Procedural Generation in Game Design

Screen Shot 2017-06-08 at 7.10.14 AM.pngProcedural Generation in Game Design is out! Kate Compton of Tracery fame writes about generative art toys; Mike Cook (PROCJAM, Games by Angelina) writes about ethical generation and also about the procedural generation of game rules; Harry Tuffs (A House of Many Doors) writes about procedural poetry generation. Jason Grinblat and Brian Bucklew (Caves of Qud) each have a chapter. Gillian Smith (Threadsteading, plus lots of cool research) writes about evaluating and understanding what’s been generated. Ben Kybertas (Kitfox Games) covers procedural story and plot generation.

The whole volume is edited by Tanya X Short (Moon Hunters) and Tarn Adams (Dwarf Fortress). And I am leaving out a lot of cool people and chapters here, but you can check out the full table of contents on the website.

My contribution — drawing on experiences from Versu, my character-based parser IF, and assorted other projects — is a chapter on characters: how generating dialogue and performances can help realize an authored character; approaches to generating characters; considerations about what is even interesting to auto-generate.


And in a related update to a previous post: I’m happy to say that the PROCJAM Kickstarter has succeeded and has now put out a call for artists to make art packs for procedural work, together with a call for tutorial authors. If their funding goes even higher, they’ll be able to commission two art packs; translate the tutorials they build into additional languages; and hit some other cool stretch goals.

January Link Assortment

Sam Kabo Ashwell has some wonderful posts on the experience of This War of Mine (1, 2) and The Long Dark: the atmosphere, the emergent narrative, the experience evoked by their systems. This bit from his review of The Long Dark particularly struck me:

Having been lost in the Northwoods before, I can say with all confidence: the biggest, scariest threat you face is that you will walk for days and days and never, ever see a single trace of human influence. Never encounter anything shaped by humanity into something that facilitates transport, shelter or food. As moderns, we are hugely, continuously dependent upon the work of other hands. That fear, the fear of a totally non-anthropic environment, is something that is almost impossible to make interesting in the purely human-made context of a game.


David Welbourn is one of the quiet heroes of the IF community: for years he’s been helping to maintain ifwiki, assembling the eligibility lists for the XYZZY awards, and creating loads of high quality walkthroughs and maps. He has an enormous amount of patience and an encyclopedic knowledge about many corners of IF history. If you have any regular contact with the IF community, you’ve almost certainly made use of some of his work, even if you’re not aware of it. I’m delighted that he now has a Patreon, which will help him with scanning and internet costs and make it easier for him to continue.


Rowan Kaiser, Austin Walker, and Alex at While !Finished wrote a series of articles on choices in Dragon Age: Inquisition, and in particular about which of those choices are emotionally resonant:

Alex writes:

One of the most difficult choices in the game, for me, happened in the Solas romance storyline, which is only available to female elf Inquisitors and therefore a minority of players. Near the end, Solas reveals the true meaning behind the Dalish elf’s face tattoos: they were originally slave markings, from when elves enslaved other elves. The Inquisitor can let Solas remove hers, or she can keep them. Does the knowledge of their origin taint them? Or are they a part of her and important to her, no matter what their original meaning? What does she believe?


The discussion of IF fanfiction brought up that there actually is some on archiveofourown: I found an alternate ending for Galatea and a prequel to Alabaster (which digs even deeper into some of the mythology around Eden and Adam’s wives before Eve). There’s also a wonderful story set in the 80 Days universe that explores some of the background of automata with souls, and the lion-like automaton of Burma, one of my favorite figures in the game. And here is an Inform game about a Fallen London character.

Continue reading

Edge article on video game AI

This has been out in print form for a while, I believe, but the content is now available online: I was interviewed for an article on video game AI for Edge. The article draws extensively from the talk I did at the AI Summit at GDC 2012, with subsequent followup. I talk about social behavior AI and its potential in gaming, and there are also comments from Ben Sunshine-Hill (whose GDC talks always leave me in gobsmacked awe) and Mike Treanor (talking about Prom Week and Facade).

Escape from Colditz as Procedurally Paced Narrative

Escape from Colditz is a board game about the German castle that during World War II became a prisoner of war camp for prisoners who had already escaped at least once from some other camp. The idea of putting all the most clever and resourceful prisoners together in an old building riddled with hiding places and odd physical quirks was, arguably, not the brightest; those imprisoned found an astounding number of escape possibilities, and the whole story became the basis of a surprisingly strong British TV show. The board game doesn’t touch on the more complex issues here, but what it does accomplish is in its own way remarkable: a skillful pacing of events that creates a sense of growing narrative urgency.

Continue reading

Drama Management with Anchorhead-based Test Case

Manu Sharma, Santi Ontañón, Manish Mehta, and Ashwin Ram are publishing some research into drama management using a simplified version of Anchorhead (with a choose-your-own-adventure interface rather than a text parser).

The full paper is worth a read (though fairly technical), but the gist is that they proceeded by taking a series of cases from players, determining which players liked which subplots. They then designed a drama manager that would compare the current player’s behavior against its cases, determine which subplots this player was likely to be most interested in, and hint the player in the direction of those subplots. The result appears to be a better experience especially for non-gamers, though some players (especially the more experienced ones) disliked being over-hinted in the direction of things they would enjoy.

When no confident predictions can be made from player predictions, the drama management model falls back on author-defined rules about what to present when.

I have a bunch of minor quibbles with particulars of the study, but found the conclusions intriguing.

Continue reading

Sims 3 in review

Endearing semi-bugs(?) aside, my recent Sims playing has been comparatively uneventful. I played a family through several generations; then, on advice to interact more with the non-active households to watch how those evolve, I set up a larger neighborhood, moved in every family I’d ever created, and spent a lot of time having my active Sims visit the ones that were running by themselves.

The result seemed to be that all the Sim families I wasn’t actively playing immediately got a new baby — presumably by adoption, since none of them acquired new romantic partners or spent any time pregnant as far as I can tell — but that otherwise they seemed rather static. The evil, mean-spirited Lars and humorless, snobbish Lisa still had fights, but still continued to live together until their family left the neighborhood entirely for no reason I could see. They didn’t seem to make changes to their houses, either. Maybe they can’t buy things when I’m not there to buy them new stuff. Or maybe it was just that they weren’t very proactive about job-finding on their own, so didn’t have the cash to make changes.*

Anyway. I’ve spent far more time playing with this than I originally intended to do, and I am still finding lots of neat and unexpected interactions; the amount of detail and care that has gone into the whole game is phenomenal. Many short-term interactions in it turn out to make good anecdotes, too — sometimes because they run into odd corners of the simulation, but not always.

I continue to think that it doesn’t tend to produce good long-term story, because there’s no arc structure. In this respect, it’s being true to what it is trying to simulate: my life also yields the odd anecdote but overall lacks narrative structure.

Continue reading