I'm jotting notes about another Game Control Summit 2010 talk: Scott Blomquist talks about Puzzle Theory, conceptual thinking about puzzle design. (Yeah, he talked about puzzle theory in 2009, too.)
- We do Information Reduction Puzzles. We get this mysterious pile of information. We want to reduce it to a secret message.
- Today we're talking about confidence and acceleration
- Confidence How much you trust some piece of the puzzle that you maybe-figured out. "I kinda think that 37 across is THUNDER LIZARD but maybe not because that conflicts with 22 down being POSSE."
- Acceleration You can go faster as facts reinforce each other. That last square on the Sudoku's the fastest to fill in.
- Not that we tend to finish filling in our grids. I don't need the whole crossword, just those ten circled letters—well, really, nine of them and figure out the tenth.
- If players need to extract some fact from one layer of a puzzle
to get to the next (or to finish) make darned sure they can be confident
of that information. Negative example of a crossword in which two
"weird" words cross.
- Maybe have more than one way to get at that fact.
- Maybe let them get to the next layer with not-quite-all the facts from this layer.
- Shout-outs to folks who think like puzzle theorists and write/talk
cogently:
- Ian Tullis for his talk from earlier in the evening
- Eric Prestemon for his Solving Really Hard Puzzles blog (with frequent guest star Jonathan McCue)
- Foggy Brume of P&A Magazine posted some "Puzzle Standards": things
you can do for a puzzle to make players think without frustrating them.
- Foggy Brume's Puzzle Standards Part 1: Don't ask players to ID more than 20 pictures; it's hard. If your puzzle leads to a clue (e.g. "Child or Roberts") instead of a word, make sure it's darned specific; don't be coy. Check your answers (in the sense of how a crossword's across answers "check" the downs they touch); give confidence. Make the puzzle fit the answer; theme.
- Foggy Brume's Puzzle Standards Part 2: Keep flavortext simple; less red herrings, please. Give solvers blanks or a grid to fill in; don't be so mysterious, give players a clearer goal. Don't just keep writing Identify-Sort-Index-Solve puzzles; Foggy's so sick of these mystery hunt staples that he coined the ISIS acronym. Play through the puzzle, even though you know the answer; think about "usability", the mechanical experience of solving the puzzle. Ordering: if it's not important, order the clues anyhow, if only to indicate that ordering isn't important; "Oh, they're in alphabetical order, so there's no information in the order".
- Foggy Brume's Puzzle Standards Part 3 Don't confuse obscurity with difficulty; kids have teh internets now. Too few "a-ha"s is bad, so is too many; it's like Ian's wow factor. Give your solver a starting point; as the Anonymice would put it, give them a crank to turn.
- Q&A!
- Brent Holman asks: Scott, what puzzle design have you done lately?
- Tinkering with some software to aid crossword construction, but it ain't close to ready yet. Keep your eye on Puzzlepro
- Alexandra asked: where should we talk about this?
- How about forum.puzzalot.com?
- Brent: People don't tend to talk about puzzles except for right after a game.
- Scott: getting the critical mass for a conversation in some virtual space. You can talk about this stuff over pizza and beer after the game, and pretty much say what you wanted to say.
- Scott: And there is some conversation, albeit not so much in a forum. Like there's some freak puzzlehunt bloggers.
- Sean Gugler: I don't tend to go looking for that stuff.
- Scott Royer: maybe we could set up an aggregator feed.
- Scott: Or a hashtag. Man searching for puzzle stuff online is hard because there's all those get-rich-quick poorly slapped-together "free sudoku game" sites.
- Corey askes: Do you have a collection of puzzles that have been
theoretically decomposed?
- There have been maybe three or four puzzles. Check out the Puzzle Theory Google Group.
- In general, not good collections of old bay area puzzles; let alone decompositions thereof. Some of the old web sites have fallen off of the internets. Joe DeV (et alia?) put together a great list of MIT Mystery Hunt puzzles. There's a good collection of MS Puzzle Hunt puzzles, but it's behind the MS firewall.
- Sean Gugler points out: maybe some of these case studies could be possible points of discussion in a forum.
