Level With Me, Alex Austin

Level With Me is a series of interviews with game developers about their games, work process, and design philosophy. At the end of each interview, they design part of a small first person game. You can play this game at the very end of the series.

Alex Austin, an indie developer based in Berkeley, CA, has steadily pumped out physics sim games since his original Bridge Builder in 2000. More recently, he made the first person sport Hockey? and vehicle strategy game A New Zero ("Red Baron meets Tie Fighter meets Mechwarrior 2 meets Joust") -- and now he's refining Sub Rosa, an ex-7DFPS about tense deals, betrayals, and car chases.

Robert Yang: You're based in Berkeley? I feel like the Bay Area indie scene keeps to themselves more than the others.

Alex Austin: We've been doing more meetups -- Tim Keenan's been organizing that. Before, it was, yeah, everyone doing their own thing, and now we try to have more weekly things. Not everyone shows up, and we're still missing quite a few indies, but...

RY: Screw 'em!

AA: [laughs] We're too good for them!... But the meetups are great, it's nice to test out each other's games, it's really useful.

RY: I wish the indie scene here was more dev-centered like that. In New York, it's more about events and talks sponsored by universities. And if they don't run events, we don't end up seeing each other.

AA: It must be hard to be indie in New York, with the high cost of living there.

RY: It's often a pretty awful place to live.

AA: We should just get all the indies together, move to Wyoming, and live like royalty.

RY: How long have you been an indie?

AA: Almost 12 years? Ever since I released Pontifex, the sequel to Bridge Builder, in 2001-- where you, uh, build bridges, with some physics and all that.

RY: How can you make something like that? What were you doing before?

AA: I went to junior college for a bit. And then I just sort of bummed around and worked on games. Then I got a job in Santa Cruz, programming FPGAs. After 3 months they laid everyone off... but the severance money, along with the popularity of Bridge Builder, let me do indie games since then, though sometimes I'm barely getting by.

RY: But you have an office in Berkeley! That's pretty great.

AA: It's been close, but fortunately I haven't had to do any outside contract work. Gish got on Steam, and on the (first!) Humble Bundle, which did really well.

RY: Wow. I can't imagine selling games to support myself. I have to teach game development as my day job.

AA: As far as day jobs go, teaching sounds pretty good.

RY: No, you're right, it is.

AA: Imagine working at a game company. It'd be pretty hard to do my own games on the side, I'd get pretty burned out.

RY: Hmm. Last I checked, UC Berkeley didn't have any games classes, but I think you should look around and teach, you'd do really well with that.

AA: [chuckles] Yeah, but uh, I don't have a degree or anything...

RY: Oh, right. That's so stupid, the whole degree thing. You're qualified.

AA: Well, I graduated back in 1994. There weren't any options for game design. I had a friend studying programming, but they had him doing 16-bit assembly, which even back then, was already pretty worthless. [laughs] So I learned a lot more on my own.

RY: That's what I found too, I learned a lot outside of school.

AA: What exactly do you teach? Game design?

RY: Right now, I teach more the game dev side of things. Unity, C#, Maya, Photoshop. Sometimes I have to teach them how to use computers.

AA: [laughs]

RY: Some students didn't grow up playing video games, but now want to get into games as a career. The other day, one wanted to make a "shoot a ball into the basket" game. They said sometimes they want the player to get the ball in, and sometimes not. And I'm like, "um, okay?" [pause] Then they asked, "so how do I do that?" ... I was speechless. How do you do that? You make the game and play it and tune it! I thought it was so obvious.

AA: Yeah, I think game design is one of those things that people think they're good at, they can do it. One thing I tell people is that it's way more about getting rid of bad ideas, than it is about coming up with ideas. Ideas are the easy part. The hard part is testing and getting rid of ideas that don't work.

RY: That's what I try to teach, but it's harder than I expected. How would you teach that?

AA: I think it just comes with experience. You can't tell someone their ideas don't work because, in their head, it totally works. It's something they have to implement and find out for themselves.

RY: Now, the first Cryptic Sea game I played -- and I didn't know it was yours? -- was "Golf?" It felt really fresh.

AA: [soft chuckle of sudden recognition] Oh yeah. That's a game I still want to finish at some point, but it's one of those "lost in development" things. That was 9 years ago? I learned a lot from working with Luke Hetherington, the artist. Before, I was very focused on the programming side. "Okay, I implemented normal mapping, so now everything should be shiny." But when working with him -- I'd put in bloom, and he'd crank down the bloom to be super subtle. People might not even think it's there, but it gives the game some really subtle soft shadows. It's not just about the technical implementation, it's also about the overall look.

RY: Yeah, restraint is a hard thing to learn. I feel like I always put too much color in things.

AA: I'm mostly been doing procedural generation lately, and it's definitely a trial and error process to keep things simple. I think I'm getting better at that.

RY: I like the simple art in your games. Somnia seems like an outlier though, with normal maps and stuff...

AA: I worked with an artist on Somnia, Aimee Seaver. It was her first game. If we had more time, I think we could've developed a more cohesive art style for it.

RY: Really? I thought it was cohesive. What's un-cohesive about it, to you?

AA: Some of the stuff doesn't mesh that well. The blocks are just plain, normal-mapped, 6 sided cubes. We did the whole thing in about a month, so we rushed it.

RY: I thought the demo was pretty good... I bet a lot of people say it reminds them of Shadow Physics?

AA: [laughs] I actually demonstrated it at the Experimental Gameplay Workshop a year before Shadow Physics, and they did theirs in 2009? I mean, they gave up on theirs, but I think Somnia's version kind of works -- specifically how you can transfer velocity between the 3D and 2D world.

RY: I liked that momentum transfer. I liked how the shadow platforming involved gravity, and it was like orbiting around the light. Like, light was a physical force? That was a good metaphor.

AA: I want to go back to it sometime, but I also have 16 other projects to finish too, last I counted.

RY: I have like 14 different Unity projects sitting on my desktop. It's hard to finish things. What helps you finish projects?

AA: I'm probably the worst person to ask about this. I haven't really "finished" a game (like, to sell, definitively) since 2007. My main motivation right now is running out of money. I have an album of 2D games that I'm finishing: a simple multiplayer skateboard game I made for TIGJam where the winner is just the last person to crash, a Tetris-with-physics game, a top-down RC Pro-AM with realistic physics game... I don't really like the polishing side of development, I'd rather experiment and expand gameplay. But I need money.

RY: That's the story of how Cactus sat down and pumped out Hotline Miami, right? He needed money to, like, live.

AA: It's a good motivation. If I had a patron, I don't think I'd take my games very far. The pressure makes you popularize the game a bit. If I didn't have that, I'd just make games myself, and never polish them.

RY: The other thing I noticed about Somnia: it's one of your very few single player games. Are you allergic to single player?

AA: There's just tons of potential with multiplayer. The album thing is actually single player games with multiplayer modes. But I'd say for me... I can't stand linear story-based single player games. The ones where dying means going back and redoing the thing I just did -- it makes me want to quit.

RY: Is a multiplayer game that different? I was trying to play CS:GO and I couldn't play more than a few rounds, it felt so repetitive.

AA: I guess I don't play too many multiplayer games either... the ones that are repetitive. Single player games... most of mine are more old school? Part of it would be -- any 3D game, it's hard to make enough content. I'm not an artist, I can't fill it with content.

RY: To me, game content is the appeal of a game. Code has to repeat, as a closed system. I think I make games more as an excuse to make places. Mechanics matter too, but to me, that action is about relating back to the place.

AA: Yes, I like worlds to explore too, and the simulations. A lot of my stuff is physics intensive. I've been working on economics simulations too, I think that's an untapped subject in games. It's crazy to look at the games from the 90s, and how their economics simulations are so much more advanced than what we do now, despite how much more processing power we have today.

RY: Economics in games... scares me. What's the appeal for you?

AA: Well... in A New Zero, all the islands have different types of goods they generate, and you can trade them back and forth. That's the basic stuff. Have you ever played Seven Kingdoms? I stole their resource system. But I ran into the same problem that hit that game -- where after a certain point, you just built giant armies of automated cannons. Eventually money becomes worthless. And in A New Zero, sometimes one team will be down to one island, and their economy should be collapsing and they're pretty much doomed, but they have enough cash to keep making vehicles. So in the next version, there will be workers to build the vehicles. You're selling goods to the workers, in exchange for their labor.

RY: That sounds conceptually interesting to me... but this sounds like the type of game design that involves going to Excel and balancing spreadsheets. I respect that but it scares me.

AA: Most of my stuff is pretty simple when you get down to it. Even the physics that I do is relatively simple. My design approach is more brute force, trial and error. I let it run and see what works. If this guy's generating consumer goods, and this guy's generating tanks, how does that balance out? One thing I want to do is this "communism vs. capitalism" thing. Where one side can change the prices around and pay the workers different amounts, but that seems really hard to balance... maybe if you're on the capitalist side, you starve the city a bit to get the food prices up.

RY: It'd be hard to reconcile this with the actual economic theories. "That's not what communists would actually do! That's not how capitalism works!"

AA: I think communism would mostly look like a small group of people at the top, deciding everything. I wonder how I'd decide which players would lead.

RY: So is A New Zero more about this, or is it more about the vehicle simulation stuff, to you?

AA: It's about all of it. I'm trying to abstract World War II. I want to create opportunities for the economy and other systems to interact -- for example, I'm thinking of adding "honor points." Each time you die or your civilians die, you lose honor points. The only way to earn honor points is if your side wins. So maybe you would bomb the other side's civilians, and they would lose points, but then they would revenge-bomb your civilians. Both sides end up in this death spiral, everyone loses a lot of points; so if you're losing already, maybe you'll just want to surrender, instead of losing those honor points?

RY: I think that sounds great... but I don't think that emerges for a while? When I was trying it the other day, A New Zero was more about getting into a plane and crashing into the ocean.

AA: The planes are pretty hard to fly. In the next version, there are more options like being infantry or tanks, those are easier to control. I want to add infantry because right now you takeover islands by building a capital building on it, and that's kind of underwhelming, so I think adding infantry lets you actually conquer it instead of clicking a button in a menu. And that's when I started thinking more about human physics.

RY: The human physics in that and Sub Rosa were interesting.

AA: It's basic but I think it's important, especially for FPS games, assuming it works? In most of these games, you're a capsule with a gun turret, right? That influences the gameplay because you turn around instantly 180 degrees and shoot someone. That cancels out a lot of tactical considerations about flanking. Did you ever play ArmA II?

RY: Briefly, before it instantly scared me and confused me, yes.

AA: Oh, I don't play it much either... but I watch ShackTac videos. You know, it's a bunch of players from ShackNews who form squads and create their own missions. Like, 100 people doing one massive mission against the AI.

RY: Kind of like wargames scenarios.

AA: It's one of the cooler things I've seen in games. Sub Rosa was based on one of the ShackTac mission types, "Dark Business." In that, a resistance team wants to trade some US hostages to the Russian team for an ammo truck -- and the US team doesn't have anything to trade for the hostages, but they have a helicopter to find out where the other two teams are doing the deal. The dynamics were really interesting one time, when the resistance decided that they were going to shoot the Russians no matter what, and they had a code word for the shooting to start. It's three teams with different goals.

One of the problems in Sub Rosa right now is -- well, team-killing -- but also just bad teammates in general. So I think I'm going to add more than 3 teams, and everyone can just setup their teams how they want. Maybe the major teams just negotiate with each other while the smaller teams do other things? I like the idea of different people doing different things in the city, sometimes you let them go and sometimes you pursue a blood feud.

RY: Yeah, I like the scale of the city. I think that makes it work.

AA: I'm also adding AI traffic, so maybe you can blend-in with AI guys. It's also just fun to weave in and out of traffic, and create some really cool chase scenes and scenarios. One of my goals with A New Zero is to encourage some of the other ShackTac scenarios to emerge, but for a broader group of people, and have that emerge from the game. ArmA II has a lot of funkiness to it, with its limited physics model. You can drag your buddy's body if he's unconscious, but he'll lay totally flat even on a hill... and if you have a better physics model, that scenario can work better. And what's great with a "human physics" approach is it's automatic; gun recoil moves your hand, which moves your arm, which moves your body.

It feels better than other games, I think. Everything else feels like a hack. One of the things I've been doing with A New Zero's human physics is: running, then leaning back, then leaning forward and jumping through windows. That's something you can't do in a lot of first person games, but we can actually model the diving and rolling.

RY: That sounds hard for players to get.

AA: As far as controls, I'm trying to keep them similar to basic FPS controls. You balance automatically. But there are some things you have to get used to. In games right now, you can just run into a wall and slide along the wall until you go around the corner -- but when you have physics, your arms crash into it and then your head crashes into it and you collapse, and it takes a while to get used to that. You feel more vulnerable, more like a person.

RY: We won't be running down stairs at full speed.

AA: Yeah, that's not even physically possible! You'd just fly off the end and crash. So it's a slower pace, but it's interesting to move like this.

RY: I think it'll be great if you manage to de-program all of us from expecting Quake movement physics... which is why I really like what you're doing with "Hockey?"... there's a strong sense of body and weight, but without the stupid body awareness we see in AAA games where you look down and see an animated torso appendage.

AA: Yes. In Hockey?, there's nothing moving around that you're not controlling. I remember in Battlefield 3, you can see your legs do a special animation when you jump a fence...

RY: And that felt weird. It wasn't my body. How did you nail the feel for Hockey?, do you play hockey?

AA: I've played floor hockey, but not ice hockey. I've never actually ice skated before.

RY: Ice skating is hard. It's so hard. I'm terrible at it.

AA: But the artist, Luke Hetherington again, he's played some ice hockey. (I believe it's required, up in Canada.) We actually made an early version while working on Golf, 6 or 7 years ago, but it totally didn't work, it was too hard to move around and it was too clunky. Then we decided we should revisit it, and I always thought first person hockey would work really well, like mapping the hockey stick handling to the mouse. There's no "shoot puck" button, everything you do is you doing it. A lot of players are coming up with stickhandling moves that we didn't even think about.

RY: And you get all the emergence "for free", just by simulating the physics of hockey.

AA: Exactly. With this, and with A New Zero, I hope these games sort of design themselves, that way. For Hockey?, we're working on a new update with more detailed player models, and you'll be able to see each others' arms, instead of a hockey stick floating out there. The body checking feels a lot better now.

RY: And that's a skill that players develop? Very kinesthetic

AA: There's been a Russian league going, and these guys are insanely good. Sometimes they'll join the North American servers with a 300 ping, and they'll still dominate the game. I don't have any client-side prediction in there. They're so good, they have the feel, that it doesn't matter. It's also been good with the North American League and Friday Night Hockey... it'd be fun to re-create a "US vs. Russia" match-up, like in the 1980 Olympics.

RY: I feel like your games should be exploding on the e-Sports / competitive gaming scene, they're so skill-based and strategy-based. But maybe it's because they don't look like typical e-Sports games?

AA: I think the new player model will help people see what's going on, make it more watchable. Also, this game's still pretty hard to learn. The learning curve is a vertical line.

RY: But DOTA and stuff is like that too?

AA: DOTA has all that momentum going for it.

RY: It did take several years to build that up...

AA: I think players are organizing and building that up themselves. Too many games focus on tutorials to teach you everything about the game. Isn't that the point of playing? Aubrey Serr, at Wolfire, says it's like a book teaching you to read itself. You should know it, you should figure it out. But we will add a practice mode for Hockey?, like dropping a puck and practicing shots...

RY: That fascinates me in a conceptual way. It's like a "real" hockey player, going to the rink and practicing their slapshot for 3 hours.

AA: The league players do have practices.

RY: Oh, and coaches are there too? Shouting things at them? And then they argue and fight... One thing that's always fascinated me about hockey is the sanctioned fighting, how fighting is a mechanic, like part of the strategy of the sport.

AA: We want to add fighting to the game, in some form... but then we'd need a penalty system.

RY: And then you'd need a referee.

AA: I think we're going to need a ref anyway, because now you can hit people in the face... in hockey, fighting is sort of a way to stop people from trash-talking, and that's a problem in online games, right? I like the idea of beating someone up and then they can't chat anymore.

RY: I like that you're willing to take risks with this. If Valve were making this, they might be saying, "no, we want everyone to have fun, all the time, always." Experimental e-sports!

AA: The core ideas in my games are always kind of simple, but yeah, I don't try to please everyone with these games. Hockey?, especially, is pretty disorienting because other players can push you around or tackle you without you doing anything, so your camera rolls a lot.

RY: If I played it on an Oculus Rift, I'd throw-up after 5 seconds.

AA: Lots of people have told me they want it, but I think it'd be one of the worst games for the Rift.

RY: It could be like the physical conditioning of actual ice hockey players, right? You have to train yourself not to get sick.

RY: Okay, so tell me what to add to this first person game, and I'll try to do it.

AA: I'm picturing... some sort of alien thing, down in the water. Maybe some sort of hatch.

RY: What would the hatch look like?

AA: One with a wheel that you turn?

RY: Like a 20,000 Leagues Under the Sea thing?

AA: Yeah, and then it leads into a decompression chamber, and you open another hatch to get inside the ship. There's a lot of machinery. And you find your way to... the... control room.

RY: What does it control?

AA: [laughs] I don't know.

RY: What color is this? The chamber and such?

AA: Hmm. Bluish-gray. Metal. And once you're inside, some red lights. Then you get to the control room... and it's the control room for the entire planet. This is getting into "Lost" territory? But you're descending to the middle of this small planet or moon.

RY: What do we see in the control room?

AA: Big levers.

RY: How big? Size of a person?

AA: Giant levers. Yeah. And some sort of screen, or portal, or... a periscope thing.

RY: What do you see in the periscope?

AA: ... Space. You see the planet you're orbiting around.

RY: What does the other planet look like?

AA: The other planet is like Mars. Orange and dusty. And you can move it around.

RY: Move what around? You can pull yourself toward Mars? Using the lever? How will the player know what does what?

AA: Trial and error.

RY: So the levers control the...

AA: ... position of the moon. And there's another planet, that's more habitable. And you can move closer to one or the other with the lever. It's a gondola type thing, a forward and a backward. One goes toward the Earth, one goes toward Mars.

RY: Okay. Then what happens?

AA: Then once you get close enough to a planet, you can go back out the hatch, and somehow go to that planet? [laughs] I'm worried this is too much to implement.

RY: It seems slightly doable for me. Are you a good judge of how much work you can do?

AA: I always overestimate how much I can do. When I started making games back in the Quake 1 era, I planned a huge first person game, and coded a software rasterizer and all that stuff -- but I never had a chance of finishing that. I was off by a factor of 10. But now, I'm usually only off by a factor of 2 or 3.

How long do you think this will take?

RY: I think I can do the environment and art pretty fast, but what will take me a while are the interactions -- the interaction for the hatch, the periscope, the actual implementation. First I'll get excited about making something, then I'll just slow down to a crawl as I realize what I have to do.

AA: I try to find something inspirational to work on, then do some of the crap work before that. "If I get this in, then I can do this later."

RY: It reminds me of an Ernest Hemingway quote; he was giving some advice to an aspiring writer, and he said something like, "when you're excited and know what's going to happen next, then stop." Because then it's easy to start working again, later.

AA: One thing that I do now: if it's past 5 PM, and I finally get something working, then I stop. If I keep going, I find out that it doesn't actually work, and then I'm there all night, trying to fix it.

RY: And you get really angry. Like you have to fix the bug, to "win."

AA: Yeah. But maybe it's better to stop when you're excited, to think about it for a bit.

RY: Thanks for your time.

This transcript has been edited for clarity and length.