New paper applies predictive processing to “what is fun?” – Raph’s Website Game Download

This paper on “Mastering uncertainty: A predictive processing account of enjoying uncertain success in video game play” is very worth a read if you are interested in the frontiers of figuring out what “fun” is. Luckily for me, it doesn’t say I’ve been on the wrong track for decades.

It does raise interesting questions given its framework — I’d love to see slot machines explained — though there is some stuff on affect that likely ties in. It also teases out some of why I have never felt comfortable with the “flow = fun” equation.


Another interesting intersection with other material would be motivations (a la Bartle/Quantic Foundry) and personal goal-setting. Players DO grind, after all, as they optimize, and tho the paper mentions people don’t get stuck in “popping bubble wrap,” they do for a lot longer than one would expect.

For me the answer to that ties back to the lemma/heuristic model of pruning possibility that is usually discussed in the context of “what is game depth.” I’ve come to see forward strategy and perception of depth as being about indeterminacy and a sense of “victory parity” tilting back and forth as we project. There’s something to tease out in that plus motivations plus this paper that could be useful in thinking about how to construct game metas in particular.

Anyway, I encourage the read. It ties nicely to other work such as OpenAI’s RND.

One of the niftiest parts of my career has been seeing pieces of my work turn up as building blocks for others (such as AI systems trying to mathematically implement Theory of Fun). Always feels good when your stuff is built on. Nothing gets to stay at the pinnacle for very long, but getting to be a piece of foundation is a pretty cool and a lot better than the alternative. 😉

Dan Cook also has thoughts on this paper that are worth a read, and open yet more rabbit holes to explore.

New paper applies predictive processing to “what is fun?”

Leave a Comment