Nobody really talked about Sleep is Death. I mean, it seriously lasted like 10 minutes on Twitter. That's a cynical way of putting it, I guess, but...

SS: No, no. And that's fair. I think a lot of people gave it a chance and did not find in it what some people did. Here's a perfect example, and this I have no explanation for -- well, I have an explanation for it, but I think it's really sad. Did you ever play Captain Forever?

That was just a wonderful game. I feel like it's a game that was made for game designers, though. Because my favorite part of the game is tuning the feel of a system, and that was like a lot of what the game was about.

So many people didn't get that. And I just thought that game was going to take off like a rocket because I had so much fun playing the game; I played so many hours of it. Its accessibility is the problem with it.

That was one of the really sticky innovations about Portal: the Valve method of like just playtesting the fuck out of it and realizing, "Hey, we are asking too much of people in this level. We're asking people to do too many things to solve one level. Let's take those two things that we thought, 'Oh, anybody could do that,' and separate them into two levels."

The thing is, while that sounds like it's kind of condescending, A) nobody knows because when you play Portal you certainly don't know what the developers were thinking. And B) no one's dissatisfied -- no one's sitting and playing Portal thinking, "Man, this game is talking down to me because I had to solve two puzzles in two levels."

SS: Well, there are two really important lessons there. One is depth first, accessibility later -- which is kind of one thing that we're struggling with on Shadow Physics right now. It's like we had these levels, that were a bunch of interesting expressions of shadow mechanics, and then people would play it, and then not be able to figure it out.

And so I went and made the game really accessible, but then Jon [Blow] made a really good point to me, which was that if you're going to tutorialize something, you have to show people why it's awesome right away; it has to be magical as well as teaching.

It can't be "Here's a bunch of stuff, and you just have to trust me that it'll get cool later," and that was the problem with Captain Forever. I was telling people this literally aloud -- like, "You gotta trust me; you gotta play this game for enough time so that you understand what's going on." And not that many people did.

But also Portal and Braid have a fundamental advantage over something like Captain Forever, which has more of a conceptual underpinning. Although it is really fun, it feels really good to drag your pieces around and arrange them.

But [with those games] it's really obvious immediately what the innovative, interesting thing is; it's like you just hold down a button and time rewinds as much as you want, or you shoot a portal at a wall and at another wall. What is cool about it is obvious; the ramifications of that are not necessarily obvious.

But now we're back to really pushing for depth in the Shadow Physics mechanics, and as I said earlier, that turns out to be a really hard thing to do. Because you are a shadow; you don't do a shadow, it's not a verb inherently. It's you are a shadow, then you run and jump and push and pull and slide and break things and...

You do what Mario or Lara Croft does, but as a shadow.

SS: Right, and so it's like a riff on all these different mechanics. And so the interesting question is, can we find depth that we need? And obviously I think we can, and a lot of other people think we can too, but that's part of what I meant about letting the game go where it wants to go.

So we have this experimental methodology in mind where we're going to take the time to explore the mechanics as deeply as possible. But by its very nature, because you are a shadow, it's not about a very specific aspect of fiddling with shadows. It may be a more compartmentalized or diffused experience, where it gets deep in a bunch of different areas, or it gets a little bit deep in a bunch of different areas. The thing that ties it together is that you're a shadow. It's a really hard problem.

With Portal or Braid specifically, the other thing that makes it really obvious with what's cool about them is everything else about them is normal. What I'm trying to say is…

SS: They were grounded in familiarity.

Yeah, exactly.

SS: They only innovated in one dimension.

You're completely familiar. You can contrast Portal against Half-Life 2 very easily to see what's different about it. And with Braid, it's the not-so-subtle nod to Mario all the time.

SS: Well, I think that that is part of the strategy that you use as a designer to get people into the game. You drop them off in a familiar context -- it's a first person shooter, it's a platformer -- and then you can let them feel grounded, and then you ease them into the weird, crazy stuff that you're going to throw at them.