Puzzle hunting
(Note: this post contains spoilers for some puzzles I enjoyed.)
I happened to be in Cambridge in time to join the annual MIT Mystery Hunt this past weekend. (Well, at least the last 12 or so out of 36 hours.) The Mystery Hunt is sometimes said to be the hardest puzzle hunt or treasure hunt in the world, and parts of it were indeed staggeringly difficult. Most of the teams are distributed around the world, with a large contingent in person at MIT and others participating remotely over the Internet, usually co-ordinating their efforts via wikis, mailing lists, chats, etc. Sometimes teams will try to co-ordinate their sleep schedules so that they have team members somewhere in the world working on puzzles 24 hours a day until the hunt is over. A modern Mystery Hunt team may have 50 members, who may well be literal nuclear physicists, rocket scientists, etc. They are in general astonishingly smart and well-educated people with an amazingly eclectic ability to solve all kinds of things.
I joined a team called Codex Dresden (cf. Maya codices). We came in third overall and were the last team to (be allowed to) finish the entire hunt.
A typical puzzle might have no instructions at all; instead, it might simply provide a number of pictures, words, numbers, or dots, which the solver has to interpret somehow (perhaps working through multiple layers). A typical puzzle I worked on was simply a list of nine large numbers; represented in suitable bases and then written in appropriately-shaped rectangular grids, these numbers were actually pictures of glyphs that represented particular numbers in nine different writing systems (including D'ni from the game Myst), and those numbers represented letters of the alphabet that spelled a word. Another puzzle involved interpreting the diacritical marks on the names of pictured items from the Ikea catalogue as Morse code. Another had three sound files whose spectrograms turned out to form a multicolored map of South America. (I was busy trying to interpret the sound files themselves directly as the channels of a bitmap image, but that turned out to be a wrong turn.) Another puzzle required figuring out the symbols used to depict various monsters in the game Angband, given descriptions of the monsters' appearances rather than their names. Each monster is represented by a particular letter, and the letters -- taken in a row -- spell out a clue. (If only they had been NetHack monsters, I would have known them instantly -- but I'm almost completely ignorant of Angband.)
Perhaps inspired by the printer tracking codes project, I spent hours transcribing anomalies in the pixel-level alignment of characters in the titles of puzzles from one round. The letters were slightly and apparently deliberately off-axis, and Mika and I managed to produce a complete database of exactly how many pixels off-axis each character in each title was. This information was never used in the solution of any puzzle and seems to have been a red herring.
I was also impressed by a puzzle involving the double dactyl poetic form (Riana take note!); unscrambling and reasoning about several double dactyls led to an instruction to compose a new double dactyl and submit it to the organizers. (There was even a reference to the idea that epics ought to be written in dactylic hexameters -- and to linguistics, Iphigenia, and Wrigley's Doublemint gum advertisements.) Several teams did solve this puzzle and dutifully sent in their own espionage-themed double dactyl handiwork.
In the final "runaround", we had tense moments in a room with a literally (mocked-up) ticking time-bomb, complete with a scary red LED display, lots of colored wires, a suitcase, and what looked a lot like plastic explosive. It looked just like something out of a movie, and my teammates had ten minutes to defuse it (based on a schematic and a simulation they'd prepared) without setting it off. Fortunately, unlike the movies, they got to try again when it didn't quite work out.
Later on I got my wish for a NetHack puzzle. As the only expert NetHack player present in person from Codex Dresden, I was placed in front of a laptop deep within Moriarty's lair on which a mysterious NetHack game was running. Unlike regular NetHack, the game restarted from a given point every two dozen moves or so -- apparently no matter what you did. You would be exploring and then you would just find yourself right back where you started from.
The intended solution was to type in an English sentence, which, interpreted as a series of NetHack commands, would perform actions that would allow you to get into a particular room. As it happened my teammates had not remembered the fact that the sentence in question was case-sensitive and had transcribed it case-insensitively, which, of course, caused it not to work right. (I kept wondering why the sentence wanted me to zap a wand when I didn't have any wands. Of course, it was really trying to get me to cast a spell.)
As I tried my own ideas and various versions of the mysterious sentence, I started to wonder whether there were any way out. No matter what I did, I found myself back in the same position, as though nothing had happened. Missing the password and stuck in an infinite loop from which there appeared to be no escape, I, like the main character in Groundhog Day, decided to try an experiment: seeing a lava pool, I intentionally dove into it to kill myself. Remarkably, it turned out that the programmer of the NetHack puzzle had not anticipated this eventuality (perhaps because players don't normally intentionally jump into the lava?) and it bypassed the restart code path, causing the NetHack program to exit and the computer to declare that I had solved the puzzle. So I'm happy to say that I outsmarted Professor Moriarty -- at NetHack.
The very, very last puzzle was a sad story for us on account of the freezing rain and the late hour, but we got through it with the help of a lot of use of the text of I Kings, something I hadn't really looked at quite that much since my bar mitzvah (parshat Chayye Sarah).
You can read about the hunt and try the puzzles for yourself at the 2006 Mystery Hunt home page.
In the puzzle hunt I saw Aaron, who joined Codex Dresden for quite some time, and got to meet SJ for the first time.
