MIT Mystery Hunt 2022 write-up

I should first introduce myself. I'm Kah Kien Ong, a puzzle hunt enthusiast from Singapore. I regularly solve online puzzle hunts with teams Tricky and Herrings, and also organize the annual Singapore Puzzle Hunt. I have been solving the MIT Mystery Hunt ("Hunt") remotely for the past around 20 years. My Hunt team affiliations are with Codex (2004-2017), and Palindrome (2018-current), and we wrote the 2022 Bookspace Hunt that I am writing about here. I am not one of the key members or leads, but could share my own personal memories and experiences from the perspective of the organizing process.

Prior to being the first team to finish the 2021 Hunt, Palindrome had been trying to do that (and earn the right to write Hunt) for a few preceding Hunts. We had finished a close second/third in the last four years, which was more disappointing for me personally. It was reminiscent of when I was with Codex, and the team also came very close to finishing first for a number of years, before finally managing to do so in 2011 and write Hunt. Nonetheless, the 2021 Hunt was a memorable one not just because Palindrome was the first team to finish, but also as it was uniquely a fully remote Hunt due to the ongoing COVID measures. So for once, my Hunt experience didn't feel much different as a remote/international solver compared to everyone else. It was really memorable as for the first time, I was able to participate in all the post-meta interactions, and even the Runaround online. However, the anticipation of running the 2022 Hunt was also mixed with the future uncertainty of a continued on-campus Hunt. So there was also some unspoken pressure on ourselves to organize a good Hunt that continues to engage the MIT populace, and demonstrates its value to the administration.

Initially, I had not planned on doing my own Hunt 2022 write-up, which is why I only put this together in December 2024. Partly, this was a way to recap my own Hunt 2022 memories again (especially when it gets close to Hunt season), and these are memories that I would probably forget over time. Another reason was that I felt it would be helpful to share some of my own thoughts on what goes behind writing a Hunt, because there are sometimes differing perceptions on how Hunt should be, and there are just no official guides out there on this. For those less experienced with Hunt, I had also previously put together my own guides on puzzle hunts, solving puzzle hunts, and writing puzzle hunts. But the majority of my write-up below covers the few puzzles that I had written for Hunt 2022, capturing some insights about the ideation and construction process.

Palindrome last hosted Hunt in 2008, and we still had a number of members with that experience, as well as a few other experienced Hunt writers (while with other teams), and writers of other puzzle hunts. We also roped in extra help from Palindrome's sister team Ange, as well as MIT Puzzle Club. Coupled with the large team size, we had enough manpower and expertise for the various roles, although it really does take an army to run a Hunt with high-production value as is the standard in recent years. So while a smaller team could be the first to finish Hunt (as demonstrated before in the past), I do feel it would be much harder for such a team to run one, even just keeping to the relatively typical/short length of Hunt 2022. From a team perspective as a whole, I thought Palindrome managed to navigate the usual stress and management challenges (e.g. timelines, resource management, logistics) of organizing the Hunt fairly well behind-the-scenes. The Bookspace theme was decided early and it was broad enough that everyone could relate to. We were quite clear on our expected audience of MIT students and small teams from the start, and did our best to cater to that. Execution-wise, the individual team experience might vary, depending on whether solvers got bottlenecked on The Ministry meta, which was an unforeseen roadblock for some, and perhaps could have been improved. Difficulty and length-wise, I felt we were largely on point as well, with the first team finishing around the more traditional Hunt end time as estimated. This allowed other and smaller teams to continue solving their Hunt during the weekend, although our communication about HQ closing could possibly have been clearer, given the absence of a consistent practice in past Hunts.

As for myself, I had decided to step up into an editor role given my experience, to help the team and the many new writers. I edited around 13 puzzles, which spanned a wide variety of topics from crochet, SQL, Bridge game, Codenames, phonetics, logic etc. I felt my breadth of puzzle experience and general knowledge, as well as my usual willingness to write puzzles on any topic, helped equipped me well with sufficient context on the different topics to guide the writers. Around 70% of my writers were first-time puzzle writers, but it certainly didn't feel that high, as they all came with great puzzle ideas and attitude. Only one puzzle went through a few rewrites, and those were proactively done by the writers in response to the feedback, resulting in a more ideal final puzzle than the original idea. I mostly worked with the writers on their puzzle execution, to ensure they were optimal for our solvers, and allowed them creative control of their puzzle ideas where possible. So I was most glad to be able to catch a last-minute call to highlight and acknowledge the names of these first-time writers in Palindrome's Hunt wrap-up video. As to be expected though, not all of the editing role was that smooth-sailing – e.g. I had to switch out the writer for one of the puzzles when the timelines were not met, and dropped another before assigning an answer, as the idea was not so sound and there was no further improvement. These are part and parcel of puzzle and Hunt editing. Firm decisions have to be made at times, and they can be difficult because we are all fellow team members. Sometimes we encounter sub-optimal puzzles in a puzzle hunt, probably because either there was lack of editing, the editor didn't do this role well, or the writer didn't allow the editor to do his/her role well.

Besides being one of the editors, naturally I also testsolved a lot of other puzzles from various writers to provide feedback, just like the majority of Palindrome. We generally had quite a healthy, diverse population of testsolvers that was sufficient, unless a puzzle had a few rewrites and needed multiple retests. This is another possible reason why a smaller Hunt organizing team might be sub-optimal - it could mean a need to rely more on a smaller pool of (usually stronger/experienced) testsolvers, which might skew the puzzle feedback and perception of puzzle difficulty/length. For myself in particular, I enjoy solving metas and testsolved almost all of them in the earlier stages. After that I focused more on testing puzzles that I had likely expertise in and could contribute more. At one point, my testsolving involvement took more of a backseat, due to my editor duties, and also as I was concurrently organizing/writing the annual Singapore Puzzle Hunt which ran in July. I forgot roughly how many puzzles total I testsolved in the end, but my position in the credits roll (which I recall was ordered by number of testsolves done) was somewhere in the middle amongst the testsolvers in the team. Other than these background roles, I had also thought about the possibility of being involved in the roleplay skits, because these had to be done in videos for this Hunt. However, these roles were snapped up pretty quickly, and anyway, I probably wouldn't have being as natural as the other team members who acted excellently in those videos. But the takeaway for anyone in a team organizing Hunt is really that the opportunity to do so is rare, and you should go ahead to try your best to get as involved as you hope to, not just to help the team, but also to feel first-hand all these unique experiences. Oh I did make an unexpected noteworthy appearance in the Whoston Round backdrop image too! We had submitted personal photos to the Art team for use in creating the backdrop for this Round, but I didn't realise that mine would end up in pretty prominent position near the top, until I saw the completed test site.

Obviously I also made it a point to write some puzzles for the Hunt, although in the end, it was not as many as I had originally planned/hoped. To me, the expectation for a team to write Hunt is both an honor and a rare opportunity to contribute back to the community, as well as to collaborate with team members on a massive puzzle writing project. So I had wanted to try to write as many puzzles as possible. Time management was not so much of a constraint for me as compared to having good puzzle ideas. Despite building up my puzzle idea folder to a decent size of over a hundred over the years, I still found it difficult to pick suitable ones from there to use for Hunt. Part of it is because of my habit (and stubborness) to only use a thematic idea to write a puzzle that fits an answer well, instead of how many writers might just take the first opportunity to write a puzzle based on an idea they already have, and use any available non-thematic answer for that. This, coupled with Palindrome having many enthusiastic writers, both experienced and new, meant that potential answers and ideas were dropping off the board much quicker than I had anticipated. Notwithstanding that, I did manage to secure a few thematic answers that fit some of my more well-formed puzzle ideas, and am still quite happy with all those eventual puzzles. I wrote/co-wrote about 7 puzzles in total, listed below in order of unlock during Hunt. It was quite fortunate that my puzzles were almost all in the first four Rounds of the Hunt, with one in the sixth Round (out of a total of 12 Rounds in the Hunt). So all of them had a good number of solves. In the following sections, I will share more insights about each of these puzzles, which includes spoilers. So if you have not solved any of these puzzles before, do click on the linked puzzle and try solving them first, before expanding and reading the respective section (which will also make more sense after having solved that puzzle).
- Too Many Toys (The Investigation Round)
- The Last Olympian (The Ministry Round)
- Crooked Crossings (Noirleans Round)
- Strange Garnets (Noirleans Round) - main writer, collaborated with other writers
- Dancing Triangles (Noirleans Round) - contributed one of the sub-puzzles
- Large-Scale Anthropomorphism (Lake Eerie Round)
- This or That (New You City Round)

This puzzle was surprisingly chosen to be one of the three starting unlocked puzzles in The Investigation round, and therefore the entire Hunt! Personally, I thought it was a little longer and tad more challenging than what some teams might expect in the initial set. There were four successive clued parts/steps to this puzzle, and each step required a minor insight. However, I also felt that it would be great for new solvers/teams in particular to have the opportunity (or be forced to) attempt this puzzle, as it was to me a good representation of a standard intuitively clued Hunt puzzle that did not require any specific knowledge or puzzle experience. i.e. It might take longer and be harder for them to solve the puzzle, but it should not be a major stumbling block (the only tricky insight was a clued aha in the first step), and they should get some satisfaction from figuring out each step and eventually the whole puzzle. Still, it was pretty stressful when Hunt went live, watching real-time the other two starting puzzles get solved by many teams, before the first solve for this puzzle finally came in. Construction-wise, the idea for this puzzle came from my daughter Athena's toys, which seemed to provide good material for an interesting puzzle with their use of letters and colors. I had hoped to write a Hunt puzzle for her (she is one of the five girls pictured), so this was a fitting theme. It was clear from the start that the puzzle should be presented in a graphical way to align with the toys theme. My idea was to mix up and scatter the toys to resemble an actual play session, since it would be easy to sort them apart anyway, and a bit of an insight when solvers notice how they match each section. The writing itself was quite smooth, and I was particularly pleased with the final wordplay references in the main solving step in step 2, though also very glad I found enough good examples of each to fit the required constraints. A shout-out too to our Assistant Art Director Joe Cabrera who helped out with the last-minute post-production color changes for the letter blocks, to make them meet the accessibility requirements.

I was initially hoping to write one of the fantasy-themed puzzles for Mark Halpin's meta, but my choice of Alice in Wonderland had been selected. So I ended up writing another fantasy-themed puzzle instead, as he said we could have a few other puzzles with a fantasy-theme as decoys. The answer I chose appeared a little constraining, which may have explained why it had remained on the board for a while. From a writer's perspective, it suggested to me the two potential related themes, which provided sufficiently interesting puzzle idea and materials to put together something. I clearly enjoyed Greek mythology, and the first part of the puzzle made use of a not-too-obvious commonality to reference the Olympians, which hopefully was a satisfying aha. The original extraction idea did not work that well, so I had to rewrite most of the word clues, which took a while given the extraction constraints and need to find good answers that are not too contrived. The second part of the puzzle required a fair bit of time researching on the Percy Jackson books, as I have not read them before. In the Discord channel post-hunt, I read about how a number of Percy Jackson fans in one of the teams jumped onto this puzzle after seeing the reference, and really enjoyed it. So I was quite glad to be able to provide some fan service, even though I am not one myself. But I have always liked challenging myself to write hunt puzzles even on topics I needed to do research on, as long as that could reasonably equip me sufficiently to write a puzzle that does justice to the theme. Just like solving certain Hunt puzzles which reference something new to me, writing hunt puzzles can also be a fun knowledge learning experience.

This is a puzzle idea that I had for maybe 6-7 years in my folder, and had been saving for the right opportunity and answer to use. So this was one of the first puzzles I wrote for this Hunt, when I saw a thematic answer for it. The puzzle idea first came about as I liked writing double logic puzzles, which basically reuses the same grid and/or clues for two different puzzles and solutions. So I observed that the completed Hashiwokakero grid resembles a Twisty Wordsearch, wtih double bridges being possibly reused paths for different words. Combining both logic and word puzzle types into one was also something interesting, and I saw similar post-Hunt feedback from solvers about this puzzle. But even though the idea was already pretty fleshed out in my mind, writing this puzzle was challenging and took some time. The letters in the unique positions need to form an extraction phrase in order of number of times they are used, and the total pool of all these letters also needed to form a set of thematic answers. So these tight constraints are akin to those of a Double Acrostic puzzle, which is how I approached the problem (with new-found respect to the construction behind such puzzles, that I had never tried writing before). Initially, I wanted to include positions with 8 crossings as well for completeness, but this proved to be too constraining, and it also would make the words too similar. I also tried reducing to a maximum of 6 crossings, but after some time, I realized that there would be an odd number of letters with odd crossings, which meant that it would be impossible to have a solution. So I ended up using up to 7 crossings, and the clue phrase letters also yielded a very thematic set of words for the Twisty Wordsearch, including one fun pun reference. The grid fit required a fair amount of fine-tuning to get the proper extraction, and to avoid alternative solutions, but eventually it all came together.

This was one of the last puzzles finalized in the Hunt, as I only decided to write this quite late, and it involved other writers, as well as a few edits and retests to fine-tune. Early on in the year, Palindrome had invited members to set up a large number of discussion channels on potential puzzle themes in our Discord server, with the aim of bringing people together to seed and collaborate ideas for puzzles, and possibly to avoid repetition in themes/ideas. This worked fairly well at bringing together members with common interests in different themes to brainstorm puzzle ideas, but not so much in cases when someone may already have his/her own puzzle idea in a particular theme to write on. So one of the channels created by someone was on The Genius Korean reality game show. And being a huge fan of the series, I was pleasantly surprised to find a number of other members having similar interest. However, the channel discussion was mostly fan banter, and there was no clear potential puzzle idea or direction to pursue. Fast-forward to August, and I was thinking of possibly writing another puzzle, and looking through the discussion channels when I noticed this one, and thought what a pity it would be not to take this opportunity to write a puzzle on The Genius. My first instinct to approach this was as usual coming up with a direction that I felt would honor/highlight the key attributes of the theme and source material. In this case, it made sense to involve and showcase the intellect and strategy games central to the show. However, making these into a Hunt puzzle is non-trivial. I then had the brainwave to position these head-to-head games as mini-puzzles that would somehow yield single-word answers, and these feeder answers were then used in a final mini-meta game. I then came up with a thematic mini-meta, which confirmed that this approach and puzzle was workable, and interesting even for non-fans. Given the original interest in the channel, I also decided to invite other collaborators who were interested to work on any of the mini-puzzles, even though I could probably have written the whole puzzle myself if needed. I'm pretty used to writing puzzles alone, but Hunt is a rare opportunity to collaborate with multiple writers too. In the end, I had a lot of fun writing this puzzle, and was really satisfied with the outcome and glad that I had gone ahead to make it happen even though it was late in the process. Special credit also goes to our amazing Tech Lead Sahil Bhasin, who coded up the interactive versions of the Open, Pass and Twelve Janggi games for this puzzle. It was really surreal to see these games come alive in this form, and solvers who were The Genius fans also said they loved these in their post-Hunt feedback. Thanks too to our Art Director Justin Ladia for the quick turnaround on the tweaks to the garnet image, which were needed for the mini-meta retests, and ended up working great.

This puzzle was actually mainly written by Shai Nir, and I was the editor. This puzzle employed 13 logic mini-puzzles, and Shai invited any members who liked constructing logic puzzles to chip in on these. The interesting twist is that these logic puzzles used a variant triangular grid instead. So I contributed a token Slitherlink puzzle, which was not difficult to construct, but was more about coming up with an interesting solution and solve path.

This puzzle idea was another one which I had in my folder for around 6-7 years, and had been saving for Hunt just because it would be less suitable for a smaller hunt. I am a chess enthusiast and play various forms such as International Chess, Chinese Chess, Janggi and Shogi. Amongst these, Shogi is arguably one of the most intriguing and complex, with its large standard board size and number/type of pieces, including promoting into new pieces fairly easily, and the ability to drop captured pieces back on the board. So I had thought about writing a Hunt puzzle themed around Shogi. And the chance discovery of the lesser known Dai Shogi variant, and the names of its pieces, gave sufficient interesting material to put together this puzzle idea. The concept is actually similar to a traditional chess problem, with the aha being finding this connection from the totally different setup using images of animals from animation films. Figuring out the solution for this chess problem is hopefully fresh for solvers, given the need to learn the unfamiliar and unorthodox moves for each piece, somewhat akin to a Fairy Chess problem. An humorous insight is that the moves seem somewhat related to the name of the piece, e.g. the Blind Tiger can move to any adjacent square except the one directly forward, while the Drunk Elephant can move to to any adjacent square except the one directly backward. I also include the special Igui ("stationary feeding") move in the solution, which is a unique move for a chess game and tricky to find in this instance, as it allowed a defending piece pinned to the King to move and capture a checking piece, and return to its original square in the same move! Overall, I was pleased with how this puzzle came together, although some solvers were a little disappointed as they thought it might have been a Furries-themed puzzle from the setup (and even asked about this during the post-Hunt Reddit AMA)!

This was another of my puzzle ideas that made use of related themes, in this case three - alternate words, alternate identities, and alternate characters in a different franchise. The first part of the puzzle and aha on finding alternate words in British English was fairly straightforward. Construction-wise, the challenge was finding enough good examples/pairs that had the necessary letters to extract out the alter ego names for the second part. This was not easy as the names were constrained by the third part, and some had uncommon letters, while others had many letters. The constraints on the third part were the hardest, as there was no canonical list of matching DC and Marvel superheroes, even though there were many examples of acknowledged similarities. So I chose only the ones who were the most unambiguous matches given across different sources, and that proved to be okay. One part which some solvers seemed to miss was the hidden extraction phrase clueing the aha in the third part. The phrase was extracted from DC superhero names based on the matching word pattern in the clues, which also served as a confirmation of the solving steps during the first part. The intent was that the somewhat contrived phrasing of the clues and the puzzle title would hint this extraction, but I knew that the aha for the third part was also not too difficult, and solvers could probably make the intuitive leap (which they did) even if they missed the helping clue phrase.

As a remote team member, my involvement on logistics during the Hunt weekend was somewhat more limited, but I was able to contribute in the online admin like hint answering. This was something I was quite used to doing, and it was probably the only opportunity to interact with many of the smaller teams, so I did my best to make the responses interesting and helpful, especially for newbies. The reality of how Hunt is solved by a lot of other much smaller teams really strikes, especially for those who have been solving only with the experienced/large teams. One slight challenge during hint answer was that with Hunt having so many puzzles, even as an editor and testsolver, I probably have only seen just a little more than half of the puzzles. Hence I mostly answered hints on the puzzles that I was more familiar with where possible, since there were a couple of hint answers during every shift. I also helped out during one of the interaction Events, and joined Mark in a Zoom call to go through the Book Reports with one of the teams. As usual, it was fun tracking the progress of solves and teams behind the scenes. I think we were quite relieved about the progress of the lead teams, who were mostly in line with our Hunt length estimates, as we had set out for Hunt to finish by Saturday/early-Sunday latest. Sitting in during the Runaround Zoom call with teammate was also exciting. Incidentally, all four of the single-word answers in the puzzles I wrote were used in the Runaround. teammate blitzed through this part, but got stuck a while at spotting the aha for the final extraction, before finally finishing the Hunt. Wrap-up was arranged for later on Sun, so there was some internal coordination to put together the slides. But it was clear that besides sharing some of the fun/interesting details from the Hunt organizing process to participants, we also want to showcase the solving teams and some of their wonderful submissions. When the Palindrome team members were asked to appear on the video before the credits rolled, that was a really poignant moment for me, because the realization that Hunt and this year-long process/project was truly over finally sank in. As mentioned at the start, I thought there were areas which were not ideal during execution, but by and large, I felt we generally delivered a Hunt and experience that solvers enjoyed. And most of all, I deeply cherished all the fun times and memories spent with Palindrome members to put this Hunt together.

If you have any thoughts or questions for me about Mystery Hunt 2022 or the puzzles I wrote, feel free to email me at puzzlesmiths@sgpuzzlehunt.com. I also have a write-up on my puzzle writing journey. Always happy to chat about hunt puzzles and puzzle hunts with anyone with an open mind. And looking forward to the next opportunity to be involved in the organizing of Hunt again!