2018 MIT Mystery Hunt Recap

After very much enjoying the 2017 MIT Mystery Hunt for the first time as a remote solver, I decided to fly over to Massachusetts this year to participate on-site for the 2018 Mystery Hunt, which kicked off on Friday Jan 12.

This year's theme was the Pixar film Inside Out, which is about anthromorphic Emotions inside the mind of a young girl. The overall story was that your team played Mind Workers inside Miss Terry Hunter, who was on-site solving the "Health & Safety hunt". Structure-wise using Inside Out made a great deal of sense, as the personality islands allowed clustering of a large variety of puzzles into several broad themes. The production quality of this year's Hunt was impressive, right from the opening skit, and the actors this year did a great job of keeping in character.

Overall I had a very enjoyable on-site solving experience, and I hope to be able to continue solving on-site in the years to come. Our team started off strong during the Emotions round, but we hit a bottleneck halfway in as we were stuck on the Sci-Fi and Games metametas for a considerable amount of time, preventing us from unlocking the last island. Our team unlocked the islands in the order Sci-Fi, Games, Hacking, Pokémon so that remote solvers would be able to assist more at the beginning, but in hindsight we might have done better had the Pokémon island been unlocked first. Oh well.

I'd like to thank Life & Order (formerly Death & Mayhem) for putting on a terrific hunt, my team and especially the members that took on admin roles for sacrificing puzzle/preparation time to make the hunt more enjoyable for others (you know who you are), and MIT for allowing this crazy tradition to continue.

And to the winners: Well done Setec Astronomy, and I look forward to next year!

Warning: Massive spoilers for the 2018 MIT Mystery Hunt below.


Here's a rough summary of events from my point of view — I'm making use of my team's log to backfill most of the chronology. Some of this summary won't make much sense unless you attended or have seen the solutions to the 2018 Mystery Hunt.

Day 1

  • Yeah but It Didn't Work: 19 minutes into the hunt, I jumped the gun by submitting our team's first (incorrect) guess "Yeah, but It Didn't Work!", based on the word TITLE(S) in the partial progress we had made. Needless to say, it didn't work, and we completed the puzzle properly as our first correct solve at 13:24.
  • Let’s Get Ready to Jumble: Anagramming names proved tricky until I recalled the existence of Nutrimatic, but by that point we were almost done with the anagrams. This puzzle was solved at 14:35, our fifth puzzle solved, and morale was high as we made our way through the Emotions round.
  • For the next two hours, I chipped into various puzzles in minor ways, such as helping fill the grid for Cross Words and helping find hidden animals in Caged. I was also enthusiastically writing up answers on the blackboards to assist with meta solving, and attempted to crack the Anger meta but did not have enough information.
  • Digust: The homophones mechanic was easy to catch onto, but it wasn't until 3 hours in that we had enough answers to work out how to pair them. A team member noted down the correct pairing method, from which I obtained REPEL based on the partial information, and this was cashed in by another team member as our first meta solve at 16:51.
  • Anger: I focused on the Anger meta for the next hour, trying to pin down the correct placement of answers in the grid and determining possible Braille letters that could be formed. We were slightly impeded by the fact that we needed to use RAGE not ANGER, but thanks to the successive solving of What's In a Name? and Roadside America and the unlocking of other puzzles to reveal thermometers, we were able to nab the Anger meta with 7/12 answers, one short minute after other team members solved the Fear meta. This later led to an interesting double interaction, where two Emotions visited our room at once and riffed off each other spectacularly acting-wise.
  • Following the double meta solve, I attempted to backsolve some of the Anger puzzles to no avail, whilst another member attempted to backsolve Jeopardy! from the Fear meta, also with no avail until much later. After some discussion, we unlocked our first island at 19:11, opting for the Sci-fi island as we wanted to avoid choosing on-site puzzles first.
  • Warm and Fuzzy: This puzzle eluded us since the very beginning, and from the wrap-up it seems like we weren't the only team that had trouble. A team member managed to solve the Joy meta with just this puzzle missing, and shortly after I managed to backsolve Warm and Fuzzy. This unlocked our second island, for which we picked Games so that the on-site heavy Hacking island could be third.
  • This Year's Hardest Crossword: At 19:50 we unlocked a cryptic that very much lived up to its title. My next 5 hours was spent solely on solving this cryptic — initially good progress was made, and with the help of OneLook I was able to contribute some of the longer entries. However, towards the end our progress had dwindled to the point that a sizable portion of our team was staring at the grid and solving one entry every 20-30 minutes. Some time in between, the Sadness meta was solved and other on-site members had gone off to do the Emotional Finale. After one of my teammates returned I finally explained what I hoped the extraction would be, only for said teammate to immediately realise what was actually going on and finish the cryptic to end our misery (I hadn't been far off!).
  • Shift: Whilst I was crossword solving, Shift was unlocked and promptly solved in eleven minutes, leading to me missing the Codenames puzzle of the hunt. Looking at it afterwards, it appeared to be a cute little puzzle.
  • Family Dinner: Also whilst I was crossword solving, our team prepared a "macaron" consisting of, if I recall correctly, peanut butter wedged between bits of muffin. HQ told us at a later date that the puzzle writer originally wanted nothing less than a proper macaron, but given the difficulty of baking one and the hour of the day there was no way our team would have been able to prepare a proper macaron easily at the time.

Day 2

  • AllSpark: At 01:40, a team member impressively solved this meta on their first guess using nothing but the flavourtext and the knowledge that the answer was a Star Trek episode. We technically had one answer for this meta at the time, but it was not used when solve-guessing.
  • Special Delivery: At two in the morning, two team members and I took a hike to HQ with a sports drink/soda water mocktail (plus an orange slice!) in hand. I wasn't involved in the solving of the actual puzzle, but thankfully we recalled the instruction phrase when asked about it during the delivery.
  • Self-Referential Mania: I wasn't involved in the solving of this puzzle but I was very amused by the concept of it. I watched as some of our team members reached the midway message, where I unhelpfully bid them good luck before heading off for the night. Based on logs, it seems that they didn't take long to finish the puzzle after I had left.
  • I came back early afternoon to find The Desert solved, and we were making good progress into the Sci-fi and Games islands. We unlocked the Hacking island next as planned, since we didn't want to leave the time-consuming on-site walkaround puzzles for last.
  • A Tribute: 2010-2017: This was a fairly routine and quick solve, and the thematic answer helped make it possible to work out the solution with only a few letters in place.
  • Bark Ode: I joined this puzzle partway, after the human portion of the puzzle had been completed and my team members had reached the first message. After a bit of fiddling with Python I managed to extract the ensuing barcodes, learning along the way that eyeballing a 2000x2000 ASCII barcode to see whether it looked like a QR code isn't exactly a good idea due to line wrapping.
  • Thanks: At around 18:00, our team hit a wall with this puzzle. They had all but one location, noticed the HER NAME IS message, and were struggling with extraction. This marked the first of many puzzles which we ended up stuck on the last step for.
  • The Robber: A few team members were able to solve the printer's devilry clues and piece together the soccer ball fairly quickly. We knew what we had to do to place the Catan tiles on the ball, but unfortunately at this point we didn't have enough answers. More troubling though, was that none of us seemed to have a good method of extraction for the final answer — ideas tossed around included taking common letters between adjacent tiles, indexing into the printer's devilry clues, sports-related knowledge (especially for Australian rules football), joining tiles of the same index by traversing the ball's surface... but none of this seemed to help. This meta ended up haunting us all the way until the end, and became a bottleneck as we were just a few puzzles short of unlocking the last island for far, far too long.
  • Worldwide Contacts: I tried to help with this puzzle, but all I could do was watch in awe as my teammates seemingly pulled the connections for each map out of thin air. One member proposed the phone country code idea fairly quickly, so our solve for this puzzle was smooth overall.
  • Texts from Mom: Thanks to a team member guessing that an early word could be MOM, and my own guess that the first word was HELLO, a quick find and replace helped roughly decode each paragraph with ease. It took a little longer, however, for us to clean up the decryption and work out the additional codes in order to fully solve this puzzle.
  • No Context: I spent quite a while on this challenge, programmatically cross-checking with another team member to make sure that we assigned each answer the correct grammar. Unfortunately though we had trouble with the diagramless part of the puzzle, and whilst I managed to work out the top few rows of the grid correctly, this puzzle ended up being backsolved once the Scout meta was completed.
  • Marked Deck: The completion of the Scout meta by my teammates unlocked the physical challenges in the Build round of the Hacking island. Marked Deck was one of the more interesting physical puzzles, and it didn't take long for my teammates to arrive at INSHUFFLE. Unfortunately though, we never ended up solving this puzzle due to one fatal flaw on our part which reportedly also affected other teams: we used an online solver for the Solitaire Cipher decryption, not realising that the Solitaire Cipher itself permutes the deck as it generates a keystream. Thus the inshuffle we performed and subsequently stared at til the very end had been on the wrong ordering the whole time.
  • Scouting Challenge: As a first time on-site solver, I was eager to go on a walkaround. With two of my teammates, we were able to get most of the required letters on the first pass through, allowing us to fill in the blanks for the solve. The Scouting Challenge passes through a skull on the floor and a display case of robotic animals, which would later hinder us in another puzzle.

Day 3

  • Is There a Draft in Here?: I wasn't part of this solve, but a group of on-site members worked consistently in the corner on this quite involved puzzle for several hours. They looked like they were having a lot of fun and the puzzle itself seemed interesting, so it gets a mention here.
  • Sports Radio: At 02:07 we bought our first solution after being stuck on Sports Radio since the beginning of the Games island. We hoped that this would help with The Robber, but as we still had no idea how the final extraction worked the unlock did little to help.
  • Feeling Cross: My only contribution to this puzzle was a quick chat with a teammate, from which we noted down "Braille crossword" on our Google Sheet and did little else to solve the puzzle. Other team members managed to work out how entries should be placed, a non-trivial feat due to the tricky 1-across clue, and this puzzle was solved at 02:16.
  • Blue Sun 6V4-178-B31 Trace Compression Block: Throughout the day, various team members attempted to guess the solution to this meta by inputting plausible Star Trek episodes. This eventually led to a call from HQ telling us to ease up on the guessing, and that they would keep marking our submissions as incorrect until it seemed like we had forward solved the meta. This meta ended up being bought with Buzzy Bucks so that we could solve the Starship Enterprise metameta.
  • Little Passages: Since it was the middle of the night and I was feeling bored, I decided to explore MIT in the hope of contributing to Little Passages. A remote team member had mapped out the whole game, but unfortunately due to my unfamiliarity with MIT I didn't recognise most of the features described in the text adventure. I did, however, recognise a "skull on the floor" and a "robotic dragonfly", which we passed successively during the Scouting Challenge, so I spent a considerable amount of time exploring the corridor that housed both these landmarks along with their surroundings. This later proved problematic as we didn't know that there was another skull on the floor at MIT, and in the end we bought the solution to this puzzle in an effort to solve The Robber.
  • Model Kit: Out of all the physical puzzles in the first round of Build, Model Kit was left untouched for a few good hours since it was clear that chemistry knowledge was required. The first thing I did upon receiving the puzzle was to take photos and document everything on our Google Sheet, which gave our remote solvers something to work on. With the aim of forming caffeine in mind, I worked with a more chemistry-knowledgeable on-site member to place the atoms and reactions. We did manage to label the atoms correctly on first try, but we were hampered by the fact that IUPAC purine numbering only seemed to apply to the rings, and CH__I_TR_S_VVY didn't make much sense to us. We remained stuck on extraction until the very last day where a fresh look at this puzzle allowed me to Wheel-of-Fortune the answer. To celebrate, we ate this puzzle once the hunt was over.
  • Thanks, cont'd: At this point we had been stuck on the extraction for Thanks for about 16 hours, whereupon a team member noticed that every line had 150 alphanumeric characters. Technically we had EXPOSURE instead of EXPOSED for one of the lines as we believed that each blank was to be filled with a word taken verbatim from the on-site location, but we realised that the latter was more grammatically correct for the given clue. This allowed me to index by class number to obtain ALLGIVING, which I passed to the rest of the team and we were finally able to put this puzzle to rest soon after.
  • Whilst I was resting, at 13:42 we time unlocked the fourth and final island, the Pokémon island. My team was dismayed to find that we had unlocked the token scavenger hunt puzzle last, and many of us found ourselves wishing that this island had been unlocked earlier as it seemed like we could have made good progress here.
  • Executive Relationship Commandments: Soon after unlocking the Pokémon island, we also time unlocked the rest of the Hacking island. I rejoined my team as this puzzle was being solved, and noted that command line flags might be relevant. A few team members and I quite enjoyed this puzzle, although I was unable to help once we reached the MIT-specific Athena portion.
  • Starship Enterprise: At this point we had all six metas (having just bought Substation 3), and all but 2 of the individual puzzles (with the missing puzzles being Message in a Bottle and Bloodroots, both with partial progress). Our team was already quite sure that resistors and circuitry were involved, and had inputted the stardates as voltages, but we did not yet know what values to input for the resistors. HQ dropped by and hinted that "there are more clues in the flavourtext than you have noted", which a team member realised meant country flags. Another member was quick to set up the circuit on a simulator, and despite missing two of the resistor values they were able to produce broad potential letter ranges, allowing them to filter through and submit our first metameta solve at 21:29.
  • Recover the Sci-Fi Core Memory: Taking a break, I watched as three of our team members nervously lined up dominoes to recover the sci-fi core memory. Due to the wire crossing problem there needed to be a point where two lines of dominoes cross over, and whilst technically the bridge our team had built didn't do its job, our solution was close enough and we were able to obtain our first orb (and marvel at the hunt's production value once more as a video appeared from within the orb).

Day 4

  • Studies in Two-Factor Authentication: As the hunt headed into early Monday morning, it was just me and another team member in our room. I soloed this cute little puzzle as the other member attempted to work on Vain Snowball again, after being stuck partway earlier.
  • You Know What's Missing: We were able to get some of the clues to show up, but the sponges fell off the end of a few of our strips, breaking the puzzle. We submitted a request to HQ for a fresh copy of the puzzle. Over the phone they asked what we had attempted, and after detailing our progress they decided to just send us the clues on the strips, although we never completed this puzzle in the end. And in case you were wondering, yes we did use urine as we were unsure whether ammonia or some other compound was necessary to activate the strips' dyes.
  • Under Control: I missed the fun of Twitch Plays Mystery Hunt, but thankfully I was able to partake in what was perhaps the most memorable puzzle of this hunt for me. I think I'll just leave my description there, since words cannot explain how amusing the experience was. Unfortunately we never ended up solving this puzzle as we were stuck on extraction. We were distracted by most of the extra text on the scroll, sifted through too many possible orderings and failed to diagonalise word-wise, although we did note that the last letters of each pose, taken in final fight order, gave the plausible but incorrect answer SENSES.
  • Don't Look: A remote solver had transcribed the Braille earlier, but not knowing knitting I didn't know how to proceed from there. Thankfully another remote solver came in asking what they could help out with, and amidst my long summary of the various puzzles we were stuck on they caught on to my mentioning of a knitting puzzle. It turned out that the rest of the puzzle was actually quite straightforward, and the remote solver was able to finish off this puzzle fairly easily.
  • X Marks the Spot: At this point it was 08:00, hunt was about to end and being the only conscious on-site solver in the room I was on phone duty, which turned out to be surprisingly easy thanks to our team's use of Google Voice. A number of remote solvers had been working on this puzzle and had gotten excrutiatingly close, but unfortunately for various reasons they had to leave just before obtaining the solution. Our team's last submission for this puzzle was a mere three letters off (out of eight).
  • It's Not Easy: Our team's remote solvers had completed the crossword grid for this puzzle with little trouble, and ditto for It's Not Normal, but we were stuck on where to proceed from there. At about 09:00, an on-site solver more familiar with MIT came into our room after some well-deserved rest and noted this puzzle's connection to MIT's Green Building. Unfortunately we were just about to leave to view the final walkaround, and as we didn't know how much more of the puzzle we had left we made no attempt to complete the puzzle in time. As it turned out, we didn't have much left to go at all.
  • Life & Order decided to host the final runaround walkaround for all teams at three different times, which I thought was nice of them as it gave everybody a chance to enjoy the hunt's finale regardless of progress. The remaining on-site members of my team joined us at the 09:15 session, and I sat at the back watching what was effectively a live game of Keep Talking and Nobody Explodes. It was clear that the final walkaround was designed to involve as many team members as possible, as audience members on one side of the room had to read out instructions for members at the front in order to determine how to configure various switches and buttons to guide the story's protagonist, Miss Terry Hunter, to the coin of the Health & Safety hunt. The event progressively got more ludicrous as some of the device configurations required knowledge of anatomy, musical chords and even Street Fighter moves.
  • The Robber, cont'd: After being stuck on The Robber for far too long, and with three minutes left before 10:00, our team called up HQ asking for hints for this meta (we actually had a lot of Buzzy Bucks saved up at this point, but nobody could think of good questions to ask or useful puzzles to buy solutions for). The one useful clue HQ provided was the need to look at antipodes, and whilst we weren't able to solve the meta before 10:00, we were able to submit the solution at 10:22 thanks to HQ still accepting solutions but merely not calling back to confirm correctness. As a team we felt that the use of "Australian rules" to clue at antipodes was rather weak, and in particular I was lead astray by the paired indices, which did not factor into the final solve. However, as a team we were glad to technically have gotten two full metameta solves for the hunt.
  • After the hunt wrap-up was over, Life & Order made the physical puzzles available for purchase. Having thoroughly enjoyed the hunt, I picked up the physical puzzle set, predominantly for the neatly cut Marked Cards and for Fowlty Towers, which we never solved but I'd like to attempt on my own some time.