Thoughts about the 2021 MIT Mystery Hunt

Over the past year, I co-led my team, the ✈✈✈ Galactic Trendsetters ✈✈✈, as we wrote the 2021 MIT Mystery Hunt. Jakob and I served as the “Benevolent Dictators” of our writing team, meaning we were in charge of coordinating our team’s hunt writing efforts, and we led or were involved in most of the big decision-making throughout the year. I’ve been involved in writing some other puzzle hunts before((I’ve been hunting since 2009, and I previously played a large role in writing three GPHs, a medium role in a BAPHL, and a small role in MH 2013.)), but I spent at least an order of magnitude more time working on this than any previous hunt or similar project. Throughout the whole process, there’s one particular question that I thought about a lot, and that I haven’t seen much written about before, and so I decided to write this post to focus on it:

What makes it the MIT Mystery Hunt?

When our team decided to run a virtual Mystery Hunt, we all agreed that it was important to make our hunt “feel” as much as possible like a “real” Mystery Hunt. We had some different ideas about what that means, and when we surveyed the community early in the summer, they gave us even more answers.

It makes sense that the community has many different answers to this question. The Mystery Hunt has been going on for a long time and has evolved a lot over the years. It’s also large enough that no participant can see all of it, and it’s easy to get sucked into the parts that excite you most and mostly ignore the rest. Of course it means something different to everybody.

For me, some of the most important aspects of the Mystery Hunt that set it apart from other puzzle hunts include:

  • Its large scale: recent hunts have had more than 150 puzzles, and teams with more than 100 people. There’s something really incredible about being on a (large or small) team and tackling such a huge challenge together((There’s an ongoing debate about whether the hunt has become too big. My opinion is that hunt size is just one proxy for something much more important: whether teams of all sizes have a good experience participating in the hunt. We tried to make sure smaller teams would have a good time by doing things like spreading a large round of easier puzzles throughout the hunt, aggressively giving out and answering interactive hints, and giving story interactions based on time instead of progress through the hunt. I think the best argument for making the hunt smaller is to make less work for the writing team. I would be totally happy with Palindrome or any future writing team making a smaller hunt if that’s what they want to do, but I don’t think it makes sense to ask the team to make a smaller hunt than they want to (assuming someone wins the hunt in a reasonable amount of time).)). Also, many types of puzzles and challenges can only reasonably appear in a Mystery Hunt because of the scale, either because the puzzle is relatively long or hard, or because it requires knowledge or skills that you can’t assume a team of fewer than 50 people would have.
  • Its connection to MIT, not just physically, but also in content and spirit. MIT’s culture and values mesh really well with the Mystery Hunt, and MIT is full of quirks and trivia and warmth that make for great puzzle content. I think Nathan and CJ explain this well. In addition, MIT provides tons of space, advice, and other resources to the Mystery Hunt for free, without which the Mystery Hunt would be entirely infeasible to run.
  • It serves as a large social gathering of puzzlers and an opportunity to spend a weekend with friends from out of town. Mystery Hunt is typically the one time each year that I see many of my ✈✈✈ Galactic Trendsetters ✈✈✈ friends in person, and I get to say hi to friends on other teams during kickoff and wrapup and the days before and after. But that’s not all. Mystery Hunt is a great communal experience, in which you struggle together with your teammates on some insane challenges, then as you walk down the hallway to grab some food, you smile knowingly at the people you pass who have been doing the same.
  • It has many different traditions, including types of puzzles that appear (scavenger hunts, runarounds, conundrums, etc.), ways that participating teams interact with the running team (phone calls, visits to the team HQ, visits to rooms set up for interactions, etc.), “puzzle-y” events and interactions throughout the weekend, and more.
  • It has a relatively detailed story which is introduced at kickoff and develops over the course of the hunt. A big collection of puzzles combined with other interactions and events provide for an unusual medium to present an interesting story.

In this post, I’ll discuss some of what our team and I did to make our hunt an MIT Mystery Hunt. I’ve tried to focus on details and anecdotes about decisions we made that I think hunt participants might find interesting. I’m going to assume the reader is familiar with our hunt already; there will be some spoilers for our hunt itself, and some for individual puzzles, but I’ll try to hide all the major spoilers.

Our ⊥IW Theme

One of the first things that a hunt writing team has to do is pick the theme. I think this timing is really unfortunate. Picking the theme is probably the most contentious part of the whole hunt writing process. It involves a bunch of smaller picture debates (“would this plot detail be more entertaining than that plot detail?”) and bigger picture debates (“is this aspect of a hunt more important than that aspect of a hunt?”). It also requires voting on which creative ideas are the best, which can lead to hurt feelings. Making a Mystery Hunt writing team — a big team full of friends, but also many strangers, who often come with strongly-held and differing opinions about what makes a good puzzle hunt — go through this process before they’ve gotten a chance to work together and get to know each other is almost cruel. And yet, since the theme choice blocks almost all the other work that has to be done for the hunt((One task it doesn’t block is reserving blocks of hotel rooms at the hotels near MIT for out-of-town participants to stay in. In retrospect, I’m very glad that we nonetheless moved pretty slowly on this task and hadn’t committed to anything by the time we realized it might be a bad idea.)), it has to be done first.

Nonetheless, I’m very happy about the theme we ended up picking. At a high level, it consists of two parts:

  1. The Story: Professor Yew is trapped in an alternate version of MIT, called ⊥IW. The buildings of ⊥IW each have an issue that we need to fix in order to bring the professor home.
  2. The Structure: The hunt takes place in an MMO representation of ⊥IW which is used to unlock puzzles, and which is itself full of puzzles.
The smobster, a dangerous adversary!

My favorite part of our whole hunt is the round theming. The Infinite Corridor is infinitely long. The clusters have confusing, foreign names. East Campus isn’t sufficiently on fire. The Charles River is full of smobsters which can only be defeated once you obtain a pirate license.

MIT is full of great puzzle inspiration, and it feels like we only scratched the surface. I still feel goosebumps listening to the solve sounds that Lennart put together for each round, and thinking about what solvers must have been feeling when the Chorallaries started singing The Engineers’ Drinking Song, the EC fire alarm went off, or the B52s started playing their iconic Rock Lobster bass line. (Okay, that last one doesn’t have much to do with MIT.)

But beyond just the puzzles, MIT was reflected in the NPCs and the problems they were facing, in the story characters’ personalities and ambitions, in the little details throughout the virtual world, and in the spirit of the whole event. I think of our hunt as a celebration of MIT, and I hope that came across to all the participants.

One issue that was brought up during the theme selection process, and that I kept thinking back to as we put the hunt together, is how people unaffiliated with MIT would perceive and appreciate the hunt. I know that personally, how much the theme resonates with me has had a big impact on my enjoyment of past Mystery Hunts. Would our theming making these participants feel alienated? Although it’s the MIT Mystery Hunt, this is a serious concern: this year, almost 60% of the participants had never been affiliated with MIT. I think the remote nature of the hunt made this a bit higher than usual. 

My hope, and a goal of mine and many of my teammates as we put the hunt together, is that the hunt taught these participants about MIT, and let them experience its culture (at least a little bit). In the end, it’s not really my place to say whether we succeeded at this or not. But, it makes me happy to see that we haven’t gotten much, if any, negative feedback about this on our post-hunt survey. When I asked about this afterwards, my teammate Amon said the most heartwarming thing:

“I wasn’t an MIT alum, but I feel like a ⊥IW alum.”

Solvers spawned next to EC in ⊥IW, but it doesn’t really feel like EC until you solve the meta and set it on fire.

The Projection Device/The MMO

To be honest, at first, I wasn’t too enthusiastic about the social/communal benefits of the MMO. I was much more excited about the MMO as a way to find and unlock puzzles, a bit like playing a Professor Layton game, and how exploring the MMO could be a puzzle in itself((I ended up helping with the navigation puzzles in four of the rounds, including helping to code the navigation puzzles in the Infinite Corridor and Tunnels.)). Among the GPH writers, we’d been brainstorming how to make an idea like this work for years, and many of us felt that Mystery Hunt afforded us the opportunity to try it out. 

Nonetheless, some of the most popular parts of the Projection Device were definitely the non-puzzle parts: the ability to see other players and interact with them using emotes and whiteboard messages, the detailed character customization options, all the different areas of campus one could explore, the scattered NPCs with varied dialogue, and so on. During our theme selection process a year ago, we had a lot of arguments about whether these “soft benefits” of the Projection Device were worth all the work it would take for us to make them. The argument was that solvers would be too busy solving puzzles and talking with their teammates in person and walking around real MIT to really take advantage of these features, and might feel FOMO if they don’t have the time to explore them all. Hearing Nathan (the “champion” on our team for these features) argue about the communal nature of the Mystery Hunt, and how these parts of the Projection Device would complement and expand on it, made me think more carefully about and appreciate the time I normally spend at Mystery Hunt doing things other than just solving puzzles. I realized how important these features would be, and I think they would have been really appreciated in an in-person Mystery Hunt.

That said, the pandemic and remoteness of our hunt definitely amplified their importance and their impact. So many people on our team understood this and helped to develop these different parts of the Projection Device, and I think the diversity of dialogue, clothing styles, scenery, and so on, was only possible because of how many different people were contributing. Even though a lot of coding was needed to put together the MMO, we also needed to do plenty of other work, including art, navigation puzzle design, scene layout, NPC placement and dialogue writing, and I think the big variety of work helped contribute to the excitement for the idea among our teammates. 

During the theme selection process last February, I wrote this list of ten different reasons why people would like the MMO:

  1. Exploring virtual MIT will be a lot of fun, especially for people familiar with real MIT. It will also directly help your team by unlocking fish puzzles((“Fish puzzles” is what we originally called the puzzles in our Students round. We designed it to be a large round of shorter and more approachable puzzles, to be interspersed among the other more intense puzzles in the hunt. The round was heavily inspired by the School of Fish round from the 2015 Mystery Hunt, hence the name.)), so it won’t feel like a “waste of time” for people on more competitive teams. Some puzzles (e.g. virtual runarounds) will specifically encourage it.
  2. It is fun to generally look at/appreciate how the structure of each round of a mystery hunt works. Normally this just consists of looking at the HTML page for the round. Here, it will instead be a few minutes of exploring the building in the MMO. This will be more exciting. I think almost everyone will want to spend some time just exploring the buildings and how the rounds in them work.
  3. If something interesting happens within the MMO, everyone can go see it! E.g., you can tell your teammates “come see this neat statue outside East Campus!”. A lot of times in typical MHs, most people miss out on most things like this. (Note: we will make sure that everything in the MMO can be experienced multiple times for this reason)
  4. Relatedly, events in the MMO can be much more widely attended and appreciated than events in previous MHs. I think we can make some really amazing virtual events. Moreover, “spectators” to a virtual event can put as much or as little time into watching it as they want. For instance, they could bring their avatar there and occasionally glance at the event while they’re otherwise working on a puzzle.
  5. The MMO lets us make puzzles that interact between real and virtual MIT, such as a runaround incorporating both MITs.
  6. Exploring the overworld will be a good opportunity to interact with friends and classmates on other teams who are hunting. Normally it’s unlikely you will see them other than at kickoff and wrapup, but now there are plenty of opportunities.
  7. More generally, seeing everyone else in the overworld will be exciting and really give you a sense of how many people are participating in the MH. It’s easy to forget that there are 3500 other people participating. Seeing all these people, and small interactions you may have with them (e.g. sending an emoji when you see them in an obscure corner of campus) will lead to a stronger sense of community and more memorable personal/emotional connections.
  8. The MMO works as a powerful narrative device and is incorporated heavily into the plot. It’s so much cooler in a hunt about alternate dimensions if you actually get to walk around in that alternate dimension!
  9. The MMO is surprising and exciting! People like to be surprised by new innovations in the MH! The MMO is right up the alley of the nerdy culture and participants of the HM.
  10. Making a virtual MIT is a great way to emphasize and celebrate the MIT-ness of the MITMH.

A lot of work had to be done, and a lot of design decisions had to be made, between when we came up with a high-level plan for the MMO during the theme selection process (and I wrote the above list), and when we finally finished making it in January. Nonetheless, I think the list above ended up being pretty accurate.

Did you find this guy behind the big dome?

Walking around ⊥IW during the hunt and seeing everyone exploring the world was a highlight of the Mystery Hunt for me. I particularly enjoyed walking around the Killian Court scene and leading participants to the secret NPC who gave out the ✈✈✈ emote.

Anyway, I could talk on and on about the MMO, but I think Nathan described it really well, so I’ll defer to his writeup and move on to some other aspects of our hunt instead.

The Story

The original theme proposal laid out the main plot points of the story that remained mostly unchanged in the final version. When we wrote the proposal, we already knew we wanted an MMO, and that the rounds should be based on buildings at MIT, and our main focus was on how to write a story that naturally incorporated those two constraints.

Our story director Lillian then took the basic and constrained story we came up with and fleshed out the characters to have depth and personality, augmented the plot points to be more nuanced and exciting, and wrote a fun script to present it all. Rather than describe the story in too much detail, I’ll link you here to watch it yourself! I got to play the role of Cameron, which was a lot of fun.

Pretty much every puzzle hunt has a story, and some other puzzle events even have stories which are quite complex. (One event I participated in which comes to mind is the Miskatonic University Game, and my impression is that this is the case for many instances of The Game.) Nonetheless, the scale of the Mystery Hunt means it’s able to convey a bigger story than most other puzzle events, at least. We also had a lot of freedom to express the story through many different media — puzzle and round pages, pre-recorded videos, live Zoom calls, emails, NPCs in the MMO, and more — which is rare or impossible in a puzzle event other than Mystery Hunt.

It’s intuitive when writing a hunt to make the puzzles first and then build the story around them. This way, you give as much freedom as you can to the puzzle writing; It’s dangerous to have the story put constraints on the puzzles, since the puzzles are already challenging and time-consuming to write without any additional constraints. Indeed, this is roughly what we’d done in the past for GPH.

On the other hand, the more intertwined the story and puzzles are, the better the experience is for participants. Writing a story after the puzzles are already in place can make the story feel tacked-on. For our Mystery Hunt, we wrote our story and puzzles in parallel. I think both aspects benefitted from doing it this way, but it was indeed quite challenging to coordinate the two without putting too many constraints on either or having them step on each others’ toes. This might just be unavoidable given how passionate we are about both parts of our hunt.

One decision we made differently from past Mystery Hunts was to have the major story interactions unlock at certain times instead of when the team had made a certain amount of progress. This had a lot of benefits. It helped with our logistics: a lot of people were involved in making the story interactions run smoothly, and it helped that we could plan ahead of time for when they would be. Relatedly, we could make sure these interactions didn’t happen at times when most participants would be sleeping, and it helped us to plan when our actors would get to sleep. It also meant we could more consistently make sure teams got to experience the story. We wanted to make sure that even teams who weren’t solving many puzzles got the opportunity to experience the story, and we did this by having a final livestreamed run of each interaction.

On the other hand, I think doing this worked against the intertwining of story and puzzles. Plot points are easier to follow and more exciting if they’re related to the puzzles you just solved, and solving a puzzle feels more rewarding if it leads to a new story interaction. I was worried about this leading up to the hunt, but in retrospect I think it made a small difference, if any. The plot interactions were still shiny and exciting on their own, and solving a puzzle was still rewarding, both in terms of unlocking puzzles and progressing through the hunt, and generally as solving a puzzle always is. (Also, if you include exploring the Projection Device as part of the story, then solving a puzzle almost always led to more that you could explore there.) The post-hunt survey results seem to agree with our decision as well: under 6% of survey respondents disliked that some story interactions were not linked to puzzle progress.

Physical Puzzles

We wanted to make sure to have physical puzzles in our hunt. Physical puzzles are something we haven’t been able to make for GPH in the past, and some of the most memorable puzzles from past Mystery Hunts are physical puzzles. I think we came up with four puzzles that use the physical medium in interesting ways. Huge thanks to Joanna and everyone who helped with the logistics of physical puzzles for the tremendous amount of work it took to pull them off. Assembling hundreds of puzzles when we couldn’t be together in the same room, and then shipping them around the country, were very time-consuming tasks that we didn’t anticipate having to do when we won the 2020 Hunt. To make it more exciting, we did this all during a period of historically bad shipping delays.

DD’s shirt design for Green Tee

I helped to write two of the physical puzzles. I contributed a little bit in brainstorming to Green Tee, a puzzle where you use the Zappar augmented reality feature together with the really awesome shirt design by DD. We contemplated having the puzzle be on a custom-printed mask instead of a shirt, but some of our teammates argued that Mystery Hunt should be a distraction from the pandemic, and participants wouldn’t want to have to think about masks during the event, so we went with a shirt instead. I’m glad we did, since the shirt turned out great and we got hundreds of additional orders for it after the hunt.

I worked a lot more on Water Bottle, a puzzle I wrote together with Jakob (spoilers for this puzzle in the next couple of paragraphs). We mailed participants unlabeled bottles of flavored water, and a key step of the puzzle involved tasting them and figuring out what the flavors are. Puzzles requiring tasting and smelling show up somewhat frequently at the Mystery Hunt, and pretty rarely in other hunts, so I was hoping this would feel like a “real” Mystery Hunt type of challenge for solvers. I’m also just generally a fan of flavored water and thought this would be a funny idea.

Hint Water, the brand of flavored water we went with, worked really well for the puzzle for a few reasons. First, sealed bottles are relatively easy to mail, especially compared to some other foods we were considering sending to teams. Second, Hint is very careful to avoid allergens, artificial sweeteners, preservatives, and so on — as far as I know, no participants were unable to work on the puzzle because of dietary restrictions. Third, Hint Water is common enough in stores that we figured some solvers could buy some during the hunt to do taste comparisons (and I’ve heard at least a couple stories of teams doing just that). Fourth, and maybe most importantly, their brand name led really naturally to the last step of the puzzle.

Probably my biggest contribution to physical puzzle logistics was working with ClueKeeper (who let us use their AR features) and Hint (who gave us thousands of water bottles) to have them to sponsor our hunt. I think the MIT brand and the huge number of participants help a lot in convincing companies that the MIT Mystery Hunt is worth sponsoring. Both companies were very accommodating and easy to work with despite some relatively weird requests we made of them. I’m very grateful to them, since we couldn’t have reasonably made either of these puzzles without their help. One of my favorite moments of the Hunt creation process was when a truck showed up to drop off 1500 Hint Water bottles at Jakob’s apartment in late December, and the bottles appeared in the background whenever Jakob and I had video calls. Huge thanks to Jakob for removing the labels from, repackaging, and shipping out all those bottles, almost entirely by themself!!

I didn’t help with our other two physical puzzles, Squee Squee and The Greatest Jigsaw, but they’re both excellent and I recommend that you check them out. The Authors’ Notes in the solution to Squee Squee go into detail about some of the crazy logistics involved in sending out these puzzles.

In A Park, Our Runaround Puzzle

I really like the idea of runaround puzzles((I’m talking here about puzzles in the hunt that involve going to places around campus or nearby, not “story runarounds” like the one that often happens at the end of the hunt leading to the coin, although the two types of runarounds have plenty of similarities. )). I’ve even written a couple before((Missed Connections from the 2013 Mystery Hunt that I wrote with Leon and Kevin, and Visitor’s Guide, a “virtual runaround” I wrote for GPH 2018.)). There’s something really nice about finding and solving puzzles in the area around you, especially if you end up discovering something new about your surroundings in the process. Solving a runaround is also a great excuse to get up from your computer and walk around a bit in the middle of a puzzle hunt. 

On the other hand, in the heat of a Mystery Hunt, I often have trouble convincing myself to actually go solve runaround puzzles. I think there are two main reasons for this. First, I really enjoy being in our team’s HQ during the hunt so that I can be with my friends and get a sense of all the exciting things that everyone is working on, so I don’t like the idea of leaving for an unknown period of time to work on a runaround. Second, runaround puzzles are the worst puzzles to get stuck on, since double-checking work often requires spending a lot of time walking back to places you’ve already been. If the solving process isn’t very smooth, runarounds can become a pain. I have solved some runaround puzzles((Back on the Move from 2015 and Little Passages from 2018 come to mind as particularly fun ones.)), but especially more recently when I’ve been trying to solve more “competitively”, I’ve mostly avoided them.

It felt important to me and many members of our team to have a runaround puzzle in our hunt, and so Jakob and I led a big group effort to make In A Park. After some brainstorming, we decided on the broad idea: the puzzle would consist of a big list of clues from lots of different parks, and to solve the puzzle, the solver would have to go to any one park and solve the clues there. 

We knew we wanted the puzzle to be fairly easy, both so that we wouldn’t have any solvers go to a park but fail to solve the puzzle, and because we wanted to make the solving experience relatively consistent across different parks; with all the other logistical considerations in making this puzzle, easy was really the only option. Jakob and I did some testing to figure out what kinds of clues are easy and unambiguous enough to solve. Jakob walked around a park near their apartment one evening while talking to me on the phone, and I looked through the park on Google Maps and made up clues to see how long it would take Jakob to solve them. A couple of those clues wound up in the final version of the puzzle. One reason we didn’t use more is that we tried to make sure solvers would have to actually visit the park to solve the clues, and not just look at maps online.

In an earlier version of the puzzle, solvers had to solve clues in their local park interspersed with clues about national parks that one needed to research on a computer to solve. We thought this would lead to a nice experience where the solver in the park would have to work together with a teammate at a computer to progress through the puzzle. In testing, we found that either the national park clues were too hard, or else the solver at the park would just solve them on their phone, so we scrapped the idea.

One of the more interesting challenges in making this puzzle was finding people in enough different places to write the clues for us. As part of hunt registration (which we put out a lot earlier than usual), we asked teams where in the world they’d be solving from. We aimed to have clues for at least one park in each popular area, so that as many teams as possible could solve the puzzle. We ended up asking some friends and friends-of-friends in order to cover a few cities where nobody on our team lives. A big debate among our teammates before the hunt was whether we’d covered enough parks so that most teams wouldn’t just backsolve the puzzle. I think we did a good job — we ended up with clues for 25 parks throughout the world, and more than 50 solvers told us in a survey that they forward-solved the puzzle during the hunt.

Since we had so many different clue writers with a variety of puzzle writing experience, we wrote detailed and fairly explicit instructions for clue writers to follow. We emphasized that clues should be fairly easy and unambiguous, not take too long to solve, and no “trick questions”. Such clues can feel boring for constructors to write, but our testsolvers usually found them interesting to solve anyway.

Getting all the clues written and compiled was like organizing a mini puzzle hunt, including its fair share of poking clue writers to send us their clues or rewrite problematic clues. We asked clue writers to take pictures at their parks so that we could attempt to fact-check all the clues. This was probably the hardest puzzle to fact-check in the whole hunt, and we didn’t expect that we’d catch every mistake, but we ended up doing pretty well: we only had to fix two minor errata during the hunt. 

So, would I have enjoyed this runaround as a solver? I think we made the puzzle easy enough that I would have, but I suspect I would have preferred walking around ⊥IW instead, where I could take a break from solving and explore without having to leave my (virtual) team HQ. Incidentally, we also had Projection Device runarounds: DD, Nathan and I wrote a ⊥IW runaround called Stud Finder, and Jakob wrote one called X Marks the Spot. Before the pandemic, we had some grand plans to make simultaneous runarounds in MIT and ⊥IW, which we had to drop for the most part, but we tried to recreate them as best we could for the endgame.

The Scavenger Hunt

Every Mystery Hunt I’ve participated in (and most before that) has included a scavenger hunt puzzle. I rarely help our team’s scavenger hunt efforts much, but I always enjoy seeing what creative things my teammates put together. A particularly memorable example for me was 2019’s Taskmaster, when the Taskmaster came to our team HQ to judge our submissions, rather than us bringing our submissions to the judge, so that everyone on our team could watch the (very entertaining) presentations and judgement. 

I helped to make our scavenger hunt, MIT/⊥IW Experimental Evidence, along with a bunch of teammates. Between our hunt’s theme and its remoteness, it made a lot of sense to theme our scavenger hunt around things at MIT. Brainstorming ideas for the list of what teams had to bring us was a fun way for me to reminisce about iconic things and events at MIT, and I hope participants got to experience something similar.

When teams had their items prepared, we invited them to a Zoom call to present them to us. I saw a number of participants join the call just to watch, even though they hadn’t helped put together their team’s items. Some participants in the Zoom calls were clearly working on other puzzles for most of the time, but paid attention occasionally. I think the Zoom call presentation was actually very effective at letting people participate as much or as little as they wanted to in the process, and I hope a lot of participants were able to get in on the fun in this way.

I was only involved in judging a couple teams’ submissions, but it was a lot of fun. Spending some time talking about each of a team’s submissions makes the process more exciting for everyone, since participants like to present what they put together, and I like to listen to their presentations. Unfortunately, this just takes too long with so many participating teams, and we ended up needing to instruct our teammates to speed up the process. In retrospect, I think we should have asked for fewer items from each team so that we could have had more time to talk about them all.

We decided to add in a “fourth wall” box to the puzzle page explaining more explicitly what we were looking for from teams to complete the scavenger hunt. We had a lot of new teams register this year, and we figured this would help save them some time and confusion and help keep us from having to reject too many submissions. I wonder whether Mystery Hunts have had issues from teams not understanding what was expected of them in the past. Broadly, it seems that only more recent years’ scavenger hunts have been focused on finding creative ways to satisfy requirements, while older Mystery Hunt scavenger hunts often had much more rigid requirements.

MIT Puzzles

Given our theme, I was never worried that our hunt wouldn’t have enough of an MIT feel to it. Nonetheless, I wondered whether we had enough individual puzzles that were about MIT.

Why is it important to have puzzles about MIT? Two more obvious reasons are that it’s a nice way to thank MIT for all the ways that it supports the Mystery Hunt, and that a big fraction of the hunt participants are familiar with MIT, so it makes just as much sense as puzzles about any topic that many hunt participants are familiar with. Another reason, that I think is less evident but more important, is to get new MIT students into puzzles and the Mystery Hunt. MIT students run the Mystery Hunt through MIT Puzzle Club and are important for preserving the traditions and culture of the hunt. What better way to get them involved than to have puzzles about their school? Regardless, it seemed to me that our MIT-heavy theme achieved these goals better than any MIT puzzle would.

I eventually realized that I wanted to write MIT puzzles, not for any bigger picture reasons, but just because I like MIT puzzles. As someone who likes and knows my fair share about MIT, I would be happy with more puzzles in the Mystery Hunt about MIT, just as I would be happy with more puzzles about other topics I’m relatively familiar with like computational complexity theory or luge or Sudoku. And so, in order to be the change I want to see in the world, I wrote puzzles about all these things.((A bit more seriously, just like most people, I just enjoy writing puzzles about topics I like, with the hope that other people who share my interests will appreciate them. That said, it’s generally a lot harder to start writing a puzzle about a particular topic rather than about a particular puzzle mechanism, and this is still true when starting to write a puzzle about MIT. See, for instance, section 2 of David Wilson’s guide.))

Not counting puzzles in the Projection Device, I wrote three MIT puzzles: Hockfield Court, Numbers (with Jakob), and New House (metapuzzle). I think they hit a nice spot where MIT familiarity doesn’t help too much to solve the puzzle, but the puzzle teaches or reminds you about an interesting part of MIT. I think it’s fair game, or maybe even desirable, to include puzzles in the MIT Mystery Hunt that require MIT familiarity to solve them, but given how much other MIT content we had, I wanted to avoid this. 

Puzzles that can only appear in the Mystery Hunt

I wrote a lot of puzzles for this hunt (somewhere between 23 and 28, or maybe infinitely many, depending on how you count). A big goal of mine was to write some puzzles that could only appear in the Mystery Hunt. I’ve written puzzles in the past for GPH and other hunts, and I expect to write more in the future. Even though we’re relatively permissive about what puzzle types we allow for GPH, there are still plenty of puzzles which could only appear in Mystery Hunt because of their size or subject matter or how long it takes to make them. I think I wrote a handful that fall into this category.

Love at 150 km/h

Love at 150 km/h is my favorite example. Nathan and I wrote this puzzle together, although he did most of the work (he did all the coding and wrote around two thirds of the dialogue). It took months to do the brainstorming, designing, writing, and polishing to put this together (huge thanks to Joie and Leiss for the awesome art and music!), and I think it just wouldn’t have been possible in any smaller-scale hunt. (I put a few stories and anecdotes about making this puzzle below, but I really recommend trying the puzzle out a bit before reading it!)

Click to reveal Love at 150 km/h spoilers

Nathan and I had wanted to make a luge puzzle for a long time. We’d tried a few times in the past without much success, but inspired by examples like the curling puzzle from 2014, we knew that it was now or never.

To make sure we didn’t forget, Nathan added a joke idea into puzzlord, our puzzle tracking software: a luge dating sim (with basically no further details given). To our surprise, our editors liked it, but we unfortunately didn’t actually have any idea how such a puzzle could work. Nathan and I spent hours trying to figure out how to make a reasonable puzzle based on luge. It’s tricky since the general populace knows very little about luge, and most of the data one can look up without a ton of work is just times and scores. We eventually decided that the shapes of luge tracks could make for a good puzzle, so we told our editors about an idea along those lines.

After some discussion with our editors, we discarded this idea because it turned out to be too similar to some other recent Mystery Hunt puzzles. We went back to the drawing board, and finally returned to Nathan’s joke idea of a luge dating sim:

Writing the dialogue was a lot of fun. Nathan and I were well-aligned on the tone and sense of humor we were going for, and we included a lot of plot points and details that weren’t super important for the puzzle, but that contributed to the feel and polish of the game. I thought our editors or testsolvers might complain about this, since “red herrings” like these should be avoided in most puzzles, but they fortunately all approved of and enjoyed these aspects of the puzzle.

The work to put this puzzle together was pretty different from any of the other puzzle writing I did, although it has similarities to making the Projection Device. Overall, I think this is the most successful puzzle I’ve written where the starting point was a topic instead of a mechanic, but it took a lot of brainstorming to get there.

Cafe Five

Cafe Five is another puzzle I wrote with Nathan. One important puzzle hunting skill, which I think I enjoy more than most, is quickly using computer tools to solve clues and puzzles. Cafe Five was, at least originally, a challenge of just doing that over and over. After finding that testsolvers usually wanted to solve the puzzles by hand instead of using tools, we tweaked the puzzles so that this was (for the most part) also a valid way to make progress.

I think a puzzle like Cafe Five could have appeared in GPH (not unlike Jon’s puzzle Race for the Galaxy from GPH 2019), but the scope of Mystery Hunt, together with the mechanics of our Infinite Corridor round, let us flesh it out into a bigger and more varied challenge than would have been possible in a smaller hunt. We envisioned this as a teamwork challenge where typically only a few people would need to be manning the Cafe, but they would have to pull in teammates to help when the puzzles started flooding in, or when a clue about that pesky intro round puzzle they aren’t familiar with showed up. My impression from feedback is that it worked well, though I do wonder whether the hunt being remote impacted the solver experience for this puzzle in particular. Having the puzzle in the Mystery Hunt also let us theme it around our favorite cafe at MIT.

Super Mystery World

Super Mystery World is a puzzle that I originally thought was specific enough that we couldn’t even put it in the Mystery Hunt. Solving the puzzle (or at least fully appreciating it) requires the solver to have a Nintendo Switch and the game Super Mario Maker 2.

Back in May we had a long conversation about it and decided it was logistically infeasible. We estimated that enough solvers probably did own a Switch and the game, but that they likely wouldn’t bring them to Cambridge for the hunt. Once it was decided that the hunt would be remote, so that everyone would be at home with their Switches, we changed our mind and green-lit the puzzle. This puzzle might be the only part of our hunt that very clearly benefited from the hunt being remote.

I really enjoyed making this puzzle together with Amon, Ben and Leo. My main contribution was making level 1-2, although Amon improved the platforming parts of my first draft of the level. I recommend playing the levels if you have the game, or watching the playthrough video we made if you don’t.

As we expected, this puzzle is a lot easier for solvers who have played SMM2 before. Our testsolving group who already owned the game solved the puzzle in a couple hours, and our testsolving group who didn’t compared it to huge and challenging puzzles from past hunts. Trade-offs like this, where you make a “niche” puzzle which is interesting or challenging for solvers familiar with the subject material, but in exchange, quite difficult for solvers unfamiliar with the material, are sometimes acceptable in the Mystery Hunt, and pretty rarely in other hunts. Which leads me to…

Complexity Evaluation

I made Complexity Evaluation, a puzzle which involves classifying problems into complexity classes. The puzzle is basically impossible if you haven’t studied this type of material before. I thought this would be fine, since 6.840/18.404 (Theory of Computation) is a popular class at MIT. Despite nerfing the puzzle at least twice in the construction process, I think I still overshot the difficulty, and either a more advanced class like 6.851/18.405 (Advanced Complexity Theory), or a lot of coding, is probably really necessary to solve the puzzle. Fortunately, MIT has enough nerds that 10 or 11 teams forward-solved the puzzle during the hunt (I know at least one person who solved it was a classmate of mine back when I took 6.851/18.405), but in retrospect, I should have tried harder to make it at the 6.840/18.404 level. I’m very happy about the idea and construction, though, and I hope solvers enjoyed the puzzle and maybe learned something while solving it!

More broadly, the fact that MIT has so many nerds means that the Mystery Hunt can have puzzles that require relatively complicated mathematical or scientific knowledge to solve. Mitchell and I wrote another one, Fun With Sudoku, that I’m proud of. Because of the composition of our team, we ended up with even more technical puzzles like this than usual.

A Collection of Conundrums and Riddles

The next few paragraphs contain big spoilers for this puzzle, so I’ve put them behind spoiler tags.

Click to reveal A Collection of Conundrums and Riddles spoilers

Similar to how companies are often willing to sponsor the MIT Mystery Hunt, companies, publications, and famous people are often willing to “plant” puzzle materials or otherwise help to write puzzles for the Mystery Hunt. One example is our puzzle A Collection of Conundrums and Riddles that I wrote together with my teammates Anderson and Colin, Scott from Bloomberg, and Zach from FiveThirtyEight.

When we started writing the puzzle, all we knew is that we wanted to plant parts of it in these two publications. We found that this is a challenging first idea to base a puzzle off of, and it took weeks of brainstorming to come up with a puzzle we were happy with.

One interesting question which arose when writing this puzzle was who to list as the “author” of the puzzles which appeared in the two publications. Both publications usually list the author of that week’s puzzle, and it would have been conspicuous not to list anyone. In order to avoid red herrings, and to clue solvers that they were on the right track, we decided to use “Barbara Yew”, the main character from our hunt’s plot. The potential downside to this is that, since these were published before the hunt started, and we’d already announced Barbara’s name on our hunt registration website, solvers might know that these are related to our hunt and solve them before kickoff (and indeed, we’re aware that at least a few teams did this). We decided to go with Barbara’s name since having a clear “this is relevant” confirmation felt more important than this concern to us.

I, at least, didn’t mind solvers solving parts of the puzzle ahead of time; it’s a nice way to let solvers feel like they’ve “cheated the system.” In a more extreme example of a similar phenomenon, some solvers also found the TikTok videos that I made for the puzzle It’s Tricky before the hunt, and at least a couple teams solved the puzzle ahead of time. I think this is hilarious, but maybe it’s something I should have been more careful to avoid. I wonder whether those teams had an easier time navigating the Tunnels of ⊥IW.


We put on some awesome events! I already mentioned above how much I like their theming as acquiring a pirate license in order to save the Charles river. We did multiple tests beforehand of each event to try to make sure it would run smoothly, and so I got to “participate” in each event at least twice, and I enjoyed all of them.

When researching past Mystery Hunt events, we found that there’s a lot of variety in what’s involved. Some events require you to solve a puzzle at the end to get the answer, while others hand you the answer as long as you showed up and participated. Some events have all the participants work together, while others have different teams racing against each other to complete tasks, or even directly competing against each other. Some events are more puzzle-heavy, while others focus more on creative tasks. 

Pretty much all Mystery Hunts require a team to attend all the events in order to win the hunt, but exactly how this is enforced has varied a lot. Some years it’s just stated that the events are mandatory, while other years, the events tie in to the puzzle structure in one way or another. It’s not so clear to me that smaller and less competitive teams are actually incentivized to attend events most years, rather than working on all the other unsolved puzzles they have instead. We tried to make sure that all teams would find it worthwhile to attend our events by testing them extensively to make sure they were fun, and by giving an exciting reward at the end: the Attack of the Lobsters game in the Projection Device.

We had to do a lot more technical work than is typical for Mystery Hunt events in order to host our virtual events. We ended up running the events in Discord servers, both for logistical reasons, and to try to give everyone a sense of community and togetherness for the events. In addition, we made plenty of Discord bots and other web tools to help coordinate hundreds of people participating in the events at the same time. We knew that not all our participants would be familiar with Discord or the other tools we were using, so Justine made some guides to teach participants what they needed to know, and we generally focused a lot on making sure our explanations of the events were clear and helpful. I wasn’t very involved in putting together the events, but I’m really proud of what our awesome events team put together.

What’s Next?

One of the best parts of the Mystery Hunt is that it’s always changing. Each writing team gets to choose what’s important to them, and there’s no need to stick with what I or anyone else thinks should be in the hunt. I know our team didn’t include some things which others consider to be Mystery Hunt traditions, and I’m excited to see which of those Palindrome decides to bring back.

I don’t anticipate playing a major role in making another Mystery Hunt in the near future, but I’m sure I’ll be involved again in some capacity one day, and it will be fun to look back on this post and see what’s different.

I want to finish by mentioning that some of my teammates have also written nice posts about various aspects of our hunt which already cover some other topics I was thinking of writing about (Brian, CJ, Nathan, Rahul; probably more are still to come). Check them out!