mystery hunt recap, 2026

maybe it’s my time to try and write a hunt recap post. i blog! i do puzzles! i’m allowed to blog about puzzles! and i’m on a team that people sometimes recognize, and somehow we’ve become a team that’s a little bit good at hunts? so maybe my written reflections will be interesting to others, and not just to me :P

here’s some thoughts on hunting this year — as always, i hunt with NES, a team that largely (but not exclusively) consists of alumni from Next House, my undergrad dorm at MIT.

obviously, spoilers for mitmh 2026 (a wonderful hunt written by cardinality!). and the usual disclaimer that all opinions on here are my own, not that of anyone else on NES or in the NES leadership team, are all made with loving intent, and so on 🧡

(not a required pre-read, but i’ve mostly been in the same mindset as when i wrote taking your medicine — you might find it interesting to read that first. or, if you don’t know much about mystery hunt, it’s also useful for a little bit of context.)

puzzles

first, a talk-through of the puzzles that i had some form of significant interaction with, which i think shows what my hunt was like in (very) broad strokes.

this puzzle has been here the whole time (bubble cove, feeder)

hooooly cow what an incredible puzzle to see to start the hunt. i immediately head to it with some other dropout nerds. had a small miststep to start (totally thought the first diagram was referencing lisa gilroy as ellen degeneres in jaws), but we quickly got past that. the sam reich cameo was absolutely incredible to boot. in hindsight, could have solved it sooner, but so it goes with puzzles. deeply enjoyed and also immediately forwarded to all the people i know who are into dropout¹ (unlocked fri 1:19pm, solved fri 2:16pm.)

architecture of flow (old bark town, feeder)

i’m no hip-hop nerd, so most IDing was not done by me. had some fun going down the alpha aerobics pathway — it sticks out in my head purely for the daniel radcliffe version. (unlocked fri 2:20pm solved fri 2:52pm.)

welcome to mit (aviaria, feeder)

helped out with this puzzle after finishing architecture of flow. most of the hard work had already been done, and i mostly helped out with a bit of error identification to get to the final answer (incorrect submissions include: deserters, desecrate, desperate, …). it really is fun when my knowledge of campus really comes in handy: love being able to look at a photo and tell you exactly which crosswalk it was take from. (unlocked fri 1:06pm, solved 3:14pm.)

jumping to conclusions (kitty city, feeder)

a joy of a puzzle. i love squinting for answers, and a puzzle that acts like a minimeta is so fun. i spend most of my time finding the right goosebumps novels and nutrimatic-ing various other components, but our team churns through this (unlocked fri 3:26pm, solve 4:08pm).

maps (kitty city, feeder)

ah, the first of two consecutive puzzles i work on where we end up hard stuck. we unlock this at 4:20pm, and i head over to do some geoguessr investigation. we’re not sure what to do with emojis, and mark it as “stuck” via nesbot at 5:16pm. it takes until 10am on saturday for someone to open the sheet and read the first letters — the puzzle is solved 30 minutes later. sometimes, you just have to remember to do the simple things.

squint your ears (kitty city, feeder)

one does feel like they’re going insane listening to things that almost sound like english. we get this at 5:23pm fri; and we mark it stuck after running state of the solve at 7:07pm. i clean up our spreadsheet and fix some errors in the early hours of saturday morning, and try to submit a hint request when they open but get promptly rejected for being “on track to finish hunt” (a cause for celebration, of course). we only backsolve it at 5:55pm saturday.

looking at the solutions, this puzzle’s a good reminder for me to just write down all the information when you’re feeling stuck. i thought about using the “time of static”, but two of them seemed like they were at the exact same time so i didn’t bother. aaaa.

acquisitions (hyperbolic space, feeder)

you know how frequently something that i do for work comes up in a hunt? never. literally never. so getting a puzzle related to auction theory was wild. i contribute various things (all-pay auction, IDing the common auctions), but we get stuck on finding the bidding strategies. some teammates get the “presidential” and “chemist” ahas after I head to sleep, and bash it out not too long after. (unlocked 7:01pm, solved 10:25pm.)²

shapes (elder drifts, feeder)

the SQUARE hole. a very fun gimmick i quite enjoyed. (unlock 8:03pm fri, solve 9:03pm). i’m far too tired to be productive after this point (it’s 2am in the uk and i’m falling asleep at my computer) and i get shooed off to bed; unfortunately, this means i didn’t get to do musical numbers (but i’m excited to post-solve it soon! <3 to my a cappella frendos)

gerrymandering (eland islands, feeder)

my first new puzzle of saturday morning. a fairly straightforward solve, and getting the logic of modes/means to work was quite fun. (unlock 9:31am, solved 10:19am). a very cute little gimmick. (it’s around here that the teammates finish off maps, and i am so relieved to be free of it.)

mixed messaging (kingdom of the puzzmon, meta)

the first meta that i help out with! (i didn’t end up helping out with any capstones, and a good chunk of my time ended up on the organizing side of NES). we unlocked it at sat 4:05am, and someone had the core aha at around 4:30am — the existence of these pairs of words is actually just incredible. i head to this puzzle after finishing up with gerrymandering, and with the help of google sheets formulas and a bunch of code, we end up solving the meta at sat 11:01am with just 6/12 answers. (this leads to many failed backsolve answers to other puzzles.)

monster mash (kingdom of the puzzmon, meta)

at sat 10:58am, someone has the core aha for this meta with CHA[KRA KE(-N)]NNEL, and i race over with some other people as soon as mixed messaging finishes. with some educated guesses about the pictures on the puzzle page, guessing some phrases that lead to plausible answers, we manage to forward solve this with just 2/9 feeders at 11:09am. throughout the weekend we get a few backsolves from this (we correctly IDed 3 more answers to get the solve), but it was a little disheartening to solve a puzzle and have it go to a meta that’s long been completed. i take a break for lunch, and the early afternoon is spent keeping track of backsolve guesses and organizing other things.

town of terror (kingdom of the puzzmon, meta)

after several false starts on this (gloucester sea monster, town of salem, …) we hit the main aha. i start chipping in around 5:32pm sat, and 15 minutes later head over to my ipad to try and draw out the paths. some others on the team get it solved at 6:24pm sat. mostly, a puzzle that juts reminds me of the deep cuts that exist in hunts.

dance of the bumblebee (atlas, feeder)

buzz buzz.

the round gets unlocked at 12:37 fri, and i start working on this at 7:20pm sat, hanging out in vc with friends and trying to decipher what the bees are trying to say to us. we slowly churn our way through different mini-puzzles, sharing our dictionaries back and forth, and i head to bed at about 9:30pm. overnight, there’s a good amount of progress, including getting to the “meta” at 11:26pm, but some incorrect words in our dictionary throw us off. in the morning, we come back and try to clean up dictionaries, but to no avail; we grab a hint from HQ at 10:54am sun. we try to clean things up a bit from the galadriel/frodo conversation, decide it has to be a (13 7) enumeration from reading order. honestly, i’ll just share some messages:

ok ok so british flying vehicle something 13 7
aaaaaa
-me, 11:53am

i thought it was the world name not the vehicle name
so that would make sense
-person a, 12:02pm

oh. hm
wait liek the world with a british empire and a spaceship ?
or like . what world
-me, 12:03pm

the (13,7) modifies world
not plane
-person a, 12:03pm

mmm where do you think plane fits in? like you take a plane from UK to this world/location w 13,7?
-person b, 12:07pm

what if “text ask you” is questions?
“british countries which write questions” or smth
-person a, 12:07pm

??? international
international as a 13 ?
-me, 12:08pm

yea i thought of it but unsure
-person a, 12:09pm

(raucous celebration keymashes, followed quickly by /solved [redacted answer])

i think the only other conlang puzzle i’d done before was ted tales from ABCDE:FG; i’d never really enjoyed the puflantu puzzles before in gphs — just not my cup of tea — but i just kept getting sucked into this. loved the slow grammar/vocab lessons becoming increasingly complicated to help us learn organically.

filial pie tree (royal groves, capstone)

we unlock this puzzle at 2:02pm sat. i hop over here at 9:34pm; in the meantime, people have had a few ahas, connected it to PIE. my big contribution is going far enough down wiktionary rabbit holes until i found roots that made sense, and helping to line things up to extract. love a good etymology puzzle! solved at 10:08pm.

connect the clans (kingdom of the puzzmon, metameta)

i’ve been keeping the little “capstone stickers” in my pocket this whole time, and when we unlock this because we solved filial pie tree, i rush over. our team quickly has the right series of ahas: games of thrones, making the cycle of animals, and then putting them onto the map. a lot of sheets squinting later (and after failing to place north facing north), we solve this at 10:56pm. not too bad for a metameta :) solving this unlocks terminus, but i get shooed off to bed before i can take a good close look. our runaround for this gets scheduled for 9:30am sunday, which provides a fun little break for those who were up early :)

mixed media (terminus, meta)

i wake up on sunday at ~6am to some huge progress overnight: terminus feeders largely solved, the terminus meta open, atlas with multiple metas solved (and multiple more to do), fate’s thread casino + the glitch unlocked, and another day of puzzling ahead. the hard work for the mixed media has already been done (i.e., the conversions back and forth between answer forms), and i get some updates from the crew who’s still up (somehow) at 6:28am on what they’re stuck on along with the one other person who’s woken up this early. we pursue the Grammy route a bit more, and find the right songs, but don’t know what to do. at 7:31am, i make the connection to lyrics, and the puzzle is solved 4 minutes later. what a good feeling to solve a late-game meta first thing in the morning.

playlist (atlas of mosaics, meta)

i spend some early-morning time (9–10am sun) trying to identify the missing notes; definitely a lot worse at this than i used to be, but i take a bit of pride when the people with perfect pitch (mostly) verify my notes after they awaken :) i’d like to think my minor contribution helped a tad, and we solve this at 10:52am sun.

point of view (glitch, feeder)

one of the handful of puzzles i looked at sunday afternoon/evening. i think we were a bit annoyed on this, as we kept trying to extract from the small grid thinking it must be relevant, too. we kept thinking in circles until we got a hint that nudged us away from that, and only using the zoomed in grid. sigh. (unlocked sun 2:36pm, solved 5:03pm)

snalc eht tcennoc (glitch, meta)

we unlock this at 5:14pm once we solve these questions are not clever, and our whole team piles onto it. a fun callback to the earlier metameta, and pretty quickly we see what we need to do. i head back to our map with some teammates, while others start applying the appropriate glitch transformations. it takes us only until 5:40pm to submit the partial answer, and 11 minutes later, we’ve solved Glitchy’s puzzle and the hunt was done. our second (and fastest) finish to date, putting us in forth!

some reflections on puzzles:

• writing this out, it’s weird to see both how few and how many puzzles i got to meaningfully contribute to (19, per the above count). on the one hand, there were 232 puzzles in this year’s hunt, which means there were 213 puzzles i didn’t actually do anything for. (sure, i clicked into almost all of them, but i contributed very little to most.) on the other hand, i had about 38 waking hours during hunt, which means on average, puzzles i worked on were solved in 2 hours — and that’s not even including the time i spent taking naps, planning out strategy things, taking breaks, etc. and eight of them were metas/capstones! but. as always, lots of hunt that i don’t get to see: i’ve still never taken the time to do an event. maybe one day…
• overall, lots of really fun ahas to be had as always. the joy of being like “oh it’s this!” is so good. now that i’m doing a bit more puzzle writing, i find myself being a little more critical of puzzles (in the “i have some critiques to give” way) in a way i don’t think i have been before. (to be clear, there are not very many of these, and for the most part they’re pretty minor.)

nes

i started doing puzzlehunts in 2019 with N3XT (a group of folks from 3W/3E); we were excited to solve 30 puzzles and a meta. the next year, we merged with another next house team (TWTW from 2W/3W) to form a next house superteam™ named NES that managed to solve a whopping 64 puzzles. we joined forces with 2MysterE (2E) in 2021 amidst the pandemic to become a supersuperteam™, and to our complete surprise, we solved 174 puzzles and placed 15th. 2022 was 18th, 2023 10th, 2024 6th/8th³, 2025 we finished and got “4th”,⁴ and 2026 4th for real. even though i was only a team member during the early years, it’s been wild to see how far we’ve come. i’m still somewhat shocked when other people recognize the name of our team. it’s weird hearing other people say that they’re rooting for us to win over the next several years. in my head, we’re still just a group of friends that gets together for puzzles during IAP, but of course, we’re now 70 people who can finish whole mystery hunts.

in 2023, i joined the “sheriffs” of nes, and started helping out with various organizational things. sign-up forms, trying to improve puzzle documentation and communication and team coordination, and, perhaps most saliently for team members, being one of the two strategy captains. this means i help run our State Of The Solves,⁵ try and keep apprised of how our hunt progress is going at a high level, and doing things to vaguely move our solving-power towards places where it’s most needed. i don’t think any single one of us is in “charge” of nes, but i think during hunt, me + katie (our other strategy captain) end up making many of these high-level decisions in the moment.

this is only the second year that NES has had strategy captains, and there’s certainly a lot that we’re still figuring out how to do. some reflections from this year:

• i think we did a pretty good job at managing the unlock queue this year. my opinion here is largely because the puzzle width⁶ was less of a constraint this year, and so we had to make fewer “hard” decisions on what to unlock and where. we started hunt with a width (“puzzmon capacity”) of 8, a width that later increased to 10 (and then 12, and then 15). more importantly, unlocking Dimension rounds later in the hunt didn’t count towards that width, effectively doubling our puzzle width (or more). i also don’t think we ever got to a point where it felt like there were too many puzzles open — maybe i should do some stats on our width throughout the hunt.

  this stands in contrast to last year, where i feel like we felt much more constrained by the number of puzzles that we had unlocked at all times, and so it made for a very different hunt experience (and lots of people wanting something unlocked at all times).

• a good amount of my time was spent trying to keep track of the state of our progress. at least 30 minutes around every dinner was spent prepping/doing state of the solve. and i think almost all of sunday was spent coordinating hints, trying to see where we needed more solves, and guiding people towards those places.

  to be clear, i do enjoy the tradeoff — it’s nice to be able to put on an organizing hat, to think through team strategies, and also just to check in on people and try and help them enjoy hunt. but just noting to myself that objectively, i spend way less time during mystery hunt actually doing puzzles than many other people on my team.⁷

• making decisions once you are more tired is always very hard. as hunt goes on, and you’ve been staring at computers for 30 hours, stuck on incredibly hard puzzles, one isn’t always in the best state of mind for things. i don’t think that we did anything wrong, per se, but i think there are certainly ways in which i could have said things better, done things with more tact, etc etc. this especially is relevant to…

• figuring out the right way to encourage people to use hints.⁸ i sent the below paragraphs into #strategy-discussion before the hunt talking about our philosophy towards hints and free-solves:

  For hints and free-solves: we’ll be keeping the same strategy that we’ve used in previous years. In essence, this strategy is one of “let’s try our best to help people enjoy puzzling”. We’ll be using the hint system as it opens (and will have a channel for managing hints,⁹ since we usually can only have one submitted at a time). For free-solves, our rule of thumb has been that if someone’s excited/actively working on a puzzle, we avoid free-solving as much as possible. We try to have dialogue between meta captains / people working on puzzle / strategy captains to see whether this is one that would help our progress if we indeed could free-solve, and always look for an “ok” from the relevant solvers that they’re ok with their puzzle being free-solved.

  There’s more “aggressive” versions of each of these strategies. The most direct one (that I’ve heard some other teams use) is that their strategy captain/meta captain equivalents will free-solve puzzles once they decide that they need it / progress would be too slow on it, even if there might be some excitement about it. That “do we need it” can also be done pretty aggressively, in a way that we haven’t done before. (Hints wise, I think it’s a little hard to be more “winning-oriented” than what we’re doing now, besides lowering the level of “stuck” where people ask hints. but also like . that’s no fun)

  of course, free-solves didn’t end up being relevant this hunt, but hints very much were. our team didn’t unlock the ability to hint until around 7pm on saturday, and figuring out the right ways to use hints on sunday was… difficult. in a sense, there’s four balancing priorities:

  1. people want to finish the hunt,
  2. people want to have the joy of figuring out a puzzle,
  3. people want to be done with a puzzle where they feel completely stuck and where nothing’s moving, and
  4. people don’t want to sit around doing nothing.

  objectively, we didn’t hint very much: 9 hints requested (with 2 follow ups for clarification),¹⁰ and one in-person nudge on the alphabet meta. but subjectively, it did feel like we ended up being fairly aggressive in terms of encouraging people to hint, particularly on sunday. (i’m especially thinking about 2 particular feeders in the glitch round that probably could have been done sans hint with another few hours of puzzle time.)

  our (as in, me + katie, NES’ other strategy captain) in-the-moment decisions boiled down to:

  1. our width was quite small (3 puzzles), meaning lots of people on very few puzzles,

  2. the glitch meta being an unknown difficulty, and us needing more puzzles to unlock it (at 5/9 solved we thought we’d need 1 more, but then we got to 6/9 and still needed another),

  3. not knowing if a metameta came after the glitch meta, and

  4. cardinality encouraging us to be aggressive with hints.

  and so on balance, items 1, 3, and 4 made us decide that hints were worth it, and led to a good amount of encouraging to hint after it seemed like people weren’t making progress, but still leaving it up to the puzzle-doers to actually write up/submit a hint.

  in hindsight, we likely could have finished the hunt without hints on those two puzzles, and so the fact that we encouraged hinting when those hints might otherwise not have been taken doesn’t feel great. i do think that, on net, taking those hints was a right decision in that moment. not necessarily the only one, of course. but so goes trying to make any choice.

• where does NES go from here? like i said above — i’ve heard multiple people (not on NES) say that they’re rooting for NES to win hunt sometime in the next several years. we’re somewhere near the top (if not first)¹¹ on a ranking that’s a combination of {recent mit graduates/low in age, does well in hunts}. about half of the people that filled out our census poll were a class of 2020/2021/2022.

  for the first time ever, our sign-up form had a survey to collect people’s thoughts on winning. by and large, people don’t feel ready yet, but are excited to do the things that would help them feel ready (i.e., writing more puzzles) — a few years ago, we started writing a NES hunt, and though momentum sputtered, we’re planning to get back on that this year with some internal potlucks and whatnot.

  there seems to be a decent amount of excitement to someday actually win. if all goes well, of starting to actively try and win come 2028+. of course, this is all thinking years in advance, and we’ll see how trying to write potlucks and hunts goes, and people would have to want to do things when we get there, and we’d have to actually win (certainly no guarantees on that), but there’s are worlds where we win in the next 5–10 years that aren’t very farfetched.

  it’s a bit weird to think about it all. people make such a big deal about winning hunt, and writing the next one, in a way that feels a bit unsustainable as a practice. (i’ve heard some stories of people that took a year sabbatical from their jobs just to write hunt, though i doubt that’s something i’ll be able to do.) sure, it’s incredibly grand in scale and there are thousands and thousands of people hours needed to write and test hundreds of puzzles. but also, between the ESP and theater backgrounds of our leadership team, i feel like we’d be able to figure it out.

as always, shoutouts to all the sheriffs (Katie, Mary, Michael, Mihir, Mindren, Zach) for being such great people to run NES with, and for putting so much into this team (non-exhaustive list of things done this year: swag, NESbot, thinking about running hunts/helping people practice writing puzzles, hauling snacks around, discussions on strategies and team goals, having a crate of NES items year-round in their house, coordinating subteams on NES to help with other things, and making NES a lovely place).

here’s to more NES. no end (in) sight :)

hunt overall

various other observations on this year’s hunt.

• so many cool physical objects. puzzmon the card game was incredible, both as a game and in terms of production value. so was the faux-tamagotchi and the hot sauce puzzle (while not for me, i’m sure it’s somebody’s idea of fun)

• the scavenger hunt giving RP which was used to unlock puzzles was a great way of making scavenger tasks feel more organic and necessary, rather than something that you only use once. they certainly fell off in terms of usefulness as time went on, but i think that is ok

• as someone who didn’t go to events, the concept of having many smaller events throughout is super interesting. it’s certainly more of an onus for smaller teams, but it let a lot more people engage in events than in most hunt years. good or bad, depending on whatever a (hunting) team’s goal is (and i do think it is a net good for the overall experience of IRL teams)

• i quite enjoyed the gimmicks of various dimension rounds, particularly land of no name (though i didn’t solve anything there) and atlas. do hope to go back to the latter and check out rounds i didn’t do

• i’m realizing that i enjoy hunt stories more when they have lots of real-life interaction. i understand why the choices get made to not do this, and that hunts without a lot of real-life interaction have their own enjoyable things, but i think i have enough mystery hunts under my belt to feel more confident in the existence of this opinion

various miscellany

• “PULL PULL PULL PULL PULL PULL PULL”

• i wore a nessie costume for almost all of hunt. it will make a return. that is all

• apparently, flame-grilled steak chips are not so bad. jalapeno candy canes are, though

• “i assume you are-“
  “yeah i’m one of the she-“
  “capable of opening that bag”

• how upsetting to see so many people wrong

very excited for next year’s providence hunt! i don’t quite know how next year’s hunt will be for me, though — i’ll unfortunately be on the econ job market, and interview/flyout season is right around January, so i might need to take a step back out of necessity. but i hope that if i do, it won’t be a big step, because after all, hunt is fun.

see you in 2027!

1. which is. a lot. look, i’m in boston, in my 20s, and am queer? yeah absolutely i know people who subscribe to dropout. it’s stranger when people aren’t subscribed ↩

2. errata: previously thought we didn’t get the ahas, but we did! wahoo ↩

3. depending on what metric you used. 8th by number of solves at end of hunt, but we were the first team to post-solve the hunt and get a virtual runaround :) ↩

4. well, 4th-ish — d+m stopped officially ranking people after 2nd, due to the volume of hints they gave out beyond that point ↩

5. our version of team-wide announcements to keep everyone apprised of events ↩

6. number of puzzles open at a given time ↩

7. something something comparative advantages? but also, i do actually just enjoy getting to be in this role. support characters ftw. ↩

8. or free-solves, though this wasn’t applicable this year. ↩

9. this did not end up happening because of only unlocking hints so late . oop ↩

10. per the cardinality AMA, among the 15 teams that finished hunt, the median number of hints was 10, and the floor was 0 (i’m under the impression that the top 2 teams used 0, but perhaps none other?) ↩

11. though wahoo!, a team formed this year by some undergrad puzzle club members, hopefully gives a chance for some younger ones to take the reins — they did very well and i’m very much rooting for them ↩