Great article. I want to call out one point:

    > The more life-like a space appears to be, the more we begin to expect it
    > to allow for the affordances we have in real life. But games only enable
    > players the actions developers explicitly allow for and implement.
    > Unfortunately, that aspect of game development doesn’t scale with
    > computing power in the way art has.

This is one of the reasons I got out of the game industry. Schatz (and many game devs I've worked with) seem to take this as an immutable fact when, really, it is a choice developers have made. They decided to make games with realistic artwork and limited interaction. The realistic art is nice, but when everything is a polygonal shell with baked in lighting, then the set of things you can do with it becomes very small.

The computer doesn't force you to build games that way. The best counter-example I know is Minecraft. In Minecraft, the art is decidedly unrealistic, bordering on abstract. But, in return for that, the world is incredibly interactive. The entire world can be generated procedurally and freely remolded by the player.

When you play many games today, you will never ever do anything a level designer and game programmer didn't explicitly implement support for. You may fail to advance to the next scene if you don't push the right buttons at the right time, but the overall play experience is not too far removed from a film (and so many game designers out there have deep film envy). That's fine, but it's not super appealing to me.

In contrast, when I play Minecraft, I play in a world no one on Earth has ever seen, and I build things no one has ever built.

But many game developers, and especially producers, don't like games like that. One problem is that, like the film industry, big game companies are making bigger and bigger budget games. The main problem becomes controlling risk. A realistic on-rails game where the player works through a single scripted narrative is a controllable experience. You can reliably estimate how much players will like it.

With emergent and procedural gameplay, it's really hard to predict if the end result will be fun or not. For every Minecraft, there's a million sandbox games that don't have everything balanced just so to get it to hang together in that magical way (see: Spore, No Man's Sky). That's a risky bet for a producer deciding what kind of game to make next.