Level With Me, Dan Pinchbeck

“Level with Me” is a series of conversations about level design between modder Robert Yang and a level designer of a first person game. At the end of each interview, they collaborate on a Portal 2 level shared across all the sessions – and at the very end of the series, you'll get to download and play this “roundtable level.” This is Part 1 of 7.

A few years back, Dan Pinchbeck and I were crowned "frontiersmen of modding" by Jim Rossignol. I felt like prom queen all over again. While I promptly sank into relative stagnation, Dan concentrated his powers into a brilliant point of light -- a full-fledged retail reboot of the 2008 Source-powered ghost story mod Dear Esther, fresh from the beach with amazing art design by Robert Briscoe, backed by the Indie Fund, and now looking better than ever. (Warning: there are mild spoilers for the endings of S.T.A.L.K.E.R: Call of Pripyat and Metro 2033.)

Robert Yang: So how's it going? What's the last thing you did for Dear Esther?

Dan Pinchbeck: We're in the middle of porting to the Portal 2 engine, which Rob [Briscoe] and Jack Morgan are wrestling with, and Jess [Curry] just re-orchestrated the soundtrack. It's weird, it's my baby, but it's an awful lot of Rob's vision too.

RY: Aren't you worried your own voice will get diluted?

DP: No, I'm not. I've resisted a cult of personality in games. If there's going to be any auteurship, then you need a collective vision as a studio and more realistically as a small studio. But as I've said with this rebuild, it's a collaboration, and now Rob and Jess have brought as much to the table as I have, and I feel really lucky that we have these people with their visions aligned.

I think what I like about games is how you don't have a clue who made most of the games you play. “What do you think of 4A Games?” Well, who's that? Metro 2033! You don't have a clue who made it, and I think that's a good thing. I don't pick up a game because Warren Spector made it. It's either good or it's bad, the same as any other product or art piece.

RY: But some people see “Warren Spector” and think it'll be awesome.

DP: Yes, but he's perfectly capable of making bad games (… in principle.) I don't care how long it took or why they made it. I just want a good experience. I don't think it's limited to any type of team. Anyone can make a turkey.

RY: I guess you're this weird blend of industry / retail, and academic / research stuff now? What's that like?

DP: It's like banging your head against the wall until you bleed. There's loads of issues – how people are employed, contracts... it's hard to make those two worlds meet. No one's tried to take a publicly-funded research project and convert it into a release on Steam, but we've always argued you need to put your money where your mouth is, that if you want to talk about games you ought to have a deep understanding of what it's like to build them. I'm very impatient with people who keep going, “Games should do this.” Well, build the things and just find out!

RY: What are you finding you can't do with Dear Esther, with the new commercial / retail focus?

DP: It certainly helps that the quality of the assets is exceptionally high. The quality of Rob's artwork, Jess's music, the quality of Nigel [Carrington]'s voice over – and vaguely so, the quality of my writing – definitely help with opening doors to someone like Valve to talk about licensing. We've kind of smuggled something quite niche or experimental on the basis that these assets are AAA quality. We're also really lucky with Indie Fund. They just really got behind it, saying ,“we're not going to tell you what to do with your vision, now that we've given you the money.”

RY: I bring it up because – you may not want to hear this – I felt like the best voice over chunks were the ones that were less, uh, “British,” I guess? I mean, I had no idea what a “semi-detached” was, for example. [In the US, we call them “duplexes.”]

DP: Right.

RY: I liked the passages that were mini-stories, incidents I could picture in my head. Then sometimes he says 10 metaphors in a single sentence and it's totally disorienting. Which can be good, but I was wondering if you were streamlining it or making it more accessible or anything?

DP: In some places we've added new speech cues, like 4 instead of 3. There is some stuff that should clarify some things, a couple new combinations of the new script that radically alter what happens, which was deliberate. When it first came out, we were looking at what people were saying in the forums and thinking, “Wow!” There was one in particular where there was none of that really in there but it was such an interesting take on things. I wanted to just drop a couple more hints in there, to see if anyone else picks up on it.

For me, writing in general – and I think there's a lot more scope for this in games – I don't think things have to make sense. We don't require logical, causal sequences in many other art forms. With Dear Esther, it doesn't matter if you don't understand anything so long as it creates the right mood. There's also a lot of rationalizing going on here; really, I just sat down and wrote it. In FPS's it's always, “save the world, save the universe, it's huge,” but I liked the idea that a game could be about something ridiculously small, with no meaning to the wider world, yet it means everything to the people involved in it.

RY: It's like lower stakes, but more personal stakes.

DP: Right, massively high for an individual. I was playing S.T.A.L.K.E.R: Call of Pripyat, and at the end of it – it's so brilliant – they go, “let's just get on the helicopter then!” And then there's no climax. It's like getting into a taxi at the end of the night, and when you think about it... that's actually quite cool in a way.

RY: It's not a cop-out?

D{“ No, it... it would be a cop-out, if it wasn't so completely unusual. No game has the balls to say, “and then you wake up and it was all a dream.” It'd just be fantastic to throw that in the player's face.

RY: Uhh... I don't know. Isn't that kind of lazy?

DP: Totally! And it's interesting that it's lazy! But everyone goes on and on how game writing is already lazy, and we haven't even stooped to the same poor tricks in film or literature...

RY: … So we should?

DP: Why not? It'd be a bit of a laugh. I guess I'm a bit of a tourist. I don't need a game to be world-changing. I'm a huge Just Cause 2 fan, a Dead Rising fan. And they just do stupid stuff to pass the time. While I do quite a bit of intellectualizing during the day, I'm also quite happy to just take a chainsaw to a zombie's head for half an hour.

There's quite a bit of rhetoric, that “games are failing to evolve” and “need to be saved” and I'm actually pretty ambivalent towards that, I'm actually quite happy that I can pick up Resistance 3 and be a complete neanderthal for a couple hours.

RY: Yeah, I agree. It's a fallacy of progress. It's science envy, physics envy.

DP: I guess it's all been hard work and a lot of fun, and when you get around to actually building the thing – like Dear Esther, putting two years into it – I kind of want to enjoy making it a little bit.

Watch on YouTube

RY: Speaking of making, I remember seeing this photo somewhere, of you and your team sculpting the levels out of clay? Is that true?

DP: Yeah, we got 4 or 5 kilos of clay and just dumped it on a table. I'm a writer, so I really just humiliate myself whenever I get into any kind of program, and I can't draw –

RY: But you're a sculptor?

DP: No! I thought... Josh [Short], who was the artist, I wanted to tell him, “I wanted to be able to see this from here” and “that from there, with this scale and angle,” and we tried this flip-pad of pictures but it looked like something out of a serial killer movie – so we just grabbed this massive lump of clay and slapped it on the table and hacked it into a broad shape.

RY: And it worked?

DP: Yeah, it worked actually! Most of the issues from the original build came from neither of us using Source before. Most of the principles of design worked, and that was indicated when Rob went in and rebuilt it and most of the level design has remained intact. The only level that has dramatically changed is the last one – it's much much bigger. Before, we built the mountain and things were popping in and out... it was just hideous. We spent 2 months trying to sort it out. Now, Rob's sent me the final level and I've just – wow. It's so far beyond the original.

RY: Actually I think that worked to your advantage. Not to disrespect you or anything, but if a more experienced modder were working on that – the conventional thinking and knowledge of technical limits would've stopped them. Being an outsider helped you make this crazy giant mountain that was destroying the engine.

DP: Yeah, you're not going to win all the time, but there is a place for... naivete. You can fail in academia, as long as it's in an interesting way. There are things we learned from Korsakovia, things that I probably wouldn't have the balls to try unless we knew some things – and most commercial games stop way short of that. Well, until Frictional [Games] came along and did everything we wanted to do in Korsakovia but better. It was kind of vindicating.

RY: What I took away from Korsakovia was that you were trying to directly represent fragmented identities through level architecture, making it difficult to navigate – and I think the result was, I don't know – you can't... do that?

DP: Certainly if – if you're going to do that, you need to spend more time and money. Korsakovia was made basically by one person in an absurdly short of time – Adam Griffiths, who's now at Rebellion. I can't keep beating him up, when I would say, “I want it to do this and this,” and he was going, “but Source can't do this,” and I said, “yeah, I know.”

The last level he made was beautiful. There's still nothing like it out there. He had the whole thing orbiting a central pillar, so you have the entire level spinning with other components spinning on a different tilt –

RY: And then Source exploded?

DP: Well in Source, if you have a room tilted at 45 degrees, all the furniture will slide through the crack between the wall and floor –

RY: Like a tiny one unit crack?

DP: Yeah, and it still does it! The level was going to be utterly disorienting. Which in a way is really bad level design, but if we can find out where the line of tolerance is, that's got to be a useful thing. It's not like we lied about the experience [players] were going to have, that we charged them for us to find out.

RY: But “gamers” still have this sense of perceived cost, like, “oh, I'm never going to download your mod again!” As if we owe them something.

DP: Well, I guess we are asking them to invest their time, and if the risk doesn't payoff, they'll blame someone for it. (Like Warren Spector.)

RY: About the player experience – I was looking through the file structure of Dear Esther, because I'm weird like that, and saw in NewDearEsther_English.txt you originally let the player choose to start on any level. Why take that out?

DP: I wouldn't want people to jump in halfway though. I'm not even sure if I'd want them to, even once they've played it. I remember on Lou Reed's “New York” album – on the liner notes, he said this album was supposed to be listened to like a movie, and explicitly, “please don't listen to tracks here and there.” Which is slightly arrogant. I always hate it when people say, “this is the only way you can consume this thing.” But if you have the capacity to go into a story at any point, you're opening the door for people to do that. For Dear Esther, it's important to have that emotional experience building up or you'll miss a big chunk of it.

RY: So in this version, are you going to allow players to save the game, then come back in a few days and forget entirely what happened?

DP: I don't know. I think the game really rewards replay in a way the mod didn't. It's more dynamic now. There's so much going on in the first two levels, to build up the subtle variations in experience, laying so much groundwork for what happens. You need that whole runway to take-off and get the whole experience.

RY: But how do you playtest this? How do you know that players are actually absorbing all this groundwork and runway?

DP: All the feedback on the forums that I looked over with Rob – we had the equivalent of 4 years of playtesting. With a lot of the interpretations that people were getting, I knew the script well enough to recognize some of the seeds sewn from level one.

RY: But would you have them fill out a questionnaire? How do you really access that part of a player's brain?

DP: As a writer, I was interested in writing something that actively resists a single interpretation. It's really bad from an academic point of view, how we can't data-mine anything or falsify a hypothesis, but we can ask, does it feel like this is working? Artists and writers work off that basis. That's a lot of the tension between working as a game designer and working as an academic. I was talking about this with Doug Wilson, the creator of B.U.T.T.O.N, about how you get these two really contradictory hats in one place. For me, the designer always wins.

RY: I think it's interesting that you're so intent on giving up all this control, but then you still want to grip a lot of it – like, “no, don't start in the middle of the game” – but also, “oh, think whatever you want of this game too!”

DP: But that's okay. If you look at poetry, no poet ever bitched about having to write in iambic pentameter. You work within those structures. It's really hard to be creative without them. If you can do anything, then where do you start? As a game designer, then – you know I have to work within this model, in terms of what players and technology can tolerate, and you have to find a vision you'd find exciting within those parameters.

RY: Speaking of models – do you still believe in your theory of ludodiegesis? Because I've kind of been pushing it hard to everyone. I want to it to catch on. Instead of saying “groovy” or “rad,” people will say, “that was totally ludodiegetic.”

DP: [chuckles] It's just academics making up phrases... but I really like “diegesis” as a phrase. You can talk about world as story, not plot or character. It has its own nebulous logic to it. Take Civilization, which is really problematic to talk about for narrative, but people do, and it's usually really embarrassing. You can certainly say Civilization has a world, though.

I just want story to be talked about as a gameplay element that sometimes isn't there. It's part of the set of tools that a game designer uses to create an experience – and it should be thought of along the same lines, as physics or AI or something more mechanical.

RY: With that in mind, let's see if this Portal 2 design portion of the interview will work at all. [Plays through the current iteration of the Portal 2 level. It takes about a minute.]

DP: Okay, so, uh, what do you want to do with it?

RY: I have no idea. It's totally up to you. Do you want to add some kind of narrative to it, or... ?

DP: What about – when you shoot a portal and you go through it, you end up where you're not supposed to be? Is that possible? Like, put a portal in that top-left window... and set your orange portal over there... and then you don't end up in the room you planned to. Is that even possible?

RY: I don't think – well, first of all, there's that HUD indicator that shows where your portal is at all times.

DP: Yeah. That's a shame, then. [Portal 2] was so linear for me. Like I had to find this one thing...

RY: … Especially the underground part, where you're looking for just that little bit of white wall.

DP: It ended up being weirdly less creative in that regard, like an FPS where there's literally only one target and one bullet. [contemplative silence] It would be interesting to have an identical second room, that isn't the first room – it's always kind of my thing, to take the player in one direction but make it something else.

RY: Okay, copy the entire thing... um, is this how you work on Dear Esther? Hover over someone's shoulder and tell them what to do?

DP: The new game we're starting – there's a huge design document with just the bare minimum. Event lists, asset lists, the scripts, the full player walkthrough... it's not, “this wall is this height,” but it's more about capturing the mood of the space. I think it's great to give them as much of a free hand as possible.

RY: Is it frustrating? Because I'm having to collaborate on a project right now, and it's frustrating me – or maybe I'm just a control freak.

DP: I thought of it as: if I want to do this, then I have to be able to communicate the design I want to people to do the work, because otherwise there would be nothing. So much of it is about finding those people.

… So what if you – um, you go into the lift, when you go into the exit?

RY: What? Which exit? I've cloned it.

DP: How many rooms have we got now? Three exits in the same room?

RY: In the top view, you can see this is where they start – then these two rooms are exactly the same.

DP: Shall we – no, that'll just be messing with people too much – when they go in the lift, they find themselves back in the same room again.

RY: [laughs] Okay.

DP: … And then they do it again! And the impression is, the first time, they'll go, “I'm back in the same room again! Ah, I get it!” Then when they do it the third time, they'll start thinking something's wrong.

RY: Like there's a puzzle to solve here.

DP: Yeah, but there isn't!

RY: [laughs] [unintelligible muttering]

DP: I was talking about this with the co-designer on the new game: how do you get people to think there's a puzzle, but there isn't – or think there's a maze, but there isn't – you're just screwing with them. How do you do that in a way that's really pleasurable? It's about finding those spaces, where you can't slide into a comfortable mode of play, where you go, “ah, I've seen this, I've been through this before.” One of the things we're doing with Dear Esther is how it all might look exactly the same, but the change in details is really subtle, kind of like a “find that one object” thing.

So maybe do some graffiti so you feel like you're going through these repetitive rooms, but it's changing as you go back through. There was... have you played Metro 2033?

RY: No, haven't had the pleasure.

DP: Heh, I'm not sure pleasure is the right word, but there's a sequence where you're running around in the sky and corridors are being built around you as you go, and you run until you see a weird bad guy – then you turn around and run until you see another bad guy – so then you turn around, 180 degrees, as if you were running back towards the first bad guy – and you can't figure out if it's a maze or not. You can either just run blindly, or stop and be aware that you're being played with a little bit.

The wall that's directly below you...

RY: Yes?

DP: If the player keep going forwards, they're in a constant loop, so they have to go back into the lift they first came in, and that's what takes them out.

RY: Wait, so they go back into the entrance elevator?

DP: Yeah, they turn around to escape, otherwise they just stay in the loop forever. They keep solving the same problem again and again and again.

RY: Well, the counter-intuitive thing with that is... the way I thought I'd do this is you choose one of the exits, you go in, you go up the elevator, then I teleport you back through the logic that controls the entrance elevator – but when you come up in the entrance elevator, you should be able to turn around and go back into the entrance elevator... and then leave?

DP: I don't know. That's really pretty simple, isn't it? But it's kind of weird.

RY: No, no! I like that.

DP: Because you just said, you wouldn't do it? So what happens when the penny drops after just going around in an infinite loop?

RY: I do like the idea that you have world state changes, to indicate that there's some kind of progress being made.

DP: What if you have some debris, down in the pit, and you just move it around a bit? And you notice it's kind of moving?

RY: Wait, so what's the solution? Or is there no solution?

DP: To this? Go back.

RY: Go back.

DP: Yeah, just have that trigger disabled – and enable the trigger after they solve the puzzle once. So the second time they go through, they can just turn around and go back in.

RY: Hm. The tricky thing is the way Portal does all the logic. These are all in instanced map files, so then you have 4 different VMF files all talking to each other... I'll figure it out though. That implementation you described will work. Anything else you want to add?

DP: How about when you go in, it's playing one audio file, almost like piped music coming into the test chamber. The second time in, it's the same MP3 (something fairly classical) and you go into Audacity and just randomly pull a few notes out. You can stack up 5 different versions of it.

You know the film Inception? It's just the same piece of music, massively slowed down. So the rooms are different, a different layer of reality – but actually it's just a slightly screwed-with MP3.

RY: Any particular song I should use?

DP: Um... “Somewhere Over the Rainbow.” The Judy Garland version for maximum freak-out value. Audio is always really great to take design in different directions without having to program and design loads of stuff, yet it has a powerful effect on player experience.

RY: But isn't voice acting and doing re-takes and re-writing the script – isn't that expensive?

DP: Not compared to modeling and programming! Certainly in professional game design, you can chew through 7 or 8 writers, 7 or 8 scripts – you can do all that for a lot less money than bringing on a programmer. And there are lot of writers out there desperate to write for games. It's not rocket science, really. It's so cheap for what it can give you.

RY: You talk about a writer like they're just another worker on this assembly line. Isn't it a case of too many cooks in the kitchen?

DP: For me, I feel an affinity to people like Ken Levine, people who are designers and also writers. That's different from doing script, from doing dialog. Tom Jubert talks about narrative architecture, how it's a process of designing the entire flow of everything, and script isn't exactly part of that.

RY: Well, can you explain that, actually?

DP: A narrative designer says we're going to string these events, make the scene change here, and change the emphasis here to craft a more engaging game experience – and at the end you just slap some dialog on there. That's the standard model.

It's what I talked about in my PhD dissertation, to say that actually those classic writerly things don't fit well in games, but these particularly do fit well. A writer isn't just a dialog generator; they're the one who says, “if we sculpt the experience in this way, then we'll get something likely more engaging.” But it isn't always appropriate, some games just make you wish they never bothered with a story at all.

Italian versions of “Somewhere Over the Rainbow”? That sounds good. Having fun with the player is always the best bit.

RY: You are a control freak!

DP: … I am a control freak. Yeah.

RY: But that's not a bad thing. I sympathize.

DP: It should be fun for everyone, really. When you get knocked off-balance by something, you tend to have a really good experience. It's possible even in commercial games.

Prototype did it well. It told me I'm not actually a human being at all (that's kind of weird) and I'm not sure how I cope with being told I'm a killer bio-weapon, a virus – how I'm playing a group of cells in a body. That's just really odd. It kept me going back to it for about half an hour, playing with my head a little bit.

RY: Isn't that a bit like Assassin's Creed?

DP: Oh, Assassin's Creed! I can't stand it!... No, [Prototype] wasn't like that at all, it didn't really change any gameplay, to go out and play some appalling “Desmond levels.”

It was more about how with one line of dialogue, you've added way more than one line of dialogue to the game, or with a couple lines of voice-over you completely spin the experience around.

RY: Thank you for your time.

Looking for the Portal 2 level? Sorry, you'll have to wait until Part 7.