The Happy Medium

Adam Saltsman’s Canabalt

Adam Saltsman’s Canabalt

Of course you have five lives

Adam Saltsman, otherwise known as Adam Atomic, is a game designer based in Austin, Texas. He’s most famous for having created Canabalt and Flixel, but he has created an impressive range of games that reveals the mind of a hyper-critical game thinker.

Saltsman was kind enough to speak with me via Skype for over an hour of his time. Though the entire transcript isn’t published here, it’s still one of the densest interviews I’ve ever conducted. Saltsman speaks about video games in a thoughtful, academic tone. He has no critical boundaries between music, books, movies, and games, which I’m sure is the reason his games are so consistently rich in tone and subtle detail.

THM: So I have an old book that I think you may be interested in. It’s called A Strange Manuscript Found in a Coppyer Cylinder, by James De Mille.

Saltsman: I haven’t read the book, but I’m familiar with the premise!

The author presents an outlandish story about an underground world, but claims he found the story in real life and he’s simply presenting it to the reader. It struck me that you would be interested in this technique. I feel like a lot of video games take a similar advantage of people’s preconceived notions of the medium.

Yeah, definitely. 

Is that why you relate to  a concept like that?

I think there are two things that I really like about it. There’s a specific conceit which is, “I’m the writer but I didn’t write this, all I wrote is the forward and I found the rest of this, which someone else wrote.”

You hear a lot about creators—like, Jon Blow talks about how in Braid and The Witness he sometimes feels more like he’s discovering the game rather than designing it. And Marc ten Bosch, who’s doing Miegakure, he describes game design as a similar process. And you hear sculptors and you hear authors talk about that.

There have been a lot of games, these “de-makes”, the term that Phil Fish coined, I think those are an actual video game equivalent of this. “Oh, I didn’t make this game, this is an old DOS game that I found and ported,” as an excuse to engage in some sort of deliciously retro, nostalgic gameplay design. I think that’s really similar to the literary technique of the “discovered manuscript”.

But I also like the sense of mystery that it implies. Not only am I, the story teller, not in this story, I’m reading it for the first time. The author just leads you blindfolded to the front of the garden and says, “Ah! here we go!” I like that feeling.

A lot of procedural games do that same thing. Like Minecraft. Notch simply leads the player to the world and says, “here you go.” Canabalt is like that, as well.

I think the types  of pleasure you get from Minecraft and you get from Canabalt are pretty different, but at the same time there is a shared desire as the creator to… on some level it’s almost selfish. I think Derek Yu has talked about this with Spelunky before. There’s a mix of greediness and selfishness, and selflessness that is really unique in a procedural game. If you’re making a content-based game, you know all the secrets, you know everything that’s in there, so maybe the game’s not all that fun for you to play. It’s maybe only fun for your audience to play. And that’s okay. But if you’re selfish, you want an exciting, interesting game to play and nobody’s made it yet, building a procedural one where you don’t know what’s going to happen yet, and you don’t know all the surprises, and you’re on the same playing field as your audience.

I also like that there’s something else going on in Minecraft and Canabalt. It doesn’t have anything to do with proceduralism, it just has to do with letting people imagine parts of the game for themselves. I, as a designer, would like to play my own game and have it be as fun as if I was part of my own audience. But at the same time I want the audience to have as much fun as I had as the designer. It’s really fun and exciting for me to imagine the whole game world. And why would I take that from them by spelling it all out? Hemingway called it the Iceberg Theory. The art of implication. Just give people the little pieces they need to imagine up the other half of the game. It makes [the player] more invested and it makes your work more adaptable and more rich, too.

I suppose that’s what’s going on with Canabalt. You insinuate what’s going on in the background and people have wild imaginations based on whatever they are doing in the game

Yeah! In previous games it was more of a process of thinking up the whole story and the whole scenario and having this rich vision in my head, and then going through and actually deliberately picking out which piece I wanted to share with people. And in Canabalt there was no richer more fully developed thing. I feel like I was able to get the effect that I wanted, engage people’s imaginations, without actually building everything behind it. Which I think is actually risky. I think that it’s important to have a solid, deep world sitting just under the surface that you don’t tell anyone about. I think that people can sense that. 

There’s that sense in the old Wipeout games for the Playstation that it’s not just a racing game about triangles. The world feels fully realized. None of its rubbed in, but you start poking around and you realize—oh, there’s a whole racing league and there’s different corporations funding the teams. It gives it this real richness that I totally appreciate. I think—either in certain cases or from certain creators, you can create the same effect on the player’s psychology without actually realizing the whole world underneath. If you can get away with it, that’s cool because it saves you a bunch of time.

Going back to the “found manuscript” trope, it feels like the novel had been around long enough at that point that people were familiar with the format and an author could subvert expectations. Do you think video games have been around long enough that designers can start doing the same?

Absolutely. And that’s been happening to some extent already [in video games]. Kojima’s famous for setting you up to expect things that he invented years before and then knocking all those dominoes down.

This is one of these fundamental things about entertainment. I think every time you read a book or watch a movie there’s this parallel process to you simply watching the film or reading the words. Your brain is perpetually constructing new expectations and then having them bear or out having them turned on their head. I don’t know if we consciously think about that in game design much. I don’t’ think I think about that.

It’s a little tricky because it’s so active and it’s so interactive. In a movie if you turn things around on somebody, they don’t have to learn new controls or get frustrated when they die because it was unexpected. They’re just there in passive and the thing kind of soldiers on. I think it can be done and it should be done, but I think it’s something that we have to play with and figure out how do we do [video game] versions of that.

I can think of two examples of you subverting player expectations in your own games. Your game FATHOM starts as a platform shooter, but then turns into a cerebral exploration game. And Happy Puzzle Party, which looks like a normal action puzzle game, but presents vague, unwinnable goals. Happy Puzzle Party comes across as a cruel trick on the player.

Yeah, that’s a really cynical game. If I have any regrets about anything I’ve built, that’s probably it. It’s just sort of mean. And I don’t think that makes it wrong or untrue or anything, but there’s already a lot of meanness in the world.

FATHOM, the whole idea with that was to set up and subvert expectations. And expectations that come from a sort of game that I like. I really like Mega Man X and I really like Cave Story, and what if something unexpected happened in one of those games?

The result was a game that, out of everything I’ve ever made, almost nobody’s ambivalent about FATHOM. I still get these really touching emails maybe once every couple months of someone who played it and they really connected with it and they thought it was really interesting and it kind of changed the way they think about things. And that’s amazing to me, because it’s not even that well done.

But there’s a lot of people who absolutely hate it. It’s just the worst thing they could imagine a video game ever being. To the point of suspecting that it’s a cruel joke, which Happy Puzzle Party is, but FATHOM is very much not intended to be.

The first part of FATHOM, where you’re thrown into an ultra-familiar side scroller world, is almost as jarring as the abstract second half.

For me it highlights sort of the challenge of really flipping around people’s expectations when they expect to be in control of the game experience. My theory is that’s what provokes such strong reactions. [I think] there is a type of player who plays to have control over something, and then there’s other people who may play a game and they play it almost submissively. Even if it’s an active, interactive game, you’re playing it to experience, to sort of be led on a tour of expectations by the author. As opposed to building something or killing everyone or destroying everything or collecting everything.

There’s a lot of space in game design to explore these sort of player/author scenarios. I think if you sit down expecting one thing and you get the other, you can really disconnect with the game in a strong way. I think some people sit down with some triple-AAA game that’s had a lot of advertisements talking about how psychologically powerful it is, and you sit down and it’s just a hardcore action game where you murder everybody. You kind of feel like, “this isn’t what I was expecting, I was expecting something that was more psychologically rather than actively modifying the world.”

The thing that was making FATHOM players upset is the game starts you out in this experience that is very familiar, and you are very active in obvious visual ways. And it immediately takes you out of it and pushes you into this other kind of game which is slower and more psychological. 

You’ve said in the past that the first game you bought was Super Mario Bros.

Yeah. I had Atari before that, but my memories are limited.

I think that’s probable the case for a lot of people. I’m wondering, from Mario on, what was the first game that twisted your expectations in the player/author relationship?

Hmm… That’s a good question…

So when I was a kid, video games were evil like rock and roll was for a while there. I had limited video game time, so I could only play a half hour a day or something like that. So a lot of my engagement with video games as a kid—for like, a decade—was mainly this very imaginary kind of relationship where I play for a few minutes, and then just spend all day drawing graph paper mario levels and horrible seven-year-old-fan art all day long.

Eighty per cent of my relationship with video games was all in my head. I don’t think that’s super rare with people that are my age, either. So games that had really evocative music—so stuff that Sunsoft was doing for the NES. They did this game Journey to Silias. It’s essentially all of the worst parts of Contra combined with all of the worst parts of Mega Man. But it has an amazing soundtrack. This really evocative, atmospheric, rocking chiptunes soundtrack. That was a game that I even a few years ago I would imagine doing a spiritual sequel full of all of the ideas that i got from NOT playing the game.

But it’s not exactly the same thing… I’m trying to think of the first time I was playing a game and something made me sit up and notice that they were breaking the rules.

I remember having this really great time with Resident Evil 4 when it came out because the the way you engaged with that kind of game at the time was: when you entered an area, your main job was to kill everything. And once everything was dead, collect the treasure and then advance. That was your basic template.

Once you get a handle on Resident Evil 4 it actually does settle into that pattern, but they very deliberately start the game in this very off-balance scenario where you’re strongly encouraged by the game system to just survive in a very basic sense. You don’t really have much ammo, you don’t really have a lot of options at your disposal, and you’re in a village full of monsters trying to kill you. So as they come towards you, instead of rewarding you for killing everyone, which you basically can’t do, they reward you for shooting a couple of them in the kneecaps so they fall down and block the rest of the mob, then running back into the house, slamming the door, pushing a dresser in from of it, climbing onto the roof, and jumping to the next building, which is a way more interesting way of engaging with a group of attackers.

I remember playing the first part 10 times and getting frustrated. “I’m doing this the right way. You kill the guys and pick up the gold.” The epiphany that maybe I should play it differently, that I should play it if i was imagining myself in this kind of scenario. Doing that and having it work is just the coolest feeling.

Did engaging with games through your imagination help you as a designer?

I’m suspicious that’s it’s the only reason I design games now. It’s obviously impossible to say for sure, but I I spent a long time dreaming up things that would be cool for games. That doesn’t make you a game designer or anything, but gave you a way of looking at games or later looking at books and movies. A question I ask myself a lot, and I’m sure lots of other people do this too, when I see a movie or play a game, something I always wonder is how would I have done this?

FATHOM can be seen as a reaction to action platformers not bothering to explore or question the tropes that come automatically. Like, of course you have five lives. Why wouldn’t you magically regenerate after you die? Nobody was asking that question, and it seemed an interesting thing to follow down the rabbit whole. And likewise, I had really high expectations for Mirror’s Edge when it came out and whatever the aesthetic virtues of the game, the way it played and the systems of the game were a really big disappointment to me. And my options were write a scathing blog post, or make my own game about running and jumping on rooftops that would make me feel how I wanted to feel while I played Mirror’s Edge.

The thing that i’m curious about is whether I would engage games and movies and books that way if I didn’t spend a lot of time having to pretend play the games in my head.

The phrase “pretend play the games in my head” must be so relatable to people who grew up with video games. I remember reading Nintendo Power and playing new games in my head before they came out.


Yeah. Let me get a number in gigabytes, here… under my video game soundtracks in iTunes, I have about 2000 songs. It’s about three-and-a-half days of soundtracks. I can just play the game in my head while I’m working.

Does that help you design? 

I like to have at least playlists for different games. If I have a prototype, I think, “Maybe the emotion or psychological effect that I would like here is the way I feel when I listen to these songs.” So I’ll go and get those songs and put them in an iTunes playlist and name it after the prototype. Then the first few days just let that loop and be in the right frame of mind and be thinking, “Is this thing i’m putting in now, does that help the player feel like that music sounds?”

You make tabletop games as well. Do those help you when designing digital games?

I’m not really certain yet… In some ways they are different. For thinking about systems, I really like how pure they are and how simple they are. If you want to engage in systems and think about designing systems and think about designing art that is about thinking, tabletop stuff seems to be the absolute best because.

The thing thats popped up to me the most is that I can usually tell within five or 10 minutes playing one of my tabletop games with somebody if its fundamentally flawed or not. It’s usually pretty immediately clear that there’s some kind of major problem that needs to be solved. That’s something that might take weeks or months if it ever happens with a video game because the systems are buried under graphics or music. Which is cool, but if you’re trying to hone your grasp of systems and system design, tabletop games seem to be the absolute best most pure way to engage with it. Plus, you can play test with a broader group of people.

People still think of board games as being a really niche thing, but with the exception of certain things we’ve put on the iPad, my mom can’t play the video games I design. She can watch it and appreciate the presentation of it, but she can’t engage with the system and therefore think about or feel the things that were put into the system. But you can sit down and play a board game or a card game even if its complicated.