The Mystery Hunt is amusing

[dr_whom]

The big story from this year's Mystery Hunt, I guess, is the restrictions and confusion around overnight solving. About a week before the Hunt, the organizers, team Left Out, announced that MIT was refusing to waive their official policy against on-campus events taking place between 1am and 7am, and so teams would have to vacate their reserved solving rooms overnight and would not even be allowed to do on-campus puzzle data gathering between those hours. After what I can only assume was an intense series of negotiations between Left Out and Puzzle Club on the one hand and MIT administration on the other hand, they eventually decided that teams with MIT student members could remain in their rooms, and you could do Hunt activities in MIT public spaces overnight as long as you wore an official Mystery Hunt name tag. (My team didn't have to vacate our room, as it turned out, because we didn't reserve it through the Hunt organizers, but there was some confusion on that point at various times.) Left Out did a heroic job making sure the Hunt could go on as usual under the circumstances, and I really want to salute the last-minute work they must have had to do.

I don't want the overnight issues to overshadow what I think really should have been the big news from this Hunt, which is this: Left Out finally instituted automated answer checking. When I started doing the Hunt, the way you confirmed a correct answer was that you phoned the organizers' official Hunt phone line and they would tell you if your answer was correct or not. At peak times you might spend a long time getting busy signals and waiting to get your answer confirmed. In 2006, when my team ran the Hunt, we switched to a system where you click a button on the website to put your team on a queue, and then someone on the organizing team would call you back to check your answer, and with small changes that was basically the system for 15 years. This year Left Out finally bit the bullet and made the change everyone knew had to happen eventually: you just type your answer and click "submit", and the system instantaneously tells you if you're correct or not. Past writing teams had resisted instituting that in part because it makes it easy to just spam guesses until you get the right answer; but Left Out had a system that imposed a time delay on answer checking if you submitted too many wrong answers for a puzzle too quickly, and that usually worked as sufficient disincentive for too much guessing. Having a fanfare play for everyone on the team who had the Hunt page open when anyone on the team submitted a correct answer was an inspired decision; it really made the answer confirmations seem like a celebratory communal team event rather than just one person's accomplishment.

Another Hunt-wide innovation I liked a great deal was the Penny Passes—the reward for attending events was a pass that would let you unlock a new puzzle in a round of your choice. In a Hunt with pretty strict unlocking—basically you unlock a new puzzle when you solve a puzzle—these were a good tool to keep the Hunt moving and keep people from getting stuck, without just handing out free answers as some teams have done in the past (which seems much more high-stakes!).

I liked the overall round structure of the Hunt—it wasn't too complicated, but I appreciated the difference between the more straightforward inner rounds and the more experimental outer rounds.The concept of a round where all the answers are emoji was inspired (and "Creative Pictures Studios" was a great title and theme for the round!), and it allowed different types of puzzles with different methods of answer extraction than your ordinary Hunt puzzle; this also interacted well with the use of automated answer checking, inasmuch as people didn't have to phone us and say "your answer of 👠 is correct." The Safari Adventure round, with multiple answers per puzzle contributing to multiple submetas in a complex way, was also a really interesting innovative round structure; the Cascade Bay round, with nested sets of puzzles, resembles the Katamari round from the 2011 Hunt but uses the structure differently, so as to have each meta use the same answers in different ways. I don't know if there was an unconventional or experimental structure to the Cactus Canyon round; the description of the meta on the solution page makes it seem like there wasn't, which if so feels a bit unsatisfying to me.

My team didn't quite manage to solve the Creative Pictures meta, and never even made enough progress on the other three outer rounds to get a sense of what their structures were or how they worked, so we really only saw about two-thirds of the Hunt. We haven't been trying to win since 2010, but over the past ten years we've moved from still being a top-5 team to a top-10 team to a top-20 team, as we get farther from our peak and other teams put more effort into improving their solving. It's an interesting adjustment to make: I'm still used to, even if we don't get to endgame, getting a solid idea of the structure of the Hunt and how its rounds worked; and this year we barely even got into the last three rounds and didn't get a chance to experience whatever interesting meta structures they contained. I was a little disappointed about missing out on some of the key parts of the Hunt that I'm used to seeing.

The Hunt opened with the literal, real-life on-stage wedding of two members of the writing team, and honestly I felt a little weird about it? The invitation they sent out a week before the Hunt was just an alteration of the invitation we had sent around for Mario and Peach's wedding in the 2011 Mario Hunt; it both seems a little inappropriate of them to literally just copy our team's design, and certainly didn't make it clear that this wedding (unlike the one in 2011) was going to be real. And so there I am unexpectedly a guest at an intimate moment in the lives of two people I don't know; I'm glad they got to have their wedding in a venue that was meaningful for them, but it felt inappropriate for them to be imposing their important personal moment on everyone who was just there to solve puzzles. The Hunt theme fakeout in the middle of the ceremony was pretty funny... but the segue from the wedding to the actual Hunt theme seemed very awkward and tacked-on. Maybe I would have felt better about having a wedding at Hunt kickoff if the wedding actually had something to do with the Hunt, since introducing the actual Hunt theme is what kickoff is for?

That said, the theme of the Hunt was delightful and I definitely would not have preferred a Hunt that was actually wedding-themed. Left Out really nailed the aesthetic of the cheesy, past-its-prime, low-budget amusement park; I loved the mismatched collection of mascots. And the overall amusement park theme allowed for a lot of fun sub-themes. I think the writers did a really good job of keeping the puzzles on-theme in each round, either through the actual puzzle content or just the flavortext. The pressed-penny souvenirs were maybe my favorite component of the Hunt: they were a vehicle for the endgame puzzle, of course, but they also put a nice capstone on each round, tied in to the Mystery Hunt as an institution by being "coins", and perfectly fit the thematic aesthetic as cheap souvenirs from a cheap amusement park; and moreover, they were just fun to make. They had a station set up in a room on campus where you would feed a penny into the press and turn the crank yourself to flatten it and print the puzzle/souvenir symbols on it. The Left Out members running the workshop when I went to get the Spaceopolis penny were kind of perfectionists, making me run four or five pennies through the press till I got one they were satisfied with (I spent five cents doing that!), but after I got back to the team's solving space I saw why: some of the other pennies we had gotten from earlier rounds were not as well flattened, and you could still see the Lincoln Memorial on them, obscuring the puzzle-relevant symbols. During cleanup at the end of Hunt, I nabbed the penny I pressed to add to my collection of Hunt souvenirs.

Here are my spoilery thoughts on some of the specific puzzles that I worked on and thought were noteworthy:

• The Trebuchet: I really like the use of puns to make this puzzle thematic: the categories of words you search for (e.g., brandy drinks, quarterbacks) are clued as projectiles the trebuchets are flinging at the castle (e.g., GRAPESHOT, THROWING STARS). The footnote on the bottom of the puzzle, "The final answer is normally abbreviated, but please submit the unabbreviated form," is an interestingly blunt way of resolving a common problem with Mystery Hunt puzzles—if there are two equally valid ways to spell the same answer, how do you make sure solvers pick the intended one? That said, given that MR POTTER and MISTER POTTER would fit the meta equally well, I wonder why they picked MISTER POTTER as the correct answer; the footnote makes it sound like they considered MR POTTER a more likely choice. (I wonder, even, whether they could have just seamlessly accepted either spelling—though I do value the principle of having a single canonical correct answer string for each puzzle.) I guess the presence of the footnote also acts as a clue that the answer is MISTER POTTER and not, say, HENRY POTTER, which seems equally reasonable given the clue.

 

• The Castle metapuzzle: A recurring principle in Mystery Hunt puzzle design is this: anything that there are exactly (or even approximately) 26 of can be used to index the alphabet. 26 cubelets on a Rubik's Cube? Each one represents a letter and you have to figure out which is which. 52 = 26 × 2 cards in a deck? Just have each letter represented by two cards. 27 Constitutional amendements? Just don't use the last one in your puzzle. So I raise a glass to whoever noticed there are exactly 26 counties in the Texas Panhandle and decided to just use that as the alphabet for the puzzle. The answer to this puzzle was OILMEN, and we got thrown off for a long time by the fact that the four puzzles we solved in this round happened to be the ones cluing the letters L, M, N, and O—so we had four counties all in a line from east to west, and they happened to be the counties that Interstate 40 passes through. So for the longest time we thought that that specific row was what mattered, rather than the Panhandle as a whole. I appreciate the flavortext cluing in for this meta for both essential components of the puzzle, the Six Flags over Texas and the Panhandle; but the words "stately, capitalist" in the flavortext really made it look like we were supposed to be thinking about state capitals.

 

• In Space, No One Can Hear You Sing: A Karaoke mini-event! I admire the feat of scheduling that must have been necessary to make this happen: it wasn't a regular scheduled event that all teams send people to at the same time; teams could unlock this puzzle at any time and try to make an appointment to participate, but five teams had to be there at once for the event to go on—I hope not too many teams were kept waiting for a fifth team to unlock the puzzle to be able to do this! The way it looked to me as a participant was just, each of us was called up one at a time to choose a song to sing from a short list; we were told our teammates would be watching it on a livestream. I went eighth, and so I sat there through some really excellent renditions of songs like "Don't Stop Me Now" and "Natural Woman" and gradually got nervous about whether my list would even have a song I knew—only one of the people who went before me had anything I would have been able to perform myself. My list had only three songs, and I picked the one I at least figured I knew about half of: "Oops, I Did It Again". I felt really bad for my teammate when I learned she'd had only one song on her list, and it was one she didn't know. We later found out that the puzzle was about the reality show The Masked Singer: the songs we sang had been used on the show, and on the livestream masks were added to our faces matching the ones from the show as well. My teammate had the bad luck of cluing a contestant who had been eliminated on the first week of the show, and therefore had only ever performed one song. The answer extraction turned out to be a little tricky—we got pairs of letters and originally read them in the wrong order, which spelled BACK RONOIANS WILD BAND, which looks pretty good but obviously is wrong in the middle, and we spent a lot of time trying to figure out which pairs of letters we'd extracted incorrectly in the middle. Then we realized maybe we'd been reading it backward, so we flipped the order around and got BALDWIN SIANO ROCK BAND—which also looks good at the beginning and end but has a non-word in the middle, and maybe it was supposed to be PIANO somehow? But someone eventually found on Facebook an obscure band called Lower the Veil whose members include people named Baldwin and Siano, and LOWER THE VEIL turned out to be the answer. I'm not crazy about that answer or extraction—if you're cluing something really obscure, you really need to use a clue phrase that's really crystal clear so that solvers can be totally convinced they're on the right track. BALDWINSIANO doesn't look much better than BACKRONOIANS, and the Google search that turned up an obscure metal band from New Jersey was a hair's breadth away from being dismissed as "this search didn't turn up anything significant".

 

• Stress Test: A phonology puzzle! I'm always pleased when my first idea about how to solve a puzzle turns out to be right. A fun puzzle and a very quick solve, and good match of puzzle structure to answer, as well.

 

• The Spaceopolis metapuzzle: My only real metapuzzle aha of this Hunt. We indexed the star numbers into the puzzle answers as intended, and spelled out __GU CRATER and then eventually _EGU CRATER—we never solved the puzzle that would give us the first letter—and some of my teammates kept saying, okay, lets find craters whose names look like that and call them in, that must be the answer; and so we called in ANGU CRATER and DEGU CRATER and so on, and all the time I'm like, the puzzle can't possibly be that simple, I think the puzzle answers—HEAD KNOCKER, SKIPPING TOWN, MINIMAL CLOUD STORAGE, etc.—all look like clues for some other group of things. Degu Crater is on Venus, so I kept trying to come up with ways the puzzle answers might clue other craters on Venus, or other things associated with Venus, or something like that. Then in the morning Noah discovered that Degu Crater isn't the only crater that fit the letters we had—there's also a Hegu Crater, which is on not Venus but the far side of the moon. At that point I was immediately like, Ohhhh, I bet the puzzle answers clue tracks from Pink Floyd's Dark Side of the Moon. And they did! And we called in the answer in like five minutes after that. The fact that Hegu and Degu craters both exist (and are both pretty obscure!) is a little unfortunate, but we did eventually come up with Hegu, and I really liked this meta.

 

• Espresso Stand: I noticed a few puzzles using this sort of cycle structure for ordering: each bit of the puzzle clues a link between two items, and that gives you a cycle through the list of items that you can use for answer extraction. So in the Masked Singer puzzle, we each did a song from one performer with the mask of a different performer, which gives us a cycle through the list of performers. In this puzzle, the label on each cup corresponds to the contents of another cup, which gives us a cycle through the seven cups. This is an elegant way to impose an order on things that wouldn't have an inherent order, but it's a little underspecified: you usually can't tell where in the cycle to start, or which direction to go. I looked at this puzzle after I got back to the hotel Friday night and was just about to go to bed, and my contribution was this: my teammates were applying the wordplay gimmick from one cup to the name on the next cup in the cycle and getting nowhere, and I tried applying the wordplay to the name on the previous cup instead and that worked better. But unless I'm missing something, it's a bit awkward that you just have to try both orders and see which one works. 

 

• Fortune Cookies: So, the way this puzzle is supposed to work is, it's a fortune cookie puzzle, so you add "...in bed!" to the end of each clue, and then you can solve the clues. But that never occurred to us, because we solved the clues just fine without doing that. Then we had a bunch of clue answers we didn't know what to do with; at some point, Ricky, our remote power solver, said, well, these words seem to clue other words that you can add a B to, so LOAN clues LEND which becomes BLEND, WEE clues LITTLE which becomes BELITTLE, RATITES clues EMUS which becomes BEMUSE, etc.; and I was like, that's really a stretch, sometimes you're adding B and sometimes BE and sometimes you're adding an E at the end as well, there's not really a common pattern. And of course it turned out he was totally on the right track, since what we're doing is not adding a B to the beginning but putting the whole word in BED so EMUS becomes BEMUSED and so on. I guess the lesson here is, trust Ricky's instincts. I assume we would have struggled with it less if we had noticed the "in bed" connection on the initial set of clues.

 

• The Scrambler: Like Stress Test, an extremely smooth solve for me. I looked at the first clue, solved that, figured out what was going on, and noticed the answer extraction after solving a few more clues, and the whole thing worked exactly how I expected it to work. So it was pretty easy, but just a beautifully elegant wordplay puzzle. I also noticed this had the exact same title as a puzzle from the 2006 Mystery Hunt; I wonder how often that happens. (It was in the Epcot Center round in 2006, and I know the title was chosen then because it sounded like a theme-park ride; I guess it really did.)

 

• The Harvard Ballooniversity metapuzzle: The concept here was just hilarious, and I really like the subtlety with which it was clued. The basis of the puzzle is "grade inflation"—take the B's, C's, D's, and E's and make them A's—but no words pointing directly toward that concept are used, at least not overtly as "flags" and "pan / handle" are in the Castle metapuzzle. It just talks about Harvard and balloons and allows the solver to make the connection and the whole thing comes together very well.

 

• Hall of Mystery: The first cryptic crossword in the Hunt! I definitely started working on this and I don't remember why I stopped or what else I went and did; I usually like to see cryptics through to the end. Looking at the solution now, I'm super impressed; "this clue has two different answers" is one of my favorite puzzle gimmicks—especially for cryptic clues!—and I wish I'd seen it during the Hunt.

 

• Plinko: The second cryptic-clue puzzle was in the same round as the first one; I feel like that's not quite great form? I prefer to see puzzles of highly specific types like this spread out a little bit. Anyway, this one I did work from start to finish, along with my teammate Mike; this was a thorny one but extremely satisfying to see it all come together.

 

• Chairiot Races: This is definitely a single puzzle which, in a not-too-much-earlier era of the Mystery Hunt, would have been an entire round of puzzles; most of the subpuzzles seem complex enough to be an entire puzzle, and it turns out even extracting the meta answer (i.e., the answer to this single puzzle "Chairiot Races") involves a mini–campus runaround after you've solved the 12 subpuzzles. Anyway, I started on this puzzle because the top-level movie title puns looked cute, and then it turned out that to access the next stage of the puzzle I had to download some Augmented Reality app and look at the puzzle through that, and I did that and exported the resulting puzzle videos to our team collaboration documents, and after the hassle of doing that I was pretty sick of this puzzle and burned out on it and went to do something else. Occasionally I would look over and see my teammates hard at work on one or another of these subpuzzles and I think by the end of the Hunt they had solved four out of the twelve of them. Overall I think this puzzle's cost/benefit ratio is too high: solve an easy surface puzzle, use an augmented-reality app to get twelve more puzzles, solve those twelve puzzles, solve the meta based on those twelve puzzles... all for one puzzle answer in a 24-puzzle round, whose meta hopefully you can solve without this one answer.

 

• Screen Debuts: A cute misdirect of a puzzle that looks like a cryptic crossword but is not; we got this puzzle right about when we were finishing Plinko and we were like, another cryptic, so soon?, but someone figured out what was going on with it pretty soon. This was one of the first puzzles unlocked in the round, and made a good introduction to the answers-are-emoji concept. We got the picture of the dog's face, and we were like, okay, is the answer PUPPY? DOG? PETE THE PUP? PETEY? The revelation that the answer could just be 🐶 made so much more sense. This is a really good implementation of the concept: this is an answer extraction that would be too ambiguous to work if the answer needed to be a word, and demonstrates how the emoji-round concept allows puzzle mechanisms that would be impossible in a conventional round.

 

• Refreshment Stand: I did the easy crossword puzzle and then went to bed, and when I got up in the morning I found people fiddling around with pipettes of Gatorate and coffee and it was the same puzzle, which was pretty hilarious. Anyway, I feel this was a less effective implementation of the emoji answer extraction method than that of Screen Debuts: the resulting image was one we thought might be a wizard with a blue hat and red hair, but we weren't sure, because the resolution of the diagram wasn't that good and especially because specific colors had been so essential to the structure of the puzzle and no font or platform's version of the "mage" emoji that we could find had red hair (including the default font in Chrome, the recommended browser for the Hunt). Eventually it turned out that the answer was the "female mage" emoji, which was somehow supposed to be clued by the long red hair (still not present on the actual emoji!) in a way that seems pretty cartoon-gender questionable in the same way that giving, like, female cartoon bears long eyelashes or whatever is.

 

• Road Trip: Another long puzzle dependent on an augmented-reality app that I got sucked into because it had a fun and easy first step. The actual puzzle components were delightful here: using emoji to clue famous books, like 🇨🇭👨‍👩‍👧‍👦🏠🏝 for The Swiss Family Robinson. But it also involved traipsing all over MIT to get the clues. And like... I am a firm believer in Mystery Hunt puzzles that make you get off your behind to explore the MIT campus or beyond; but the problem with this one is that there was no reason, from a puzzling perspective, to involve these campus locations. You just had to go to a bunch of rooms around campus to get the clues for no reason other than that the augmented-reality app wouldn't give you the clues unless you went there; there wasn't anything interesting or entertaining about going to these locations or finding anything there. I spent most of Sunday morning just walking around doing this puzzle and it felt like it could have been a lot shorter and more interesting without the runaround.

 

• Obscure Movie House: An intense and very challenging cryptic crossword—it's hard when you don't know the length of the answers or where they go in the grid, let alone where the clues begin and end!—that was very rewarding to solve. The top right corner of the grid, where MISSING and FLASHLIGHT are, was the last part we got, and so instead of finding the final clue phrase TAIWAN WEDDING CHURCH we spent a lot of time on the blind-alley PAKISTAN ARACHNOPHOBIA CHURCH, which is a little more entertaining but unfortunately doesn't clue the correct answer.

 

• Old West Revue: A great language (or, well, writing-systems) puzzle. The solution page recommends some online methods of attempting to solve this, but we did just have to find people on our team who could read Arabic, Cyrillic, Devanagari, and so on, and then stare at e.g. the Telugu alphabet and try to puzzle out the rest.

Puzzles I worked on and liked, but don't have much to say about: Moat-er Boats, The Scottish Display; Goldilocks; Gallery of Tomorrow; Sand Witches; the Witches' Hut metapuzzle; The Holy Cup of the Raven-God; Teacups; Dog.

Congratulations to Left Out on an excellent and very entertaining Hunt; I just wish I'd been able to see more of it!

Comments

[landofnowhere]

Could you move the comment on the Creative Pictures round to spoilers? Llamas had not yet figured it out.

[toft]

This was neat to read!

[doctorskuld]

I was saddened, but not surprised, when Left Out announced that MIT was enforcing stricter rules about hunting overnight and use of the MIT rooms. It very much feels like somebody new came in, did a risk assessment on the MIT Mystery Hunt, and went, "OH HOLY SHIT," and then laid down the law. There was a lot of speculation on the Codex mailing list of how this was going to affect the runaround (many of which are initiated in the hours of 1-7am), but I guess, in the end, they managed to get MIT to walk back some of the new rules.

I can't imagine how freaked out Left Out was, because it seemed like there were so many more physical puzzle components and items this year than years previous (I am usually remote, but it felt like more). They must have panicked. ...any sane Hunt runners would have.

I liked the automatic answer checking, which I guess frees up the on-site team to do more with interactions and giving out physical puzzle components. My guess is that this will be the death of the old telephone answer confirmation system, which is sad, because it was one of few things I could do for Codex when I was remote and they were running the 2012 Hunt.

The emoji thing really messed with us the first time we got it to, we kept on trying to answer WISHBONE for Screen Debuts. It'll be interesting to extract puzzles now that pictures are more of an option than they ever were before.

We never solved the Creative Pictures Meta, even though by the end of the Hunt we did have Cactus Canyon and the first two levels of Cascade Bay open. I felt the Creative Pictures meta was a little too unconstrained as a first step, and that was really annoying. We had figured out the concept, but even with two-thirds of Creative Pictures Studios solved, it's just not possible to do that meta.

I felt that one of the weaknesses about this Hunt, everything else that was amazing about it aside, was that for many puzzles I felt there was just one layer too much to solve for, or the complexity came just too late into the game. Safari Adventure was difficult with multiple answers, as was Cascade Bay with multiple metas.

Maybe I'm grumbling, because Codex used to be a Top 5 team and we've since slipped down to a Top 10 team, but I admit I was surprised not to see Metaphysical Plant up top anymore. Were you also one of the ones to foolishly believe the previous years' runners when they set down an "optimum" team size? Anyways, if you or any of your friends would like to join Codex for next year's Hunt, we would welcome you. I think we are itching to win again. We have forgotten the pain and suffering that is running a Hunt, and might be willing to give it another go.

ETA:
I am really annoyed by the solution to Hall of Mystery, because I did have them cut out (on photoshop) and was rotating them around, but the first orientation that most people try, L+AKE, O+F, DE+ED is so underwhelming, and the INITIAL+LY likewise. *headdesks*

Agree also on the too-high cost/benefit ratio of Chairiot Races. I really thought that was too much for a round that is impossible to backsolve with a meta that is too unconstrained at the first step.

[gwillen.livejournal.com]

Regarding "In Space, No One Can Hear You Sing" -- when Rage did it, they just had us and one other team participating, and just had people run through multiple times. And in fact, we sent three people (one as an alternate, due to not knowing how long the event would take, against the time constraints of someone who _really_ wanted to sing), and they just let all three of them participate. So I think they planned up front to be pretty flexible about it.