better served to whom? That's the question, I think. Better served for a different audience? Yeah. Or, you know, for some segment of the audience that backed it, but maybe a big chunk of them not enjoying it? Sure. I think it's a dangerous game to play. Without the people who are super passionate about the games that inspired this, we wouldn't have been able to make [Pillars] at all.

Now, I can say that I think there's this mythological, objectively better game that does a lot of different things. Maybe it's not class based. Maybe it's turn based. Any range of other things. And I can say that I would make a game that more people generally might like, or that I personally might enjoy more, but I don't think it would be responsible of me to just throw all of that to the wind. When Feargus pitched me on the idea of doing the pen-and-paper game, I said, "I will do that if I can tell everyone up front that I'm going to do whatever I want; that it is not going to be class based; that it is not going to have a d20 base; that it is not going to be compatible with D&D; and that I will listen to people, but I am absolutely not going to bend over backward to make it fit their vision of what the game should be. He said, "That's fine."

So, I told people that, and then I started showing people the pen-and-paper rules, and some of them started getting really bent out of shape. I said, "Sorry you don't like it. I'm not changing it." Because that was the expectation: I said from the beginning that I was going to do whatever the hell I wanted. If we were to do a new game where we said, "We're going to make a roleplaying game that's going to be kind of like this [model of game], but ultimately we're going to change a bunch of stuff if we think it's going to make the game much better," then that would be a different conversation.

I don't think that with Pillars 1 or II, [deviating far from Infinity Engine games] would have been fair.

Craddock: You might not be ready to talk about this since Deadfire is recent as of this interview, but one thing we've talked about is where you'd like to CRPGs go. I'd like to revisit that. Where would you like to see this style of CRPG go, or how would you like to see it evolve, post-Deadfire?

Josh Sawyer: I don't know. I don't know that this style of RPG needs to evolve that much. I think this is the traditional RPG. Any sort of modifications... I guess the thing is, if someone wanted to make a game that was isometric and turn based, like a game similar to [Pillars of Eternity] and has real-time with pause, it doesn't have to be class based. Tyranny had a system that wasn't class based. I think that's totally viable and cool. You can experiment more with the number of parameters, and you can also look at making mechanics that revolve around cycling companions in and out of a party.

One thing that's an ongoing struggle that we have to deal with is inns and camping. They're things that feel very appropriate for D&D, but it's difficult to incorporate those mechanics into a CRPG that doesn't resort to save scumming, or necessitate things like repopulating enemies after you rest; otherwise resting can feel pointless. We've tried two different revisions: Pillars 1, and Pillars II. I still don't think we're all the way there in terms of making a tight cycle out of [resting in camps and inns].

If you look at a game like the tabletop RPG Torchbearer, they have an extremely, extremely tight time economy where resting is precious and dangerous, to the point of really being anxiety inducing. Also, Darkest Dungeon: These are games where the whole game cycle is built on this downward grind that you have to close yourself out of by retreating and resting. But the thing is, that kind of feels like it's in conflict with the nature of the Baldur's Gate, Icewind Dale, and Pillars games, where they're open-world, you're exploring and can come and go as you please. It's not the same as delving into a dungeon where you're on this tether.

So, it's a tricky balancing act. Dungeons & Dragons, as the name implies, pioneered this dungeon-diving mentality where you're trying to survive and scrape by, which is really something Torchbearer tries to [capture]. But RPGs are in a difficult position because some of them feel like mechanics are in conflict with the way in which a world is laid out. These are things I think should be either integrated—like if someone makes a game that isn't Pillars of Eternity, that's not trying to emulate this mix of somewhat-in-conflict mechanics... Because that's the thing. There are a lot of things in old D&D [rule sets] that are in conflict, and my job as a designer is to try to keep that spirit alive while making games not obnoxious.

If someone wants to, they can make an isometric game that's entirely focused on dungeon delving, even more so than Icewind Dale, and really focuses on this tethering-survival concept that makes resting a core mechanic. Or they can make a game where resting isn't really a part of it. It's not what the game is about. That would be a really different feeling, but you can do it, and it could still be a roleplaying game with choice and consequence, including certain classes and races. You would just think about your economy of time and resources differently because you don't have a rest move. Stuff like that.

Hello, Meet Lola