Story Time With Valve’s Erik Wolpaw, Pt 2

Yesterday, you probably read the first part of my chat with Valve’s Erik Wolpaw and Double Fine’s Anna Kipnis. If not, it’s right here– but FOR A LIMITED TIME ONLY. By which I mean until the Internet ceases to exist, which, you know, could happen someday. Anyway, in today’s installment, we branch out a bit from yesterday’s story-centric beat. Valve’s newfound love of wearable computing, virtual reality, heaps behind-the-scenes info on Portal, crowd-sourcing, and more are all on the docket. OK, there wasn’t actually any sort of docket involved. I’m not entirely sure why I said that.

RPS: During a panel, you talked about a bunch of stuff that got removed from the final product of Portal. The whole Turret Wife thing, where Chell would marry a turret and then it’d stalk her – Slender-style – for the rest of the game. Stuff like that. Which was hilarious. Can you talk about any other stuff you’ve taken out of games? Or projects that Valve hasn’t gone through with?

Erik Wolpaw: I talked about something today, where you were going to actually meet Cave Johnson at one point in the underground. He was just going to be a… They had actually uploaded his AI into this crummy cardboard box, basically. You find it and there’s a ledge that you have to get onto to get out of this room. It’s too high for you to jump on, and because it’s a game, you can’t mantle on to it. So you have to unplug Cave, effectively killing him, so you can use him as a stepstool to get up onto this ledge. We just never quite worked out the gameplay for it, so we cut that.

The Turret Wife thing, everybody liked it, but there’s just cuts that need to be made in order to ship a game. At some point we sat down and we thought about it. We said, “This is funny, but it doesn’t move the story. There’s no gameplay thing that we’re cutting. If you cut this, the story still makes sense. It hangs together. Let’s make the decision not to put it in.” It would have been a huge amount of effort to do that Turret Wife thing, problems to solve. At some point we were just like, “Good idea, but it’s not worth spending another year working on the game to put this in.” Especially because there’s no story beat that it really fulfilled. It was just this funny extra bit.

RPS: How often does that happen? Would you say that happens more often at Valve than it does at other companies? Where you say, “This is good, but not quite good enough?”

Erik Wolpaw: It happens often. Again, game design, especially at Valve, is iterative. You’re always throwing stuff away. Usually because it’s not working or it’s not working well enough. Sometimes the amount of work to get it to work just feels like it’s not going to be worth the payoff. You’ve got these more promising leads to work on. It was similar on Psychonauts, and I assume it’s similar on a lot of games. Probably more likely to happen at Valve because, again, we can throw stuff away and we’re willing to take our time. We get a lot of feedback and try to react to that feedback.

RPS: I semi-recently got to talk with Michael Abrash, actually.

Erik Wolpaw: Yeah, super smart guy.

RPS: No kidding. He’s basically paving the way to the future with VR and wearable tech. How much interaction is there between what he’s over there doing and what you’re doing – if any?

Erik Wolpaw: There’s some. We look at each other’s stuff. At some point, we try to cross-pollinate – get everybody thinking. So they’re not designing the hardware in a vacuum without thinking about what the applications of it would be. We also will bring game teams down to just look at it, to say, “You know what, in your spare time, think about this stuff. Does anything occur to you?” Valve’s open enough. If somebody has a great idea, they’ll go down there to Michael or Jerry or any of the people working down there and say, “Hey, why don’t we try to mock this up and see if it goes anywhere?”

So yes. A lot of it’s informal, though. It’s just people talking with each other. The big thing at Valve is… It’s almost like the standard greeting. If you see somebody you haven’t seen in a week, you’re like, “Hey, what are you working on?” And it’s not, like, “What are you working on because I need to know so you don’t get fired.” It’s more like, that’s how you figure out what’s going on. You just talk to people.

RPS: I got to mess around with Oculus Rift recently, and – while shooting demons in Doom was cool – I only got really excited when I started thinking about how it could open up new genres. Have you given any thought to that kind of thing? 

Erik Wolpaw: Not yet. I haven’t gotten to put the Oculus on yet. I went down to look at it one day, and it was broken or something. I haven’t been back down since. So no, I guess. I looked at some of the AR stuff, but I’m not thinking too much about it. Other than, in the back of my mind, just like “What can we do with this?”

RPS: So, since both of you work at two of the most unique studios in the industry, and Erik’s actually worked at both, how are they different? I mean, it seems like Double Fine’s very Tim-driven, while Valve is just this amorphous human blob of self-motivation. Somehow, though, you both end up with very focused, extremely well-realized worlds and characters.

Anna Kipnis: One thing is that, the way that Double Fine is now, it’s very different from when Erik worked there.

Erik Wolpaw: Yeah, I think it’s a very different place now.

Anna Kipnis: For many reasons. We’re doing multiple projects now, which is actually a similarity with Valve. Whereas before it was one team. And also, the other difference is, we have our shit together [laughs]. We didn’t as much at the time. Pretty early on we were still trying to figure things out, as a company, while we had a really tough deadline for the project. This is probably little-known because [Erik] gets so much credit for the writing he did there, but his official title on the game was the gameplay programmer. That was his title at the company. It happened that we wanted Erik to write stuff as well. I had similar duties. We were expected to do a lot of stuff on the game. I think now, at Valve, they think you should work reasonable hours and have reasonable expectations.

Erik Wolpaw: Yeah. Valve’s good at maintaining a home/work balance.

Anna Kipnis: Double Fine’s definitely good at that now.

Erik Wolpaw: When I was at Double Fine, it was more top-down than Valve is. Tim was the fearless leader of the whole thing. There are some benefits to that, I find. When you work on something – probably any project, but especially a creative thing – for two years, at the two-year mark you’ve lost the thread of whether this is good or bad or what’s happening. It’s reassuring to have – if you’ve got a good leader, which Tim is – someone who says, “Yeah, this is good. We’re pursuing this vision that I have. I still believe in it.”

I guess this isn’t super-profound, but that’s why leaders exist, right? They do this. They rally the troops. And when there is no leader, there’s a little bit of everybody… You have to find your own motivation. It can be a little bit harder when there isn’t somebody who is that leader.

Anna Kipnis: Not only that,  Tim is also really good at reminding people why these projects are cool. Some of these projects are very long. Psychonauts, from the time that Double Fine started to the time it came out, that was five years. There were a couple of people who were on it the whole time. It becomes difficult to keep track of the fact that, “Oh, we’re working on a really cool project.” Psychonauts, I think we were still laughing at some of the jokes…

Erik Wolpaw: I don’t remember that. I don’t remember much laughing. [laughter]

Anna Kipnis: You weren’t laughing at your own jokes. You were laughing at…

Erik Wolpaw: It’s just hard. It’s hard to maintain perspective two years in, much less four years in to something. Double Fine’s structure, at least the way it was structured at that time, did help with that.

RPS: Right. That was another thing that fascinated me during your panel. You said that both you and Chet had playtested your game so many times that you started playing it in other languages so that you didn’t have to hear your own jokes anymore.

Erik Wolpaw: Absolutely. Both of us discovered that early on as a good way to still help with playtesting, but not have to listen to any of the dialogue anymore. Once we know it’s hooked up right,  we’re looking for other bugs, and so we test the foreign versions. At some point in any project, I guess – especially with lockdown – what you basically hear is all the flaws and all the missed opportunities.

Anna Kipnis: I still can’t play Psychonauts.

Erik Wolpaw: Yeah, I’ve never gone and played Psychonauts. I probably should. By now it’s been so long, I should just do it.

RPS: How do you not just get burnt out on the whole process after a while?

Erik Wolpaw: Mainly because you forget, really quickly. Because I’ve had the great, great luck to work with amazing companies and amazing people… I haven’t released a game that’s been a huge disappointment. You get that positive reinforcement when it comes out. I want to do that again.

It is exciting, and Anna can back me up on this, to ship a game. To get in that final mode where it’s almost done and you’re shipping it and everyone’s working together to ship it. There’s this communal thing that’s really exciting. You want to re-create that. It’s actually nice that Double Fine is doing the smaller games, because you get that shipping feel more often.

So yeah, part of it’s just amnesia. You forget about the difficult parts and you just want to do it again.

Anna Kipnis: Especially if the game is actually gonna do well. That’s all Erik knows.

Erik Wolpaw: That’s the awful thing… Well, to a certain extent. We haven’t released any huge, huge stinkers.

RPS: So you mentioned flaws and regrets that you pick up while playing your own games again and again. Are there any specifics on, say, Portal that you can think of?

Erik Wolpaw: I can always think of a line that could have been better. I’ll think of a line or a phrasing that would have been better if I thought of it that way. There’s always regrets. I can’t think of specific ones. But that’s probably true of anything. You pull the trigger and you can’t take it back. Eventually you have some regrets. You could have done something slightly differently. Having said that, I am willing to let it go. It goes out and people like it. It’s fine.

Anna Kipnis: The thing is, the design process on Portal was really meticulous. They tried to hone in on exactly what they wanted to do. It’s not like one of those things where they were constantly fighting deadlines or anything like that.

Erik Wolpaw: Yeah. We had enough time to really work it all out. Especially on Portal 1. And it was also a small team, so there weren’t as many moving parts as there are in some other games.

I think part of it is just the Narbacular Drop guys. Jeep [Barnett] and Kim [Swift]. They’ll tell you this. Every game that they tried to make before that was just too ambitious. They decided, “We’re going to take one simple concept and milk it for as much as we possibly can, because that’s something we can actually do with a small team.” I think that kind of laser focus on doing one thing well benefited the final product. We didn’t feel the need to have portals be this thing that… One level is portals, then we give you bullet time, then we give you a gun, then it’s back to the portals. It’s going to be all about this fairly simple concept and exploring what you can do with that.

RPS: Portal 2 was a lot more elaborate. There were tons of really neat options, but it felt a bit more disjointed to me. 

Erik Wolpaw: I’m not really sure. Is that true? I mean, the gels were… I feel like we did slowly integrate everything. Again, we never changed it. You have two buttons, you can lay down two portals. Everything you do is two portals. I think we did do a pretty good job with that. Again, people can disagree. But that was certainly in the forefront of our minds. Maintaining a simple interface and trying to explore some different things you can do with portals. Yes, we introduced the gels, but the only way to really manipulate them is through the portals.

RPS: We were talking about the process you go through whenever you’re coming up with ideas, what sort of writing you do, what narrative techniques you end up using. There’s the song at the end of Portal. Where’d that come from?

Erik Wolpaw: It was Kim and I. Kim actually started it, because she mentioned Jonathan Coulton. Like, man, it’d be cool to find some way to do something with him. Just like a lot of ideas, that sparked me to think, “Oh, man, God Hand had a song, and that was so great.” And then Kim and I started talking. We’re like, “Oh, we could put a song at the end of Portal!” If it hadn’t been for Kim finding Jonathan Coulton, there would have been no song.

RPS: So it was just a wouldn’t-this-be-cool moment?

Erik Wolpaw: Yeah. Not planned. Just opportunistic. A couple of ideas jell and this could fit in the game. We could actually do this.

RPS: There’s that experimental side of Valve, and then there’s your focus on extreme polish and making sure something’s nearly perfect before it comes out. Basically, you’re working on two opposite sides of the development spectrum. How do you balance that out?

Erik Wolpaw: It’s case by case. Playtesting will work some things out. In the case of the song, it really didn’t. It was kind of our own intuition of what was working and what wasn’t, because the song just kind of appeared, basically. We weren’t playtesting the song. But there’s lots of ideas that you can get out in front of people pretty quickly and start getting feedback about whether this is working or not. You can fine-tune it.

At some point, you just have to abandon it if people are looking at it, and not getting it or… You start to be able to read the playtests and figure out that it’s not actually working, whatever the idea is. Some of it’s just being a responsible co-worker and not pushing things. Not making a big deal if it’s pretty clearly failing and failing and failing, and you can’t articulate – either in words or in some sort of playable piece of game – why it will work. It’s something that happens, but it tends not to become a problem. And again, I think largely because of the pretty strict playtesting regimen.

RPS: Another thing that both Double Fine and Valve are doing that’s really on the forefront is all the community stuff. What I find fascinating is that you’re approaching it very differently. With Double Fine, obviously, you guys did the Adventure and you Kickstarted it – basically making Kickstarter this huge force for games. And then with Greenlight, Valve is taking a similar core idea and removing the funding aspect.

Anna Kipnis: What I can tell you about the Kickstarter thing and working on the project, I realized… You never actually see what it’s like to go through the process. How shitty games are until they came out. We wanted to involve people in it, to maybe even help steer it in a certain direction with their input. We really wanted to show people what it’s like to have a game be in production and what goes in to making it. Because I think people just don’t know. Maybe if people knew they would be more inspired to make games themselves. They would know what we go through to create demos for E3 or something, to actually finish it.

In general, just demystifying the whole process, I think, benefits everyone involved. With Greenlight, it lets people steer the direction… I don’t know how to put it. But people can show interest in a game that maybe even Valve didn’t know would be really successful. That was the benefit of Kickstarter. They didn’t know that something like that existed, and that they could directly fund the game. Somebody with a lot of money didn’t have to decide for them what games they would want to play in the future. They could directly get involved. I think Greenlight is along those same lines. Except of course people aren’t paying for it.

RPS: It seems like some people are wrapping their heads around how development works, though. I mean, you still have large groups that rage vomit all over their keyboards any time they see a bug or glitch, but you have others who got in on the ground floor of stuff like Minecraft.  

Erik Wolpaw: Obviously, I did see some e-mails floating around that people are just putting up jokes, basically. But the Workshop stuff – people submitting items for TF2 or DOTA – that’s part of the reason we crowdsource it. The community quickly ferrets out what’s not real. What’s a joke and what’s real. [Note: This interview was conducted before Valve announced its $100 initial Greenlight fee.]

Anna Kipnis: I think, though,  that the general public now is still pretty unaware of what actually goes into making a game. I still feel like if someone posts an early demo that’s not polished and it doesn’t show well, people won’t be interested. Even on Greenlight. Maybe Greenlight will work towards reducing that problem, but I’m not sure. Even with Kickstarter, you need to have a pitch video that shows your game well or shows your ideal really well.


  1. Lemming says:

    Very insightful, thanks!

  2. Ultra Superior says:

    My name is Gabe Newell.

    I hire the best talent in the world. I have the world’s most amazing developer studio, where the great minds of gaming industry gather to sleep for years without making any game.

    They’re free to make hats, from time to time. Unless they’re too tired.

    When some of my many employees wake up from their slumber, they are allowed to pick a game they want to have made, and I buy it for them.

    That’s how we develop things. Put your dreaming hats on, my children. Float in the half-life of eternal slumber.

    For grandfather Newell loves you all.

    • hello_mr.Trout says:

      well, it’s a good thing you’re around to keep them grounded :/

    • zeroskill says:

      Games Valve released since 2004:

      Counter-Strike: Source 2004
      Half-Life 2 2004
      Half-Life 2: Deathmatch 2004
      Day of Defeat: Source 2005
      Half-Life 2: Lost Coast 2005
      Half-Life 2: Episode One 2006
      Half-Life 2: Episode Two 2007
      Portal 2007
      Team Fortress 2 2007
      Left 4 Dead 2008
      Left 4 Dead 2 2009
      Alien Swarm 2010
      Portal 2 2011
      Counter-Strike: GO 2012
      Dota 2 2012

      Blizzard since 2004:

      World of Warcraft 2004
      WoW: Expansion 1 2007
      WoW: Expansion 2 2008
      StarCraft II 2010
      WoW: Expansion 3 2010
      Diablo III 2012

      • Ultra Superior says:

        Valve is good at releasing games. Not at making any new games of their own. Real titans amongst men.

        • zeroskill says:

          Yeah, I also aboslutely despise how Valve hires people from the community (modders, students attending the DigiPen Institute), giving them jobs, and putting world class artists and writers by their side so they can make the game of their dreams.

          I must be an utter and complete moron.

          Valve should take example by Blizzard, who let their world famous modder of Warcraft 3 (ehm..Icefrog) live off donations for over half a decade while still selling copies of Warcraft 3 based on the success of Dota, until a company that isn’t led by complete idiots actually hires the guy. Yes they should do more of that indeed.

          • Ultra Superior says:

            Why would you despise that ? I think it’s great.

            However is that mutually exclusive with making a new installment of a game serie that their fans really want? Does that exclude any chance of creating a new interesting, ambitious IP?

            I mean, who else should be leading the charge, if not Valve. Instead, lets just reiterate counter strike!

          • ResonanceCascade says:

            There is a bit of a point buried deep in there, though, which is that Valve would really benefit from shipping a fully original game in the near future. The balance has fallen too far to the “re-do stuff from the mod scene or buy a company that is making a cool game and release it ourselves” end of the spectrum.

            People carry on about Half-Life, which I’d also love to see, but what about a crazy new game to blow our minds once again? They’ve got people at the company who used to specialize in that kind of thing.

          • MadTinkerer says:

            @Ultra Superior: I’m 99% sure zeroskill is being ironic.

    • JackShandy says:

      Valve is not your bitch. I can’t believe this needs to be said so often.

  3. JackShandy says:

    Great, honest interview.

    And man, it boggles my mind how shitty games are before they come out. Like, a painting, you do a sketch, if it looks good you start the real work. An AAA game, you’re working over two years before you can see that sketch. It’s impressive that anyone can make anything.

    • PleasingFungus says:

      The thing is that the graphics aren’t the core of a game – the play is. You make a prototype, a sketch of a game, and if that’s fun – unpolished, obviously, but fun! – then you make the game. It’s going to be hideous, but you’re not basing your judgement off whether it looks good – you’re judging whether it plays well. So, that’s a better analogy, I think.

      • JackShandy says:

        Sure, but making a proper software prototype takes just as much time as making the graphics. Getting portals to work perfectly is hard. Not only that, but story-centric games; so much of it is dialogue, timing, voice-acting, the way it fits in with every other part of the game…

        I’d imagine Portal wasn’t funny until late in development.

      • lumenadducere says:

        You’d be surprised. I was in QA at Sony for…8 months or so (last September to May of this year) and sometimes gameplay is incredibly shitty for a long, long time. It takes a tremendous amount of balance tweaking, level design changes, and AI adjustment to get gameplay to actually be fun. And this is on titles that are generally considered to be a critical and commercial success, let alone games that aren’t regarded as being good.

        Honestly the biggest thing I’ve learned since then is that the general gaming populace has no idea how real game development works. It’s an annoying process full of adjustments and changes, and from a QA standpoint nobody listens to you and it feels like the work you’re doing is worthless. But at the end, when things just click and you see titles start to come together…it’s a really oddly satisfying feeling. Even in the testing bays where we loathe the game we’ve been seeing for 8+ hours a day every day for months on end. It’s satisfying just seeing a bug that you wrote a report on being fixed, so I can only imagine the feeling that the actual devs get.

        Before that point, though, games are such utter shit that nobody in their right mind would enjoy them. And this is in every way possible, from graphics to gameplay to story. No one aspect is worse than the other – they all suck in equal amounts. I think it’s one of the only mediums in which the final product goes through such radical changes in the span of 6 months or so.

        Needless to say, I’m no longer at Sony as a tester. There are many reasons for that, but one of the biggest is that you have to deal with such a bad product for so long that by the time it actually gets good you hate it with every fiber of your being. No tester I know has ever played a game they’ve worked on once it’s been released, even though we all get heavily discounted (or free, if you’re in the core testing group) copies. Those all wind up going to Gamestop for trade-in credit. So if you’re in the San Diego or Foster City areas in California and you see used copies of Sony titles at your local Gamestop on the first day or two of release, you can likely blame a QA tester at Sony for that. I’d imagine it’s similar at wherever the Sony offices in Europe are as well.

        • Skabooga says:

          Hmm, most fascinating. I always imagined playing games professionally would drive me to develop and aversion towards them, don’t even want to think about what it would do if it were a crappy, unfinished game I was playing over and over again.

        • dragonfliet says:

          This is also why novelists get that tired face when you bring up their book that just came out (even though they HAVE to talk about it for sales), but will light up if you ask them what they’re working on. They’ve worked on that damned novel for years, most of the time with it being garbage, and by the time it gets good and is polished enough for release, they’re woefully sick of it. And by the time the publishing mechanisms get around to, you know, getting it in print, they’re six months (or more) into their latest project that they actually care about.

  4. baby snot says:

    I loved Portal 2. The second act is wonderfully, tragically haunting. But the gels still felt shoehorned in. And no amount of great writing can obscure that.

    • Skabooga says:

      I am foolishly enamored of the gels if only because of the delightful changes they instigate into the music as you glide around on them. I spent a couple of minutes doing nothing by just zooming on the gels the first time I encountered them for that reason alone.

    • absolofdoom says:

      I completely disagree, I thought the gels worked well and were a welcome addition.

      • baby snot says:

        I agree that the gels worked fine and were a fun addition as you say. But a fair bit of exposition went into explaining their part in the story of Cave Johnson and Aperture Science in order to obfuscate the break in gameplay continuity between the first and second games.

        • dragonfliet says:

          While I can see where you’re coming from and understand that there is, indeed, some sort of continuity break, I didn’t feel like anything was shoehorned. Since Portal is a series that exists for two reasons (puzzles and comedy), it was a perfect mix for me–new mechanic means an opportunity for a new set of jokes, and by golly, Cave Johnson was awesome.

  5. Totally heterosexual says:

    Love these talks.

    Though I will have to bitch a little about the fact that Chell looks too young in that picture.

    • Skabooga says:

      I’d say mid-to-late twenties, which is around the age I presumed Chell to be (physiologically-speaking). Do you see the picture as looking younger, or is my estimation of Chell’s age off?

      • Grey Ganado says:

        Her mind is also most likely that of a grade school student, since she got captured on the Bring Your Daughter To Work Day where she presented her science project which can be found in Portal 2.

      • Totally heterosexual says:

        From what i know, Chell is 30-40. Thats how she looked in portal 1 but in portal 2 they suddenly made her into a supermodel and that pissed me off.

        Ok maybe that picture is fine. maybe im just throwing a hissy fit >:X

  6. tomeoftom says:

    This was great.

  7. Deston says:

    This was an excellent discussion Nathan, very interesting to read. Such candidness is always refreshing.

  8. Shralla says:

    lol Erik Wolpaw. Even the snack room girl won’t talk to him, why would you want to?

  9. JoeX111 says:

    Man! I was so psyched when I read this:

    “We were talking about the process you go through whenever you’re coming up with ideas, what sort of writing you do, what narrative techniques you end up using.”

    I love it when people talk about their creative process. Especially writers. So open ended. So worthy of discussion! Then the question became:

    “There’s the song at the end of Portal. Where’d that come from?”


    • dragonfliet says:

      I think that question is actually really worthwhile. Not only is it one of the high points of the game, but it comes out of freaking nowhere. It is interesting to see where such a left turn comes from, creatively, and the answer was just as interesting as any other discussion of writing, imo.

  10. Bremze says:

    Gene’s Rock-a-Bye is indeed a great song that fits in perfectly with the tone of the game.

  11. The_Great_Skratsby says:

    That was great.