We won the 2026 MIT Mystery Hunt

This post contains unmarked spoilers for puzzles from the 2026 MIT Mystery Hunt.

I am not in the custom of writing recaps for puzzlehunts. This is mainly because a) I’ve done so many, and b) I don’t always have interesting things to say about them – or not things I couldn’t better express by writing a hunt of my own.

But winning Mystery Hunt is a large enough personal achievement to be worth documenting. Not to mention that I want to get these memories down on paper before they escape.

If you don’t know, the MIT Mystery Hunt is the largest puzzlehunt event of the year. Hundreds of teams, some of them numbering 80+ or even 100+ people, compete for multiple days to solve (150+, even 200+) devilishly difficult puzzles. The team that wins writes for the next one.

Here is the précis: I hunted with Providence, and we won. We are so excited to write this year.

This will be a play-by-play of what I did during hunt: the puzzles I worked on, the experiences I had, what I saw and thought was cool. It will, however, be light on my general analysis. If I think something is great, or not so great, why would I post about it here when I can directly help to adjust the next hunt?

Thursday

After a safe flight from PIT->BOS in the early morning, I arrived to campus at around 2pm. Unlike most of my peers, I did not have a hotel room planned; rather, I had been promised an air mattress by Annie. (Thank you, Annie!) She would be busy all day with Cardinality-writing-the-hunt-things, though, so I had some time to kill before I could drop my bags.

I ate lunch at Saloniki, and spent the afternoon and evening meeting up with people. I played Blue Prince alongside a friend in the Banana Lounge, caught up with people from teammate and Level 1 Spicy at Vester, spent hours talking to members of Providence and LAF in the Le Meridien lobby, and then ran over to the Marriott lobby to meet with some UE and Society of Spirits people before they went to bed. (In fact, I saw many more people than I usually do on average during Mystery Hunt! I am very thankful!)

I say this all to establish the fact that I was pretty tired by the time I made it back to campus. My phone claims I walked about 10km over the course of the day, with all of my luggage on my back, and I believe it. I was very relieved to make it into a place to stay. While I was waiting in the lobby for Annie to show up and badge me in, I spent my time idly reading the copy of The Tech that had been placed on the table in front of me.

Thankfully, my quarters had been carefully sanitized of puzzle spoilers. I went to bed around 1am on the loudest air mattress our species is capable of building, and slept fitfully. I was very excited for, and nervous over, what the morning might bring.

Friday morning

Then followed six successful hours of sleep. Boston was unusually bright and clear then, with very little snow, and it was rather refreshing to walk around despite my tiredness. After my traditional yearly Flour visit, I headed for the Providence HQ and arrived around 11AM. We had been placed in the Grier Room, a large conference-style area in Building 34 that could support up to 120 inhabitants. The space was (and I cannot stress this enough) absolutely excellent. There was plenty of space for every member of our team, and it was very easy to walk around and check on what everyone was doing. Last year we had been split across two classrooms, and struggled greatly. This year we could do full-team meetings whenever we wanted. There were also large amounts of wall space for people to work with markers, multiple TV screens, and a small kitchenette adjoining. It even had its own wi-fi.

I helped set up the space, including taking down a small dividing wall, and then got set up for the hunt. At 11:45 we headed downstairs for…

Kickoff (12pm Friday)

Kickoff was great fun. We were very excited to see what Cardinality had meant by:

Team coordination is especially important this year. Between kickoff and hunt start, there will be a video going through some of the systems particular to this hunt in detail. Roughly though, captains should be aware that teams will need to make decisions about, for example, what to unlock and how to manage teamwide resources. Teams may have many options for what they can do at a given point in hunt, especially early on.

So when we learnt what RP meant, we were ready to handle it. We trekked back up (from the 1st floor to the 4th) to our HQ, buzzing about the theme and structure. Then we connected someone’s computer to the TV and did a big watchthrough of the instructional video together as a team.

The way the system worked this year is that

1. Unlocking puzzles consumed a resource known as Research Points (RP)
2. Gathering RP required completing scavenger hunt-like tasks, which were not puzzles
3. You could choose which puzzles to unlock with your RP, up to a maximum of 8

Providence has a position known as Operator, who calls in answers to all puzzles, submits hints (not that we did so this year) and does all resource consumption. Basically, whenever teamwide state is manipulated, the Operator is the one who does it. It was very easy to add spending RP to the Operator’s tasklist. Based on our experience last year and some quick discussion, we also added a channel called #puzzle-unlock-requests. Anyone on the team could post a discovered (but locked) puzzle that they wanted to work on. The Operator would work down this list to unlock puzzles, relying on their own intuition/opinion/alphabetical order/whatever if nothing was currently in the queue. It was an incredibly efficient way to handle unlocks with so many people.

At about 12:50, our captain turned up with nametags, which were obviously a puzzle. I helped with some initial observations (although we were very, very slow to understand what the puzzle was actually about – more on that later), and nabbed one of the Silksong nametags once we established that there were duplicates.

Friday afternoon (1:05pm - 6pm)

Another reason I don’t tend to write recaps is because my style of puzzlesolving is to always be moving: I buzz about the room, trying to find the path of least resistance, a puzzle where I can best be helpful or make the most progress. I’m not one of those people who finds a feeder and sinks their teeth into it. (I save that energy for the metas.)

So, as with every year, all hell broke loose with puzzle release. I spent a few minutes floundering around the site, trying to understand how everything is structured and what we had available. Unusually there were three avenues for me to help with: research tasks (for earning RP), playing around with an actual video game built into the hunt, or solving regular feeders.

After bouncing around and pecking at a few clues for A Wanderer’s Colorlog, I landed on Replacements. I immediately noticed the connection to the Tech copy I had read just the night before, and with that me and some remote solvers devoured the puzzle rather quickly. Very cool puzzle.

After that, a member of the team returned from HQ with our first physical puzzle: Novelty Store. (A custom Tamagotchi – sorry, TaMITgotchi – toy!) I missed the early work on this, but once I worked out it was in the room with me I immediately went over to watch and help out. I think I was minimally useful: I helped transcribe some of the Morse code (for some reason I was the best at this out of all the people working on it), and worked on the sleeping minigame, but otherwise I mostly took a backseat to observe.

While this was happening, there was a small group of people working on the nametags on the floor nearby. We had made an early incorrect observation that this puzzle would be about altered names for famous people (Angel(a) Martini, M(C->D) Hammer, etc). Unlocking the puzzle, Watchtower, did not help very much in this regard. I would drop by occasionally, to see if people had made any progress… it would be some time until we did so.

A new round, Kitty City, unlocked, so I spent some time helping to discover puzzles in that area and complete the sliding challenges, before jumping on the Aviaria capstone that just unlocked. We solved it breezingly quickly (about 15 minutes), and were then stunned to discover something I truly did not expect: a Dimension!

The Operator and a few others had a quick discussion about the small info tab on the Dimension. It told us that teams interested in coin-like prizes should work on concurrently. We determined that we were interested, and that we had plenty of RP to spend (a solid set of the team had been doing research tasks more or less as we unlocked them), so we decided to open up Land of No Name.

<sidebar: Unlock Structure>

It’s very rare that I am blown away by a structural element of a hunt before 3pm on Friday. The exceptions I can think of are unlocking the Perpendicular Device from MH21 (a huntwide mechanic), and maybe breaking into the Factory in MH23 (which top teams did at around 3pm as well). Neither of these were cool rounds, but rather broader hunt structural elements that could be placed early.

What I mean to say is the Dimension unlock structure this year was genius. Instead of saving all the cool rounds to the end (where fewer teams appreciate them, and top teams are more tired), they can get unlocked in parallel so they’re more spread out. In a year of incredible gimmicks, I think this new unlock system was what impressed me the most.

Land of No Name, in case you don’t know, is a round with a brilliant gimmick. All letters in it are rendered as totally unreadable “scribbles”. Thankfully, as you solve each of the 26 feeder puzzles in the round, you unlock new readable letters, Wheel of Fortune-style. I would love to speak endlessly on how intricately this round is put together – consider the way some puzzles are mostly images / symbolic, to give you a toehold! – but I think Herman Chau’s post about the writing process says it better than I could it.

Approximately nine million people started devouring 🍓🏔️🎮, so while I knew the content I found it difficult to help. I spent about an hour wandering about LoNN puzzles and helping with the ongoing Novelty Store work. At some point we decided that we might need more steps, so a member of our team resorted to pacing around our giant classroom, while working on other puzzles on their phone.

Eventually I was tired of not helping, so I sat down next to Zach Barnett and demanded we work on the same thing for a while. We chose to work on Maps, which was a fun little puzzle, and got it done in about 25 minutes. Then he went off to do something else, and I worked on Capoo’s Diary (fun dataset to skim through), Nine Lives (I was not helpful), Spilled! (ditto), and Triple Threat (very cute!).

Providence is a huge team, and it’s hard to capture everything that happens in the room succintly. In the time that it took the previous paragraph to elapse, the nametag crew realized what they had been missing (“how did we miss that?”), people completed many different research tasks, a small shrine to LeBron is created on the other side of the room, and we unlocked the Hyperbolic Space round.

Friday evening (6pm - 1am)

Somewhere in here I took a break and eat dinner, but I’d be lying if I said I remember exactly when that was.

Me and Barnett switched to working on various Land of No Name puzzles, including Mass Confusion, Like Unlike, Not A QR Code, and probably more. Barnett made a very early observation about the first and last letters, so when we unlocked The Alphabet just after 8:30pm we feel like we have a decent handle on it. While we were all researching Logo and L-systems, I pitched the substitution puzzle that I thought the system would use (and turned out to be right). This meta would surely fall soon!

(Spoiler alert: it would not.)

I helped poke at Not a QR Code and Acquisitions as the other feeders in our two unlocked Dimensions got solved, but it seemed like more software-inclined people had both covered, so I switched over to the recently unlocked metas in the Kingdom. Some maniac (Paolo, I think) had already solved Let Her Cook(book)! with 2 out of 4 feeders, so I had the scent of them and knew they were tractable. I teamed up with Kylee to do The Labyrinth. It was a very fun metapuzzle and an excellent exercise in meta-matching. We also had a decent sense of what was happening in Mixed Messaging, but we weren’t aware yet that we lacked quite a few of the feeders.

I looked at the Atlas of Mosaics round long enough to determine that it was very cool, then jumped on Electrical Circuit, since I knew exactly how it worked. There I got embarrassingly stuck on extraction before another team member (also named Thomas G) bailed me out. I switched to working on Point of Divergence, but the campus closure time loomed large; so the remaining team members and I tidied up the room a bit, and we headed out.

Thomas metaterminal, 11:26pm
I’m waiting for a way to spend RP outside of puzzles (which they brought up).
maybe a casino

Friday night (1am - 9am)

The problem with staying with an MIT student who is also on the writing team is that your schedules don’t line up. Annie said she would be busy until maybe 3am. Theoretically I could go with her back to the dorm at 1, but I was wide awake and wanted to keep working with the night crew.

Thankfully, a teammate bailed me out with a spare bed just south of the river. I texted Annie to let her know that I would be staying somewhere else tonight, and also that I hoped the hunt was going well for them.

Zach Zagorski (yes, this is why I use their last names) had offered up their hotel room as Night HQ, so a small little group of us about 8 strong trekked over to Le Meridien to resume the fight.

<sidebar: The Night>

Providence likes winning, but I would be hard-pressed to say that it likes winning more than it likes sleeping.

In 2025 in particular, we took huge hits in productivity overnight, since we did not have a very codified night crew to keep moving through metas while everyone else rested. By contrast, Cardinality kept pushing on metas, and pushing, and pushing… and on Saturday night they swung the hunt in their favor.

Though the competition was less fierce this year (I’ll talk about that later), it was very demoralizing to stall out in such a big way. Providence made a concerted effort to put together a dedicated night crew who would stay up during campus closure and keep our progress rolling forward.

Have a night crew, kids. It helps.

(See also: Trevor’s more extensive discussion on the same topic.)

Crammed onto Zagorski’s small hotel couch, I watched Point of Divergence get solved, helped round out the solve on I Think That’s Why They Call It That, and then worked on The Quaking Earth with both some in-person night crew and dedicated remote solvers. (We also looked at more puzzles, including Lights Down Mode, but didn’t make a great deal of progress.) The room had a very pleasant energy: much cozier and chiller than the big, somewhat frantic buzz of the whole team.

We all had a nice little laugh about the final Ancestry cluephrase, and then I managed to insta-backsolve A Fake Artist Goes To Eland Islands while alt-tabbing between it and Adding and Abutting. (“Hey, this game is just like [answerline]!” “What did you say? Call it in!”) After the latter was solved, I posted in the Discord:

Thomas metaterminal, 3:52am
Adding and Abutting was our 100th solve!

We had also been mulling over our hunt scope, since at this point it felt we were already a decent way into the hunt and it seemed like we had all the Kingdom puzzles unlocked. If there were only about 180 puzzles (lol) we were decently far ahead of where our progression usually would be, which would suggest that we might finish the hunt by tomorrow evening. Zagorski and I discussed the RP count, and we made the calculation that there would probably be six dimensions:

Zach Z, 3:59am
80 kingdom puzzles + 8 kingdoms to unlock (2 apiece, 16) + 9 capstones (2 apiece, 18) + 10 kingdom metas (2 apiece, 20) = 134, leaving 18 for unlocking dimensions (3 apiece)

It was observed that there might be a Kingdom capstone or other puzzle that could eat RP, but for now this was our guess.

4am came and went, and so the night crew decided to call it, and I caught an Uber across the river to the place I was staying. I strongly appreciated not having to sleep on the air mattress again.

Time passed.

Saturday morning (9am - 1:30pm)

I slept more than I meant to (four hours), and woke up at 9. (A luxury!) I decided to take the 30 minute walk over the Charles River to campus, since the fresh air and exercise would likely wake me up.

Boston is very beautiful, you know.

I greeted the morning crew, who were much perkier than I was, and made myself some tea before settling into puzzles. I was also informed there were hot sauces, which I smelled but did not taste. (Side note: Da Bomb’s whole concept is hilarious.)

After doing exactly one ID for Da Bomb I then switched to helping with metas. Alphabet was still open, after being stared at all night (more on that later). Hyperbolic progress was also advancing. I felt I’d be most helpful in Kingdom, though. I helped wrangle and send out some MIT students for the in-person component of Chart Your Course, and started to make guesses for feeder assignments and layout in Town of Terror. I got to be the transcriber for a teammate observing Shabbat, which was an incredible experience: we were doing Filial Piety Tree, and they were speaking and thinking faster than I could type. With their help we knocked it over very quickly. Incredibly impressive.

It was around now that my partner arrived to campus, so I took a break to show her around the hunt, and also eat food and drink water.

Things that happened while I was taking that break: our last Kingdom meta was solved (Town of Terror), the Kingdom metameta was unlocked and solved, Not a QR Code was solved (we had all the feeders now!), and a physical paper construction of the Hyperbolic Space round was underway.

Saturday afternoon (1:30pm - 6pm)

We got the metameta interaction from Cardinality, who we also interrogated for information. (Or more accurately, they asked us if we had any questions, to which the answer was yes.)

Cardinality: You’ll need to befriend some strong Puzzmon.
Us: How many Puzzmon do you think we’ll need?
Cardinality (doing hand motion): Oh, about six or seven.

Cardinality: We see you’ve been working on The Alphabet and Hyperbolic Space metas. How’s that going?
Us: [internal screaming]

Clearly they thought we’d do fine, and left after we completed the interaction. The hard confirmation that there are (at least) six Dimension rounds is welcome, and there is some speculation that we might be able to finish by midnight Saturday if we keep progress up.

I had been working with the Hyperbolic Space crew, and at this point we were confident enough to begin backsolving, which triggered the meta unlock. Simultaenously, we unlocked Fate’s Thread Casino, which was immediately swarmed by hungry solvers who weren’t working on the vanishingly small number of Kingdom or Atlas feeders.

<sidebar: How Round Unlocks Work>

This is information I asked Cardinality for during our final runaround, though it has since been corroborated by authorial statements online.

The way Dimension unlocks work is that there are three “slots” for rounds to occupy, and you can only unlock another round if you have a free slot to fill. A round takes up a slot if you have it open, but do not have the meta unlocked. The Kingdom also takes up a slot until the final capstone meta is solved. So when we unlocked the Land of No Name meta, we freed up that “slot” to unlock Atlas of Mosaics; and when we unlocked the Hyperbolic Space meta, we freed up that “slot” to unlock Fate’s Thread Casino.

Additionally, Dimensions must be unlocked in order; if you have one pending, you won’t get another until you choose to unlock it.

I think this idea works rather nicely. Previously round unlocks have been gated behind meta solves, but that runs into a problem if the meta is hard and takes some time. It’s a better rule of thumb to unlock new feeders for a team to work on if one of their other sources of feeders has run out. In a round, that usually means unlocking its meta or metameta.

I kept working on Hyperbolic, but switched to the Kingdom capstone meta as soon as it was unlocked, and helped with observations about the relevant source material. (Nice final reveal!) I also think this puzzle counts as a physical puzzle, because the stickers we had been collecting ended up being used, in combination with a map that someone had printed out. This got solved fairly quickly, clearing yet another slot for our fifth dimension, Terminus.

I got a little distracted here, bouncing around with different puzzles, but I remember working on some of the Trends minipuzzles, trying vainly to understand the Terminus round gimmick, and helping a bunch of Brown Puzzle Club members monkey-typewriter the Kingdom capstone interaction (which does not count as a runaround!).

But it’s at about 6pm that the really interesting stuff starts happening.

Saturday night (6pm - 1am)

(teammate), 5:51pm
We know from finishing Kingdom runaround that two other teams did it before us.

The way the Kingdom interaction worked is that each team which completed it was permitted to make an element of a mural, and place it up on a wall. It made intuitive sense that as teams did the interaction, they would add cumulatively to the mural, so that by the end of the weekend the mural would be “complete”.

When members of Providence first arrived, the wall was totally blank. However, Cardinality added two squares to the mural as the interaction went on (before we put up our square). We could take this in two ways:

1. We are the third team to finish this interaction. The two squares are from teams who completed it before us (maybe just before us).
2. We are the first team to finish this interaction, and Cardinality is putting up some squares as an “outline”. (The squares do seem to be less interesting and on the edge of the mural, so this explanation holds some water.)

It’s the first inkling of placement that the team had received all weekend, and either way the result was promising. After thinking about it some more, the team generally settled on the idea that we were probably in first.

But it was nothing compared to the next thing we heard:

(teammate), 5:57pm
We have an opportunity to ask questions of Cardinality
What do we want to ask?

The King of Puzzmon had deigned to allow us to ask a question: to which they could only answer yes, no, and “it’s complicated”.

What did we want to ask?

There was some furious debate. The immediate consensus was that we wanted a hint on Alphabet, but from there opinions diverge. We had been successfully making paths with the long nested strings of letters, and then mostly just staring at them for the last 12 to 20 hours. Simply put, we were in extraction hell. We discussed whether we wanted to confirm our strings of letters (“no, their lengths are confirmed by the puzzle, we know that they are correct”), whether the paths linked together or always started from the same place, whether instead they started from each of the mysterious letters already on the grid, whether we treated each letter in the string as an instruction, or instead each answer in the string as an instruction…

(different teammate), 5:59pm
they said they’re getting impatient
we have undefined number of questions. probably more like time limit
too late
they said it’s time for us to go

Well, that was a fun two minutes.

<sidebar: The Runaround Questions>

Did anyone actually use these questions for something useful? LAF mostly burned them on silly lore, and I haven’t seen anyone else say on the record that they asked a question that helped them solve a puzzle. Surely someone must have.

Maybe not. It seems weird to me that hints would be offered, on a very tight time limit, to a group of people who (by definition) aren’t in HQ and haven’t looked at puzzle progress for a bit.

The Terminus puzzles continued to fall, and I spent my time mostly alternating between the Alphabet and Hyperbolic Space meta groups, trying to push both towards a solution. I was pleased that my suggestion of how the geometry would work out (with the creation of the cubes in the hyperbolic polyhedra) was bearing fruit, and we were extracting letters slowly but steadily.

But soon after, at 7, the team gathered together to see a short presentation on our Alphabet progress. We had learned the hard way from last year that, if a meta was stuck, we desperately needed as many people on the team as possible to look at it; the fresh eyes often came with suggestions that got us moving again. (It has been suggested that our inability to make metas accessible to fresh solvers contributed to our silver medal last year.) So the microphone in the classroom was booted up, and members of our Alphabet team walked laboriously through everything we had tried. The rest of us ate burritos.

A few minutes after the presentation had concluded, and everyone dispersed to their own puzzles, a great cheer went up in the room. What happened – had we solved Alphabet?

No, no. We had solved Hyperbolic Space. (Hooray!) Keep working on Alphabet.

Hunt Length

Hunt is long. Some people think hunt is too long. Other people disagree.

What do I think? I think that, as someone on the winning team this year, I would not have been upset if we had finished on Saturday night. (I think cutting The Glitch round, and nerfing Alphabet/Hexagon, would have brought us comfortably into that realm.) Otherwise, finishing earlier on Sunday morning would have been pleasant.

The problem is “when the top team finishes” is an entirely deceptive metric. We finished this year with two gentle nudges on metas. Last year, Cardinality won before noon Sunday… and then no other team finished until late Sunday evening (us), and nobody outside those top two teams would have finished at all if not for strong hinting from HQ. 2023 and 2024 both required significant amount of hinting and free answers for any team to finish.

I have a whole separate post about whether hunt length is increasing. Ignoring 2023 and 2024 as 2013-like outliers, I think that yes, hunt is getting longer.

But top teams are also getting stronger to match. (Historically top teams might also be getting weaker, due to member loss or greater sleep requirements, but I don’t have strong evidence of this.) This is why Dan Katz and Eric Berlin both said hunt was too long this year, despite the fact that someone found the coin earlier than in the most recent hunts they ran (2019 and 2022). If they’re on a team that is not matching the increase, the hunt feels longer for them.

As Alex Irpan said, more pithily than I:

The 18 year old crack solvers who’ve spent formative years reading devjoe’s archive are now finding each other on Discord and shitposting their way to victory. I can only imagine how weird this must be for people less online.

source: MIT Mystery Hunt 2026 by Alex Irpan

As someone in that generation, I can confirm.

Do I think this is a bad thing? …it’s hard to say. I think hunt could stand to be a little shorter. I also think a Sunday afternoon finish is well within the historical baseline.

Soon after we won, I polled the Providence team on what their preferred “coin is found” time was. The result was almost evenly split between Saturday night and Sunday morning. Hunt length is also notoriously difficult to predict. When the entire writing team changes out every year, and the competitive scene evolves so quickly, how are we to know?

The Competitive Scene

Manifold is a “prediction market”, a kind of web app where people wager fake-currency on the outcomes of real events. Leaving aside my personal feelings on these companies (and Manifold seems less predatory than most, given that you cannot earn actual money from the results), the prediction market for Mystery Hunt winners is an excellent way to quickly poll sentiment on what teams have a shot. Last year, the odds of both Cardinality and Providence winning were about 40%. (Manifold seems to have a recurrent issue where extremely likely outcomes are underestimated, and extremely unlikely outcomes are overestimated; I think the odds of either Cardinality or Providence being the winners was 90-95%, not 80%.) This year, Providence hovered at about 50%, with an oddly selected number of other teams taking up the remaining slice.

To be completely honest, though: there was absolutely no way that Providence was not winning this year.

This is not because of competitive advantage or anything like that. We did end up being the strongest team by a narrow margin. But my point is that, even if we had not been, we were not in danger of not winning, because to the best of my knowledge no other team willing to win was on track for a Sunday finish without significant hinting.

If Galactic had decided they wanted to win, and gunned for it, they could have beaten us. I think the Manifold market is correct in considering that as a potential outcome. But given how the sentiments of all the top teams actually lay, even if Providence was running 5 hours behind where we were, I still think that we would have been given the win.

Perhaps I’m wrong, and in the absence of competition another team would have decided to pick up the mantle. But Mystery Hunt this year was not a competitive event.

Is this a bad thing? I don’t think so. It is certainly not be the first time this has happened in the event’s long history, and given the reward incentives for first place, I think this occurrence happening once every so often is natural and inevitable. Some competition is healthy. But even if Mystery Hunt becomes a series of carefully-coordinated handoffs, between effectively pre-agreed running teams, is that such a bad thing?

Maybe it is. This seems like something that other people will have stronger opinions on.

LLMs in Mystery Hunt

Now

Reality has temporarily returned to normal. We are just starting to handle team logistics. I’ve returned to Brown’s campus, for my final semester of study, and doing so feels… weird. The commuter rail over, which I caught with other soon-graduating puzzle friends, felt psychopompic: shuttling us from this joyous, welcoming, gamified space, back into the visceral strictures of reality.

I wrote this post mostly out of order, so right now, it’s the first day of class. I’m typing this paragraph up as I wait for one to begin. And to get here, I walked from Friedman to the engineering building… past the statue of Marcus Aurelius and the Soldier’s Arch, past the library, past the reporter presenting on-camera and the flowers heaped on the snow. When I walked into the building, I discovered that the vast majority of it had been walled up.

People have jokingly offered their condolences for winning Mystery Hunt. This is a fairly old joke: a tongue-in-cheek acknowledgement of the incredibly large amount of time that writing takes, an effort which is often thankless. I think people who have not yet written do truly, genuinely underestimate the complexity and staggeringly large amount of work that an in-person puzzlehunt – even one on a smaller scale, like Brown Puzzlehunt – requires. There are entire teams of people that work on things that have nothing to do with puzzles, or art, or the website.

But I am getting tired of people saying this. Yes, Mystery Hunt is a staggeringly large endeavor, and one that should not be embarked upon lightly. It is also a thing of gargantuan beauty, a design of human craft and care with the express purpose of bringing others joy. Puzzlehunts have brought me such joy, such companionship, that it feels wrong to pretend that this opportunity to write is anything other than a gift.

Save your condolences.

Maybe the previous generation of people don’t want to write for Mystery Hunt anymore. Galactic, teammate and Level 1 Spicy, TTINNL, D+M; maybe they all feel their day is through. I shan’t begrudge them that at all. There is no obligation, no necessity for people to write, especially after they have written so much.

But I don’t feel the same way. I am excited to write this year. And I hope there is a whole new generation of people behind me excited too, who shall step forward and advance the culture of Mystery Hunt in new directions; and perhaps we shall be the vanguard of a new wave of teams, new talent. I hope so.

A few years ago I told a friend that I write fiction, create puzzlehunts, and design experiences because I feel compelled. It’s not for fame or fortune. It’s certainly not because it’s easy. But I have some drive, some itch to make things, and the only thing that tames the voice is to produce. Perhaps I am alone in feeling this way. Speaking only for myself, though, I think I shall keep making things until I die.

Because we are, for now, humans. And while I am human I would like to do something that only we can do: make art.