My understanding is more about systems than games, but an understanding of games will be able to emerge from it.

Notation fully unfolded, a system can be understood as a set of states and connections.

A system is a graph of nodes and edges.

All other theoretical models of games are simplifications of this full system-graph view. Koster's "Any atom that involves risk must always have a failure state link" is a statement of simplification. This notation erases all states that do not contribute to a structure of choice-making and risk-taking, and therefore encodes an implicit definition of what a game "is."

In Koster's proposed notation, movement around a safe screen in a platformer is not an atom. As a result it cannot actually capture what games are. The notation excludes significant detail from consideration.


The map is never the territory, but we must remember to regard it as a map. The "Why?" slide fails to address the simplifying effect and presents this notation as territory.

I believe that I have in mind the abstract concept of the whole territory.

It was the immediate next two slides at which I stopped dead. They read as follows:

QuoteCost of failure

Syntactically, atoms must
always have a failure state
link, even if said failure is
only an opportunity cost.

QuoteWhy?

Any atom that involves risk must have
at least a binary result.

   This is why we do not consider
   moving a checker piece without a
   capture or a setup to be an atom.

   I've notated these using red arrows
   rather than blue.

One of the slides in Koster's presentation began to give me a feeling. It reads as follows:

QuoteClarifying dimensionality

Depth
   Literally, the depth of recursion

Breadth
   Literally, the amount of parallelism

Size
   Literally, the amount of sequentially chained atoms.

(Isn't it nice to finally know what these
   mean?)

It was the last parenthesized rhetorical question. Though I cannot transmit the feeling, it was a kind of dismissive, disgusted, superior feeling I associate with the thought that I have recognized someone has produced a simplifying (but inaccurate and destructive) lens and is presenting it smugly as the truth. I was not having that explicit thought specifically at the time, and I'm not suggesting that Koster is smug - either here or in general - but I was having the feeling that goes with the thought.

The moment of understanding occurred while perusing a Raph Koster presentation linked in How We Design Games Now and Why as follows:

QuoteDesigner Raph Koster highlighted the imprecision of natural language as a tool for designing gameplay, and proposed we develop a graphical notation system for game design

(This sentence and this link were provided in context among many other designers' statements and proposals. Here I have included only the one that contained what sparked the moment of understanding.)

I began to read Understanding Comics, not for the first time, on January 31. I finished on February 4. Then I began to read Katherine Neil's How We Design Games Now and Why.

On February 4, Paul Diaz blote thus:

Quotesorta tangential, but this reminds me of Katherine Neil's article, because of sheet music being this abstract design tool which is something games do not have: https://medium.com/@haikus_by_KN/how-we-design-games-now-and-why-bcbc1deb7559

link to bleet

On January 31, Alex Beachum blote thus:

Quotestumbled across this this thread because i'm rereading Understanding Comics and decided to check if anyone has tried to Scott McCloudify video games...idk if this is that but it's interesting! also i'm curious, would your non-player-centric definition include a game that literally plays itself?

link to bleet

existing is an experiment. to think of the end of experimentation is only to think of death. after the dream of death comes the dream of the mountain... i have learned many things i can't put into words! my purpose is not to convey what i have learned, but to use it all.

to move on is not to forget.