Metapuzzles, Backsolving, and Short-Circuiting: A Study of Three Puzzlehunts

(This post will include limited commentary, and potentially spoilers, on three puzzlehunts from earlier this year: the MIT Mystery Hunt, which can be found here), the Galactic Puzzle Hunt, which can be found here, and the Cryptex Hunt, which can be found here.)

Hi all! Since we last spoke, Mystik Spiral was (were?) accepted to Miskatonic University, and my team is very excited. You can see our application video (and lots of others) on the Miskatonic site. In addition to an obvious bias toward ours, I really enjoyed the videos from The AMI-Gos, Boneless Chickthyologists, Friday the 13th Part VI, The Gray Old Ones, and Innsmouth High School Scuba Squad. I have friends on all those teams, which probably made them extra-amusing.

I’ve promised to discuss backsolving policies in the Mystery Hunt, which was going to be Part 6 of my Mystery Hunt recap, but then I participated in two other puzzlehunts that dealt with backsolving in very different ways, and I thought it might be useful to compare and contrast. Of course, that means I’ve been assembling an epic post in my head for weeks, which makes it harder to get things on the page. Let’s see if it lives up to my personal hype.

First of all, let’s get some terminology straight, since I know some people think of “backsolving” as solving any puzzle with partial information, but I think it’s more specific than that. I’ll define backsolving as the practice of solving a metapuzzle (or at the very least, figuring out how the metapuzzle works) and using that information to confirm a correct answer to a feeder puzzle without solving that puzzle. I’ll define short-circuiting as the practice of solving a metapuzzle and understanding how it works, but managing to do so while missing a significant number of feeder answers. (Note that short-circuiting often leads to subsequent backsolving.) Short-circuiting sometimes results from wheel-of-fortuning, a common puzzlehunting strategy where the solver figures out that a puzzle or metapuzzle is generating a series of letters, and the solver manages to guess a lot of missing letters by pattern-matching.

Okay? Okay.


What elements of this puzzlehunt encouraged backsolving?

In the 2017 Mystery Hunt, solving metas was a very powerful thing. We intended quest metas to be a mechanic through which power teams kept themselves from being bottlenecked, but if you solved quest metas quickly (not to mention if you solved character metas quickly and used them to help you solve other character puzzles) you could cut through the Hunt like a hot knife through butter.

There were multiple exploitable features of that structure, but one was that strong teams solve metas faster, so if solving metas opens puzzles faster, the rich get richer. This was one of the motivations behind a structure that based unlocks exclusively on puzzle solving rather than on metapuzzle solving. The problem with this is that if you solve a metapuzzle and don’t get a bonus for it, it massively incentivizes backsolving, as we discovered when many many teams (starting with Left Out, who short-circuited the first meta with ten missing answers, which is insane, by the way) started calling the remaining answers in for everything in sight.

What, if anything, did this puzzlehunt do to discourage backsolving?

We didn’t have any explicit policy (certainly not communicated to teams at large) about backsolving. When we wrote the Hunt, we thought the “mingled metas” mechanism would effectively reduce backsolving, because teams wouldn’t know where to put their orphan answers. But a lot of us on Setec feel that calling in an answer twice is totally part of Hunt, and calling it in nine times is tres obnoxious. A lot of teams don’t feel that way, and in their defense, if the goal is to win and you have no way to know other teams aren’t doing that, it’s hard to argue that you should place artificial constraints on yourself.

In practice during Hunt, when we felt a team was attempting backsolves to an abusive extent, we sometimes put them in the “penalty box” by giving them a stern phone call and then not answering their calls for a while. We’ve done this in past years and it usually only affected a few teams… but this year we hit the trigger too quickly and then felt compelled to treat other teams the same way out of fairness. Yet based on comments on an earlier post, we didn’t succeed in consistency here. And there’s a question of whether penalizing people for doing something we didn’t tell them to do is fair. I think there were balls dropped here.

What, if anything, did this puzzlehunt do to discourage short-circuiting?

We thought the mingled metas would make it tougher to solve metas in general, particularly in that you might not know how many feeder answers to expect. We also ended up with more metas than usual where even when you knew how the meta worked, you needed most or all of the answers to carry the process out. This was not intentional on my part as meta captain, but I suspect some of our authors had the 2017 Cleric round in the backs of their minds.

What (in your blog host’s opinion) should be changed?

Backsolving is a part of Mystery Hunt, and I don’t think that should stop. How much backsolving is appropriate is an incredibly controversial question, and I encourage you to voice your opinions in the comments. I’ve already said some of what I think in the “two backsolves vs. nine backsolves” example above. But there’s a whole lot of gray area between 2 and 9… where’s the line? And how do you set up a policy that limits abusive backsolving, but doesn’t penalize teams (especially casual teams) if a puzzle is legitimately challenging for them and they need to try multiple answers?

I would recommend, at the very least, two things to future Mystery Hunt constructors. First, if you’re going to have a backsolving limit policy, make it clear to solvers as part of the Hunt instructions. Don’t make them guess, and don’t assume their sense of etiquette matches yours. Second, there are two things Setec pretty clearly agrees should be prevented by the Hunt software; it should not allow teams to have two pending answers submitted for the same puzzle, and it should not allow teams to have the same answer submitted for two different puzzles. Should teams be able to spam the answer confirmer? Maybe, maybe not. But I maintain that submitting two things that can’t both be right, and cutting in front of everyone else before you find out if either of them is wrong, is a jerk move. Our tech team realized after Hunt started that this would probably have been easy to implement. We just didn’t think of it, because we didn’t realize that SOME OF YOU ARE MONSTERS.


What elements of this puzzlehunt encouraged backsolving?

To start with, the fact that you had to solve every single puzzle. I know this because… well, let’s back up.

The hunt was run through ClueKeeper and built around a collection of puzzles embedded in a fake magazine. Incidentally, I really liked this combination; in a puzzle set like this, one of the potential annoyances is not knowing if this page is one puzzle or two; having a separate interface that listed which magazine elements had corresponding puzzle answers (and their enumerations) felt like it made the playing field more fair.

Last year, the Cryptex Hunt (eventually) got a couple of full posts from me in this blog… that was back when I had small amounts of free time. So this year, instead of a full post about how I failed to win Cryptex Hunt, you get a partial post about how I (co-)WON Cryptex Hunt! Wooooo! This was not something I expected when I fell asleep. I solved a bunch of puzzles right away when the hunt began on a Friday evening, but I started to stall out, and the leaderboard showed a big lead for [flips my Jangler/Projectyl two-sided die] Projectyl, with several other solvers slightly ahead of me. I didn’t feel close enough to finish before passing out, so I went to bed, expecting a winner to be crowned by the time I woke up.

One wasn’t. I guess you know that since I already told you I won, unless you thought I somehow finished the hunt in my sleep, which might be a better story, so feel free to skip ahead and pretend that happened. Actually, in the morning I noticed no one had solved the puzzle I believed was the meta, and I had some ideas on how it might work, so I decided to put more focus on that. Also, Jackie was now awake and able to contribute, and she made some progress on puzzles I wasn’t getting anywhere on.

I thought I was placing enough answers into the meta grid that my idea must be right, but I proved not everything could go in. Then I realized that if one of the letters was misplaced, everything could fit; so I pinged the main hunt author, and he confirmed that the grid was wrong, which was released as an erratum shortly after. That fix was enough to solve the meta, and we were the first to do so! But nothing special happened! I poked my head into the escape room Slack and asked if solving the meta meant we won. Nope. The goal was to solve every puzzle. (Errol seemed surprised… I’m not sure if he was surprised the meta could be solved without all the inputs, or just that anybody would try? Neither surprises me, but this is what I do.)

And so the backsolving began. Because enumerations were given and we had a few letters in some answers, we were able to guess two of them right away. Another took a bunch of guesses, because the answer we wanted, while common, was not an option that came up on OneLook (it was a title). One answer resisted backsolving, and we had a large chunk of that puzzle solved… Eventually Jackie cracked the last step of it, and we finished our complete.

So to answer the question I asked in the first place, if a hunt allows you to solve a metapuzzle but won’t let you be done unless you go BACK and SOLVE all of the puzzles that fed into it, then yes, that incentivizes backsolving.

What, if anything, did this puzzlehunt do to discourage backsolving?

Since the hunt was on ClueKeeper, there wasn’t anybody on the other end of the line to throttle excessive backsolving. CK itself does have some auto-throttling; if I submitted three wrong answers in a short period of time (guess how I know this), it would lock me out for a minute. What effect does that have on a solver? I can say that it slows you down, but it also means you feel no guilt whatsoever submitting an answer 61 seconds later. That may or may not be the desired effect.

What, if anything, did this puzzlehunt do to discourage short-circuiting?

Nothing, really. But if you did short-circuit (as we did), the “gotta solve ’em all” requirement made that accomplishment less helpful, unless you complemented it with backsolving.

What (in your blog host’s opinion) should be changed?

Mileage will vary here, but one of the things I find satisfying about puzzlehunts is that it’s a mad dash from Point A to Point B. Good metapuzzles don’t require you to have all the inputs, and the process of solving around the missing data is something I really enjoy. The goal of solving everything on the page feels more like a crossword than a puzzlehunt, where you’re not done until you’ve filled in every box. I enjoy crosswords, but I enjoy them more for the themes; once I know the theme, completing the grid is only appealing to me as a timed competition (or practice for timed competition).

In a structure like last year’s Cryptex Hunt finale, where you only see one puzzle at a time and you need to solve each one to unlock the rest, it makes sense that you have to solve everything to finish. But in an event like this one, where everything was provided at once, finishing the metapuzzle and then having to go back and clean up was a bit weird. Backsolving as a strategy to advance in a puzzlehunt feels like part of a game; backsolving to check the boxes to be declared the winner was anticlimactic. (Apart from that objection, I thought it was a well-designed event that was slickly presented and led to an exciting race that kept changing leaders. And I’m excited to have a gorgeous cryptex in our house soon!)


What elements of this puzzlehunt encouraged backsolving?

The Galactic Puzzle Hunt is a lot like the MIT Mystery Hunt on a smaller scale, which makes sense since it was created by a Mystery Hunt team. As I understand it, the 2017 GPH was actually created because Galactic Trendsetters had time on their hands after the 2017 Mystery Hunt ran short. Add this to the fact that the 2009 WALL-E-themed GPH was clearly inspired by the Escape From Zyzzlvaria theme my team used for the Mystery Hunt that year, and I think we can all agree that I am the spiritual founder of the Galactic Puzzle Hunt.

Anyway, there are a lot of gonzo puzzlehunt ideas, in terms of both puzzle content and overarching meta structure, that have rarely fit anywhere other than into the Mystery Hunt, and I admire the fact that GPH throws a lot of these ideas at the wall and most of them stick. I love both the Puzzle Boat and Mark Halpin’s Labor Day puzzle suites, but I think GPH has established itself as the premiere online-only puzzlehunt in terms of quality and ambition.

Anyway, my point is that GPH is a lot like the Mystery Hunt. There are a lot of reasons backsolving is beneficial during Mystery Hunt. So there are a lot of reasons backsolving is beneficial during GPH.

What, if anything, did this puzzlehunt do to discourage backsolving?

GPH has what I think is a brilliant policy for wrong answers. The various Australian online puzzlehunts (MUMS, SUMS, CISRA we hardly knew ye) often allow either 100 answer attempts per puzzle, or per day. This is a very reasonable cap if the puzzles are clean; they aren’t always, and I remember at least one Aussie puzzle on which my team spent 50 guesses before getting it correct, and we weren’t backsolving. I don’t remember it fondly.

GPH instead allows only 20 guesses per puzzle, but they note in their rules that if you use 20 guesses, and they judge you used them legitimately in an attempt to solve the puzzle (as opposed to, say, trying the name of every US state), they’ll give you more. This seems like a sane way to curb abusive backsolving attempts while still allowing for the possibility that a nasty puzzle might call for a high quantity of submissions. I think 20 may still be a high limit for an event with live submissions, but the GPH model works just fine for an electronic-submission event.

Having said all this, backsolving is only relevant if you have unsolved puzzles by the time you solve the meta. That didn’t happen a lot in this year’s GPH because…

What, if anything, did this puzzlehunt do to discourage short-circuiting?

My team, Killer Chicken Bones, was in second place near the end of the weekend before two things happened: (a) most of my team disappeared (half the team knew in advance they’d be leaving for two separate work trips in Hawaii, and another had separate work obligations… all this was before I had an ugly health scare later that week), and (b) due to the length of the hunt, the organizers started passing out meta hints and free answers like candy. All of this made me a lot less engaged in the event  and led to our team not finishing, but we probably would have kept more momentum going if we hadn’t gotten to the point where we had four puzzles left to solve, and NO METAPUZZLES AVAILABLE. Despite the fact that we had two of these left to solve. What?!

As it turns out, the structure was designed such that you would get to see a meta once you solved a certain (pretty high) proportion of the puzzles going into that meta. During the hunt, we had no idea this was what was happening; we just thought they required a lot of total puzzle solves to access the metapuzzles. Later I heard that one of the last metas we opened was the first meta another team opened. This certainly blocked both backsolving and short-circuiting. You can’t solve a metapuzzle with 6/10 answers if you can’t even see the metapuzzle!

This was especially problematic with this year’s structure/theme, which revolved around learning a constructed alien language. (I could go on about that for hundreds of words, but this is closing in on 3K already. In short, I liked the concept, and I enjoyed associating concepts with words/roots; I did not enjoy trying and failing to work out the tense/case infix conjugations, and ultimately I thought the depth and detail of the language design was too big for its britches.) All of the metapuzzles involved digesting certain aspects of this language, which took time. I find that in most puzzlehunts, I look at the meta(s) early and spend a lot of time thinking about it in parallel as I solve other puzzles. That wasn’t possible here because of how late the metas opened. Once we finally opened one, we had to start from scratch. Ultimately, we failed to solve the Artists meta, which certainly had the most work to do after opening it, and if it had been available earlier, we could have used our resources more efficiently.

What (in your blog host’s opinion) should be changed?

There is a traditional “eighty percent rule” for metapuzzle writing. The idea of this is that a good metapuzzle should be solvable with any 80% of the answers and ideally not much less. This forces teams to solve a reasonable number of feeder puzzles (and thus not short-circuit the meta too drastically) while allowing them to solve around a puzzle that might be broken, too hard, or simply not their cup of tea.

This is a hard balance to nail (I’ve failed at it many many times) but the solution is not to hide the metapuzzle until teams already have 80% of the answers. What makes a metapuzzle special is that it’s initially impossible, and it gradually becomes more approachable as you obtain more feeder answers from elsewhere in the hunt. If you don’t show teams the meta until they have enough answers to solve it, it’s not really a meta… it’s just one more puzzle. Albeit one that you have to solve to advance, and thus one you wished you’d had more time to look at.

GPH has done really cool and creative things with their unlock structures, and I have faith that they will keep trying new stuff (until they win Mystery Hunt and get to try stuff on an even more epic stage). I hope, at least in terms of what I like about puzzlehunts, that they consider this “don’t show them the meta until they have most of the answers” mechanic to be something that did not stick to the wall.


Well, first of all, I can’t believe you read this far. The hunts described above are three different hunts of different lengths run for different audiences, so it’s not surprising that they all dealt with metapuzzle “enforcement” differently, and the results varied. I wrote one of these hunts, I won another, and I crashed and burned in the third, and my instincts on how constructors should deal with backsolving and short-circuiting were sharpened by all three experiences. What are your opinions and instincts about it? Let’s chat about it in the comments. I’ve put a few hours and a few thousand words into writing this, so I’m going to bed, and I assume that by the time I wake up, Projectyl will have won the blog post.

2019 MIT Mystery Hunt, Part 5: Acknowledgments

(This is a recap/review of the 2019 MIT Mystery Hunt, which happened in January. Puzzles and solutions can currently be found here. These posts will contain many spoilers. Maybe not so much this one.)

It’s an atypical Thursday night at Chez Dan, as Jackie is on an overnight trip to NYC, and I’ve been alternating between solving logic puzzles and catching up on wrestling. (Okay, my portion isn’t that atypical.) In any event, this seems like a good time to write up my penultimate Hunt post, which may be of less interest to puzzle enthusiasts… I’ve said what I want to say about Hunt content, and now I want to thank some people who made Hunt happen.

I should start by saying that everyone on Setec (listed here) contributed to Hunt and everyone was important; I’m not writing this post to try to identify everybody or try to single out who did the most work, but rather just to tip my hat to the people who contributed especially to my own Hunt-writing experience.

Setec picked up a whole bunch of new members (from Illegal Immoral & Fattening, Left as an Exercise for the Reader, and Om [some number of Noms that I can’t remember]) this year specifically for construction. Setec was not what I would consider a small team when we won, but we weren’t big either, and the Hunt has grown to a scale that I don’t think we could have written without recruiting. So thank you to Sami Casanova, Chris Cieslik, Jesse Gelles, Avram Gottschlich, Albert Lin (actually returning to Setec after a long time away), John McLaren (repeatedly called Australian John despite being the only John on our team), Matto Mildenberger, Julia Tenney, Julia Urquhart, Rebecca Vessenes, Mike and Sandy Walsh, and anybody else I may have missed. Without you, Hunt likely would not have happened, and I hope some of you stick around to solve with us in January.

We had our executive structure mostly figured out before the new folks joined us, which probably made it difficult for them to wiggle their way into big-picture roles. Chris Cieslik thankfully insisted on being involved in the nuts and bolts of releasing and ordering puzzles, and he built some simulations that provided a nice counterpoint to my approach of “my instincts from running Hunt before say this feels right.” These are the advantages of having an experienced game designer in your ranks… I now realize I’m too late to plug his company’s Kickstarter, but you can read about the successfully funded game here.

I’d also like to credit Rebecca Vessenes for taking control of the midgame (the birthday party interaction) late in the year along with Justin Werfel. I led the midgame construction in 2017, and a big priority was making the whole thing run efficiently, since I knew we’d be executing dozens of times with varying groups of people. My biggest worry in handing it off was that whoever took it over might not be as organized, but Rebecca did phenomenal work in gathering and constructing the props and equipment, and more importantly, idiot-proofing the organization and instructions so that anybody could portray the Fool and their handler. Given that none of us had done any event planning with Rebecca before, we realy lucked into having the best possible person doing this.

Speaking of Justin, I don’t know if he noticed it at the time (it might just have been my being pompous), but the first year we hunted together with Setec, our personalities didn’t gel well at all. But I’m glad we’ve now had the chance to construct together, because it turns out he’s a great collaborator, and in particular, as I said in my story post, I really appreciate the character development (and scriptwriting) he contributed to weave the April Fool into the plot. In my hands, the Fool would likely have been a generic chaotic evil villain (that’s what I do), but the story we told was a lot more nuanced, and a lot of that is Justin’s doing.

Scripts are only as good as the people who perform them, and so thank you to Marisa Debowsky, Philip Loh, Josh Oratz, Greg and Margot Pliska, Scott Purdy, Julia Urquhart, and particularly Jesse Gelles and Steve Peters for bringing the kickoff to life. I had never seen Jesse perform before, and I cast him based on meeting him for a couple of days at our construction retreat, and he surpassed my expectations as Jack. In contrast, I’ve seen Steve perform tons of times, as we did a lot of musical theater together at MIT. Given that Steve gave me my first significant role in college (as Adam/Noah in Children of Eden), it was nice to finally get to direct him in something.

I cannot even fathom Matt Gruskin’s computer skills. Certainly he contributed to the website and made Hunt happen from a tech perspective, but I lost count of the number of times we needed an interactive implementation of a puzzle or a simulation of a tester saying “what if I asked the coding guy on my team to do this?” and our default approach was, “Someone ask Matt.” Frequently I’d tell him something was no hurry and still get a response in less than an hour that was exactly what I wanted. I was sitting across from him in HQ on Sunday morning when the Hunt website went down–conveniently right after TK, our head of tech, had left to sleep–and I watched as he got on the phone with TK, and they calmly identified and fixed the problem. (To be fair, when I say calmly, Matt may have just hidden his panic well, and I only heard half the conversation, so for all I know TK was screaming.) Arthur C. Clarke said that any sufficiently advanced technology is indistinguishable from magic, and as far as I (with my pitifully amateur coding skills) am concerned, Matt and TK are sorcerers that put Mystereo Cantos to shame.

Speaking of TK (Focht), I believe that after 2017 I said he was one of a couple of people who kept me sane through the year. That was true this year as well, and since I was running metas and he was running story/structure, we had to interact a lot. TK and I often have matching perspectives about Hunt philosophical issues, and the nice thing about having a person on your wavelength like that is that when they do disagree, you tend to trust them. I know I’ve done it before, but at this point when I imagine writing a Mystery Hunt without TK on my team, I curl up in a corner and cry. Okay, to be fair, right now I do that when I imagine writing any Mystery Hunt. (I’m done for a while, for real this time.)

Almost last but by no means least, there is Tanis O’Connor, who was in charge of the whole dog and pony show. Unlike my creative rhythm with TK, Tanis and I frequently don’t approach things the same way by default, and in 2017, we had a significant number of arguments and personality clashes. So to Tanis’s credit, when we won the Hunt again, she immediately initiated a conversation with me about how we could work together without getting on each other’s nerves. We still didn’t always see eye to eye, and we still had a few disputes throughout the year, but we collaborated much more productively than last time, and I suspect that where we met in the middle was farther from her comfort zone than it was from mine. I am deeply grateful for her flexibility and her patience, and that’s just in dealing with me; as editor-in-chief, she had to keep watch over all the puzzles and deal with an entire team worth of constructors with idiosyncratic creative processes, and she made it happen. If it weren’t for Chris Morse, Setec Astronomy wouldn’t exist, and I’m glad he let me join the team back in 1998(!) when I was still in high school. But if it weren’t for Tanis O’Connor, neither of the last two Setec Hunts would exist, and I hope people appreciate all the time she’s volunteered.

Also, I don’t know if I talked about this in an earlier post, but when I unsuccessfully tested Funkin’ at the retreat (in its original form that involved mini-CDs rather than USB sticks) I thought the concept was a logistical nightmare. I insisted that we should have a website version ready to deploy if we didn’t get the donuts as intended, which I honestly didn’t think we would. I eventually leaned extra-hard into my role as a “donut truther,” repeatedly referring to “the donuts that won’t actually exist” and consistently generating rude gestures in response from Tanis, who was guaranteeing she could make the donuts happen. Hear ye, hear ye, internet. The donuts happened and were one of the most memorable elements of Hunt for many teams. Tanis was right, and I was totally wrong.

Finally, I want to thank Jackie Anderson, because if you’ve never lived with me in a year when I was writing Mystery Hunt, you have no idea what she’s been going through. (I heard at least four accounts of executive committee members’ spouses, some of whom are on Setec and some of whom aren’t, looking forward to late January when they’d finally get their husband/wife back.)

Thanks to everybody I mentioned above, and everybody else on the team; if you pick a name on the credits list, I could tell you something they did that made this a better event. But more than 1600 words in, I’m going to call this done, get some sleep before Galactic starts tomorrow evening, and hopefully moderate a backsolving discussion in a future post before the Hunt is three months old. Cheers.

Some Non-Mystery-Hunt Housekeeping

Hi folks! I remember when I said I was a bad blogger because I was busy writing the Mystery Hunt. Turns out I’m just a bad blogger!

I have two Hunt posts left that I intend to write; most of the Hunt-specific content has been expressed, but I still want to thank a whole bunch of people, and I still want to spark a hostile debate about backsolving. But in the meantime, there are tons of exciting puzzlehunt events happening, so let me unleash a quick rundown:

The competitive portion of the 2019 Cryptex Hunt has concluded, but the hunt is still finishable, and well worth solving both for the puzzles (very elegantly presented in magazine form, bringing back pleasant Small Town News memories) and the potential prize (random drawing for a cryptex if you finish by the end of March). You may recall that in the 2018 Cryptex Hunt, I had an early (narrow) lead until the finale exploded due to technical difficulties. This year, I managed to win despite more minor technical difficulties (a nontrivial typo in the metapuzzle that has since been errata’d) so soon the most valuable aesthetic item in my house will be puzzle-related, which seems only fitting. I’m not sure what the current most valuable aesthetic item in my house is. Does an electric piano count? What if I never play it?

This Friday the Galactic Puzzle Hunt kicks off. After the last couple of years, GPH has become an event I seriously look forward to, in the same echelon as Mystery Hunt, Puzzle Boat, and Mark Halpin’s Labor Day puzzles. The puzzles have been high-quality, but the constructors have also been very creative with puzzle structure and story, and my Mystery Hunt posts should make it clear that that interests me and I want to subscribe to their newsletter. I’ll be solving with Killer Chicken Bones, which will be extra-challenging because some of my teammates are leaving mainland America after a few days, if they’re not eaten by The Thing (or Cookie Monster?) first.

Applications are currently open for August’s Boston-area Miskatonic University Game. I really hope to participate in this, so I encourage you to submit applications only if your application is slightly less good than my team’s application. If I don’t get in, maybe those of us who are rejected can have a party that weekend where we play Arkham Horror and sulk.

There was a new P&A this past weekend, continuing the trend of releasing issues at times that are awkward due to soccer… the release was at 12pm, and the Revs home opener kicked off at 2pm. We solved eight puzzles (out of twelve) before the game, and during the game I put the puzzles away as promised while the Revs gave up a goal and the supporter groups took it out on each other. During half time, I saw that seven people had solved the meta and freaked out… Then we got the meta aha and submitted the meta answer to make the top ten shortly before the second half. We were triumphant. The Revolution were not. (Note: Am I the only one who actually pronounces the name of the magazine “P and A”? I feel like I’m running into more and more people who say, “Have you solved Panda?” and now I’m wondering if I’m the crazy one. If so, that should improve my chances for Miskatonic University, right?)

Also on the horizon but further out… MUMS! DASH! Maybe a BAPHL or two (their website says May and July)? Will this blog ever feature non-MIT content that you couldn’t just get for yourself by clicking some links at Stay tuned, true believers!

2019 MIT Mystery Hunt, Part 4: Puzzles

(This is a recap/review of the 2019 MIT Mystery Hunt, which happened in January. Puzzles and solutions can currently be found here. These posts will contain many spoilers.)

Happy Presidents’ Day! Hopefully you’re celebrating in the traditional way: invading Thanksgiving Town and Valentine’s Day Town. I will celebrate by talking about puzzles I was an author or co-author on in this year’s Hunt.

Puzzles I Wrote

I almost made that heading “Puzzles I Wrote By Myself” to differentiate from the next section, but honestly, nobody wrote any puzzles in this Hunt by themselves. We had a very hands-on editing staff that helped refine ideas, give advice during writing about both puzzle mechanics and content, and suggest revisions before and after testing. My puzzles were definitely improved by the editors who worked with me, and I hope I helped with the puzzles I edited (beyond the handful where I ended up butting in enough that I was promoted to co-author).

* I Knew Weird Al Yankovic, And You, Sir, Are No Weird Al Yankovic (Halloween)

In 2017, I wrote a puzzle called Before & After MASH that I was pretty happy with. It consisted of a bunch of alternate lyrics I sang to TV theme songs, about subjects that overlapped the titles. I heard that some teams played the songs for their entire HQ and solved it “shout out” style, which is the ultimate compliment. I was definitely hoping to accomplish that again, and at some point I randomly wondered what the lyrics to “Shingle Ladies” would be (and answered that question for myself in two minutes), so I decided that was the thing. Whenever possible, I tried to end the snippets right before the title would be sung, in the hopes that one of those listen-alongs would result in people singing the answer together.

I limited myself to single-word artists (no “The” allowed) and tried to use the most well-known artists I could stand. This forced me to use Vengaboys, because while I loved V.A.S.T. in college, they’re (he’s?) way too obscure. I expected to use Radiohead, one of my all-time favorite bands, but the fact that many of their songs don’t include the title in the lyrics was problematic; I was complaining about it to fellow Radiohead fan Tanis and she suggested Roxette, which was one of my favorite tracks to record, so props to her. The background music is all from karaoke tracks I pulled from YouTube, and I recorded vocals over them in my bathroom using Audacity. The worse I sound on a track, the more takes it probably took to sound as decent as it did… my singing voice is sadly not what it once was. At some point we decided to alphabetize by original song title, and I was horrified that this put Kansas and Styx first, which are definitely my worst performances.

When I submitted this, I also suggested an alternative puzzle called Your Own Medicine, in which I would write parodies of non-parody Weird Al Yankovic songs, like “Skipper Dan” and “Frank’s 2000 Inch TV.” This option did not get chosen, because (a) it’s infinitely less appealing than what I did write (except to a small niche group of superfans), and (b) a lot of Al’s originals are artist style parodies, and I’m not sure that a parody of them wouldn’t sound more like a parody of the homaged artist. (“Is that a new version of “C.N.R.” or “The Denial Twist”?)

* Spinning Tops (Halloween)

I think this was the first puzzle in the entire Hunt to graduate from testing. Most Hunt puzzles either start as an outside (or theme-related) idea that is in search of a best-fit answer, or they start by using an answer from the answer list as inspiration. This was the latter; I knew SBARRO was in an early round and thought I could whip up a pizza-themed puzzle pretty quickly, so I did.

The original slicker title to this was Topspin; unfortunately, we discovered in the first test that TOPSPIN is a transdeletion for TOPPINGS, and that sent the solver in the wrong direction. I’m still not in love with the revised title, which doesn’t have much of a ring to it.

* DK8: The Turducken Konundrum (Thanksgiving)

Coming up with the title was easy. The rest was hard.

As soon as I knew we were going to have a Thanksgiving round, I figured we should have a turducken konundrum. I wasn’t sure what that meant though. Initially I figured the solver action would involve stuffing objects into other objects as one does when making a turducken. But this was unsatisfying on many levels. DK6 was about cooking. DK5 involved a bunch of random objects, and lots of previous duck konundra involved moving stuff around grids or another playing surface. I wasn’t really sure what would stand out about this particular set of wacky instructions, which meant I wasn’t very motivated to write it.

Then at some point I had a revelation: a turducken konundrum isn’t a konundrum about things inside of things inside of things… It’s a konundrum inside a konundrum inside a konundrum! I posted this on Puzzletron on September 16 (the post began, “I HAVE HAD AN EPIPHANY”), and started brainstorming various elements, such as who would be solving the Hunt on each “level.” But it was still two months before I found time to sit down and actually start writing steps. Once a draft was written, in place of my traditional “test where I look over the solvers’ shoulders at every step,” I re-solved the whole thing myself in order to write the solution. Then it got two group tests in Seattle and New York, as well as Jen Berk solving it herself on a spreadsheet, which is apparently a thing she likes to do, and we decided it was ready to roll.

I’m pretty burnt out on writing duck konundra (and have been for years), but I’m very happy with how this one turned out. Given the meta theme, if this turns out to be the last “official” duck konundrum, I think it’s a good way to wrap things up.

* Crosscut (Arbor)

I don’t know where this idea came from, but I liked the idea of reuniting scrambled haiku, and if the solver was reordering sets of nine bits, I figured iterating that step on nine-letter movie titles would make sense. Some early testers were very grumpy about filling “gibberish” into the crisscross grid; taking something meaningful and intentionally making it scrambled felt wrong to them. My favorite tester quote: “The twist is clever. Is it ‘fun’? I don’t know. But I respect it.”

* Connect Four (Holi)

I test-solve for Grandmaster Puzzles, which is one of the best sources for abstract logic puzzles in the world. (The only reason I don’t simply say it’s THE best is that I don’t regularly solve Toketa?, which is also supposed to be of extremely high quality.) I asked Thomas Snyder shortly after we won Hunt if I could hijack a week of puzzles to use as Hunt puzzle fodder. Thomas gave me a lot of latitude, under the condition that the puzzles were independently solvable and met the site’s usual quality standard.

Several other GMPuzzles testers are also Mystery Hunters, so a tricky obstacle was not making it totally obvious in advance to them that this was part of Mystery Hunt. I attempted to hide this by using a pseudonym, though things got awkward when the testers rightfully questioned the randomness of the letter assignments and started trying to “improve” them. As it happens, one GMPuzzles tester told me after the Hunt that while he was sleeping, the team had found the connection and solved all the fillomino from scratch, not realizing that the tester had already solved them earlier (and had the official solution file).

* Gone Guys (Pi)

I still watch WWE wrestling regularly (I have Raw on now as I write this), but I’ve gotten really into New Japan Pro Wrestling in the last couple of years, thanks to their global streaming service, and I’m fascinated by how much work goes into the booking of the G1 climax tournament. NJPW is a company where wins and losses matter–when a champion gets beaten in a tournament match, the opponent is expected to get a significant title shot at some point–and all the point totals have to add up appropriately for the final day to matter, yet without earlier results making it obvious who will be in contention for the final. Compared to a WWE where currently the chairman rewrites the major shows hours before they air, it’s amazing how much planning (and math!) is put into this every year.

Anyway, I was curious how few results it would take to determine who was who in a tournament block, and that was easy to turn into a puzzle (written entirely by hand, so I knew there was a valid solving path that a human could work out). I heard from lots of non-wrestling-fans who got the aha, knew nothing about who the wrestlers were, but still really enjoyed the logic aspect. And I also heard from one wrestling fan who thanked me for writing a puzzle specifically for him.

* Ore Aft (Pi)

This was another “inspired by the answer” construction, as I asked Tanis for a handful of answers that hadn’t been claimed and one was notably part of a sequel subtitle. As it happens, I’d drafted a Fourth of July + Christmas meta that involved subtitles of fourth movies in series, and that hadn’t made it out of committee, but I still had a huge list of fun movie subtitles. I didn’t spend too much time on this puzzle or think of it as a highlight, but a lot of people cited as a favorite after the Hunt.

* Standardized Mess (Bloomsday)

This was kind of a bold construction, in that I was at first not sure how to clue the aha, and then I decided I would thematically clue it with every single question on the quiz. I was hoping for an effect where solvers would not see this for a while, and then all of a sudden they’d see everything at the same time (kind of like the 2001 Hunt endgame, where there was an entire deck of tarot cards hidden around the room with five in plain sight, and we spent 45 minutes staring at the five, before someone found one card, and then we found a bazillion cards all at once). Judging by Eric Berlin’s blog post, Palindrome had an experience along those lines, but the delay was a lot longer than I intended. Sorry, folks.

An awkward anecdote: In the first test, I made a construction error so that instead of flipping four of the grids, all the grids actually appeared in the orientation that would yield the answer, which they eventually found without doing anything with the quiz. Some members of that group tried to be diplomatic, but others were pretty clear (assuming the quiz was a giant red herring) that they thought the puzzle was dumb. And I agree, that version of the puzzle was extremely dumb!

Puzzles I Helped Write

* Haunted (Halloween)

When Alex Rosenthal found out he was going to be able to do a TED talk about Hunt, a bunch of editors got thrown together to quickly brainstorm the best way to embed information for a 2019 puzzle. The tricky thing was that we knew people would see the talk before Hunt and would look for puzzle information. I think I was the one who suggested a puzzle with a clear answer, so that those people would think they were “done,” but that could be solved differently with a clue phrase clarification to yield a different answer (and the answer AMBIGUOUS was perfect for this state of affairs). I came up with enough details for the mechanic we eventually used that Alex graciously invited me to be a co-author, though he did most of the heavy lifting on clue construction (and, of course, giving the talk.

I initially suggested the title Hunted, which we used in the first test, and the testers thought since it was in Halloween, Haunted would be more thematic. IN retrospect, maybe we should have used H(a)unted.

If you’re looking for additional pro wrestling content, yes, that is a Kevin Owens shirt I’m wearing in Alex’s TED talk!

* If At First You Don’t Succeed (Thanksgiving)

Setec held a weekend retreat in September to try to jumpstart testing and construction. As part of the retreat, Chris Morse and I led puzzle construction seminars, and my seminars had a format where Tanis gave me a handful of unclaimed answers and we brainstormed puzzle ideas for them. This was very productive, though I ended up as a co-constructor on most of the ideas, so it made my plate very full in a short period of time. (Fame is Fleeting and Compromised also arised from these sessions; I edited both but didn’t co-construct.)

This puzzle was basically triggered by the observation that the two words of the answer started with the same trigram, and that the remaining letters could be clued independently. We considered giving this without the accompanying trigrams, but given that it was in an early round, we wanted to err on the easier side. If it was in a late round, we probably would have done the opposite.

* Twelve! Eleven! (New Year’s Day)

I like puzzles with lots of mini-ahas even once you know how the puzzle works, and this was one of those. I recruited Jackie to help brainstorm list ideas, and to help me narrow down a larger list of lists to the best set of options.

* Middle School of Mines (Arbor)

This was mostly Philip Loh’s baby (a sequel to his 2017 puzzle School of Mines), but I gave the suggestion of creating a final logically solvable “ubergrid” where all the absent instances of one number would be used as an extraction. Then I wrote that final puzzle, which was enough to get me an author credit.

I now find the fact that he called this “Middle” School of Mines troubling. Philip, we’re not winning again. Stop that.

* Rules of Order (Holi)

Of the three puzzles I co-wrote with Jackie, this was the one that was primarily engineered by her, which makes sense because arithmetic dynamics is what she does for a living. She had the idea of getting solvers to identify Julia sets, and I suggested also identifying Julia Roberts movies at the same time, and we weaved those ideas together into the final product.

The original draft gave the movies in extraction order, but shortly before testing, I had the unnerving realization that if you knew nothing about Julia sets but identified all the movie character names, you could do a computer search for ten-letter strings that pulled a letter from each name in order, you could narrow things down to a couple of options, one of which was clearly thematic and likely correct. To combat this method of short-circuiting, we came up with the “order by magnitude” mechanic, which ensured you’d need to know which letters you were taking *before* knowing what order they should be placed in. Hopefully this stopped people from bypassing the math.

* Protection Plan (Pi)

This was another retreat brainstorm based on the answer ZOEY. Somebody remembered that Zoey Bartlet had a fun Secret Service codename, and in doing some research we discovered that Secret Service codenames of fictional characters is a fun and extensive data set. Someone suggested connecting the data set to the game Codenames, and that’s when it started to feel like a full idea to me. There had recently been both a Secret Service codenames puzzles (with real codenames) and a Codenames puzzle (emphasizing the red and blue elements) in Hunt, but we decided the mechanics in this puzzle were different enough from both of those to justify its existence.

* Letters From the Battlefield (Patriots’)

Mark Leach, a fellow Midnight Rider, beat us to putting an MLS puzzle in Hunt, but once I knew Patriots’ Day was a round, Jackie and I were definitely writing a New England Revolution puzzle. Spelling an answer in chronologically ordered away game locations in consecutive years was very constrained, and once we had the game set, Jackie and I both generated lists of notable game elements, but then I wrote the majority of the old-timey game descriptions. This was tedious. I hope this was more fun to solve than it was to write.

Kelyn Rowe shows up in a lot of these descriptions, partially because he’s my favorite soccer player, and partially because he just happened to do a lot of notable things in these games. He was traded this off-season. 😦

* Broken Concentration (Bloomsday)

This was another retreat brainstorm, with Matto Mildenberger leading the charge in suggesting that this answer should be a “concentration” game where the undersides of the cards change every time. (A good way to brainstorm Hunt puzzles: What variation on Puzzle Type X will cause the solver to swear at the constructor thirty seconds after opening the puzzle?) Most of the mechanism got hashed out in that room, and then we just needed to implement it. Sincere apologies to anybody who revealed all the pairs in the hope of getting extra information at the end.

* Getting Digits (Bloomsday)

Chris Morse wrote this puzzle as a 2017 backup puzzle that didn’t get used, although in that form it was just a list of clues for each segment. At the beginning of the year, I wrote a manifesto about how to avoid both Identify-Sort-Index-Solve and “list of clues” puzzles, so if we used this, I really wanted it to look like its own thing, so as an editor I proposed the “scroll readout” presentation we ended up using (and the START clues that would add an alignment step), and Matt Gruskin made it look awesome, as he did with many puzzles (Broken Concentration, for example).

Chris was indisposed during much of the period where we were writing the START clues, so I ended up writing those and got promoted to co-author.

* Keeping Tabs (Bloomsday)

Yet another retreat brainstorm, and yet another Matto inspiration, who was the one to note that UKULELE has four unique letters and four strings. He wondered how many other instruments have this property… enough, as it turns out. We ran into a bit of a block trying to figure out how to clue the bank words in a way that felt like it belonged in the puzzle. Someone suggested a series of couplets being “sung” by the ukulele player, and I predicted that was too constrained to be well-written. Justin Werfel quickly proved me wrong with a great set of rhyming clues.


2019 MIT Mystery Hunt, Part 3: Metapuzzles

(This is a recap/review of the 2019 MIT Mystery Hunt, which happened last month. Puzzles and solutions can currently be found here. These posts will contain many spoilers.)

While I very slowly answer questions you don’t have, it might be useful to know that my Setec teammates and I answered questions you actually did have on Reddit last week. You can read the discussion here.

In the last post, I talked a lot about how we developed an intricate web of metapuzzles that all fit together into a nice closed map. One thing I didn’t talk about, which people have asked about here and there, is how we tested for the one element of our Hunt that was arguably most variable: solvers’ ability to determine which answers went with which puzzles.

The short answer is “not that much.” The Mystery Hunt is unique, and there’s no way to tell how 100 teams of varying sizes and skill levels will grapple with things until you actually find out in real time. And as noted earlier, we had to test metas as we went in order to know what fit on the map and what other metas still needed to be written; so when testing Halloween-Thanksgiving, we couldn’t show our testers all the other Halloween and Thanksgiving answers because we didn’t know what they were yet. One thing we did do for some tests, usually a second test after things had gone okay the first time, was introduce decoys (fake answers) in both rounds as we revealed answers to testers. But it’s one thing to have noise when you’re focused on a single meta.

The second thing we did was much later; a few weeks before Hunt, Tanis tracked down a handful of puzzlehunt-savvy solvers in the Seattle area who were not planning to participate in Mystery Hunt. (This is hard to do, as most of the solvers with the level of experience to succeed at Mystery Hunt are excited enough that they find a way to participate for real!) I’ll reference this below as the “Seattle Test”… We basically fed solvers a steady diet of new answers and had them focus on metapuzzle solving. At that late stage, we had no ability to cut metas or change puzzle answers, but we could use the results to decide if presentation should be tweaked, and predict how people would use solvent (as well as testing the website’s ability to process solvent).

The group had about four or five solvers present on average for one afternoon, which is not a good gauge for how the Hunt is going to go; as it happens, they solved three metapuzzles (and backsolved a few answers) but didn’t get very far beyond working on the metas that were in the “Town Square” (the five initial holidays) or connected to it. We did get enough data to conclude that (a) our early metas might be a little bit harder than we thought, and (b) figuring out which answers go where might be as well (though to be fair, this was something Chris Cieslik always thought would be more challenging than I did, so hat tip to him).

As a result of this, we tweaked the flavortext on a few of the metas to make them slightly more approachable, and we also decided to give out solvent a bit faster than we initially intended (probably about twice as fast). The Seattle Test group were initially skeptical of the usefulness of solvent, but once they started playing with it, it definitely seemed to help them make progress when they were otherwise stuck, which is what we hoped.

Metapuzzles Primarily Written By Me

Halloween/Thanksgiving: This was the first meta I wrote, and quite possibly the first meta anyone wrote; I came up with the mechanic in February and came up with a way to fit it thematically to either of the two Hunt themes I was advocating for! The combination of ABO and ternary seemed natural enough that I figured it must have been used somewhere else, but I couldn’t find a previous instance. The initial version had two key differences; the ordering was alphabetical with answers starting with letters from C to L, and the answer was A NICE STAKE, suggesting that the resolution was actually a way for the Thanksgiving Townspeople to ward off the vampires. The ordering changed due to overlap with the Halloween/Valentine’s mechanism, and after the Seattle Testers had difficulty noticing the new length-based ordering, we added a sentence hinting it. As for the answer, it was too guessable with few answers in testing, and we also decided our resolutions should be more peaceful… I believe Todd McClary came up with the excellent pun we used instead.

Valentine’s/Arbor: I had been interested in using the names of trees in a meta, and when one of our finalist maps needed a Valentine-Arbor connection, I came up with the idea of carving initials into the tree names. Initially I was only going to add two letters to each tree, but that turned out to be too constrained. It was also hard to make good bigrams out of the letters that formed good answers, so eventually I decided to put in shell elements (the X’s and O’s) that explicitly reordered the letters. The Seattle Testers didn’t make progress on this meta, but when I explained the answer to them I got worried that the mechanism might be too much of a leap, so I suggested adding a plus sign in the middles of the carved hearts to suggest the missing initials. TK didn’t think this was necessary, and ultimately I think he was right.

Holi/Pi: I’m not convinced the Trivial Pursuit connection here was my own; I think there was a test solve of a different Pi Day meta where testers were brainstorming connections to pi/pie, and someone mentioned Trivial Pursuit, but it didn’t lead anywhere. I kept thinking about it, and given that our meta interpretations of Holi were color-based, “changing” the colors of special edition TP categories seemed like a fun mechanic. (I also like both word-association metas and metas where a bunch of normal looking answers actually belong to the same data set, and this allowed me to play with some of each.) There were a lot of more unusual editions I hoped to use in this meta, such as World of Warcraft and Lord of the Rings, but it turns out TP changed orange to purple at some point, and so to use the “classic” set of six colors, I was restricted to fairly early editions. The exact form of indexing is dangerously close to being a “what am I thinking” puzzle… we tweaked the flavortext a lot to try to make it clear how the solver was meant to interpret things.

Some Comments on Metapuzzles Written By Others

Halloween/Christmas: I won’t get into as much detail about the metapuzzles other folks wrote (or even mention all of them), but I’ll share a few anecdotes/opinions here and there. The authors of this meta found that it required either ROBERT EDWARD CROZIER LONG (who is, basically, some dude) or IN THE SHADOW OF ZEKROM! (the name of a Pokemon episode) to conceal a TWO ZERO. We settled on the second, but nobody noticed until after a Halloween Pokemon puzzle was written with the answer NURSE JOY that we had stuck two Pokemon-themed answers in Halloween. Since that was the first round where metas would split, we wanted to avoid red herrings there as much as possible, so I insisted on switching back to Mr. Long, and agreed to write a puzzle with that answer myself.

Halloween/Valentine’s: Great mechanism, glorious pun answer. As Brian has said elsewhere, this was initially a Halloween/President’s meta with the answer HAIRIEST TRUE MAN (and using two separate grids, I believe).

Thanksgiving/Presidents: We saw in the Seattle Test that this had major red herring potential… Many of the turkey names contain food items, and once you convince yourself that it’s a food meta, how do you convince yourself that MAY THE FORCE BE WITH YOU goes here and FRANCIS BACON doesn’t? After that test, we made a subtle flavortext tweak to refer to “their” annual tradition, in the hopes of getting solvers to think about what presidents do for Thanksgiving. As noted at the wrap-up, the original answer proposed here was simply GEORGE WASHINGTON CARVER, and our testers guessed it purely from the flavortext (and implied answer length), so we changed it.

Valentine’s/Presidents: This was initially a Valentine’s/Thanksgiving meta (with references to “stuffing” and an initial answer of CANDY CORN BREAD) but that pair didn’t fit into our final map. But we liked the mechanic for an early meta, so this got rejiggered for Presidents’ Day. The mechanism stayed mostly the same from conception to delivery, but we did change the presentation of the extraction hearts from a line to a circle to make it harder to wheel-of-fortune the answer. The Seattle Testers struggled with this, causing me to propose changing “romantics” in the flavortext to either “couples” or “pairs of lovers”… I advocated for the latter, but the authors thought that was too much of a giveaway, and we compromised. This played slightly harder than intended, so I still wish we’d used the stronger hint, but ultimately it was fine.

Bloomsday/Arbor and Bloomsday/Pi: Those of us keeping an eye on the big picture got very nervous about the Bloomsday metas, particularly because the answers in Bloomsday were arguably the hardest to divvy up. The answers that went with Arbor were only notable in that they’d need a lot of Scrabbly letters to form a pangram; the others were odd-length, but in the initial version, there was no grid to even suggest that odd lengths would be reasonable. Leading up to Seattle, I was trying to convince the authors that we should give solvers the shape of the first grid… the Seattle Test didn’t get close to Bloomsday, but given the difficulty of the other metas, I thought this was a necessary change. We still thought Bloomsday’s metas would be a stumbling block, and our focus on that left most of us blind to the real bottleneck…

New Year’s/Patriots and Holi/Patriots: Although Patriots’ Day opened before Bloomsday, New Year’s/Patriots was the last solve for both contender teams. This was not because teams had trouble deciding which answers went where or even how to solve the meta; at some point Sunday afternoon we visited Left Out and Palindrome, and both teams were able to tell us exactly how the metapuzzle was meant to be solved. It was just very hard to do so with incomplete information. (There’s been some speculation about whether we intentionally wrote metas that required all or close to all the answers… we did try to avoid metas that could be short-circuited with less than half the inputs, but requiring completes was never a goal.) For a while, Left Out abandoned traditional solving and tried calling in various clever puns… little did they know that this was the one meta where we didn’t push very hard in this direction. The answer we used was a play on Lexington, MA vs Lexington Ave, but the rest of the phrase was kind of arbitrary. Guy Jacobson did say at some point that he thought they could make any 11-letter phrase work in the Game of Life setup… We just didn’t come up with anything of that length that worked better. Much of the Hunt was a dead heat between Left Out and Palindrome (though Left Out more frequently led on metas solved), and a big part of the final difference between them was the ability to work out where the answer should go and reverse-engineer an input that would complete the answer.

As for Holi/Patriots, I was an early advocate of this meta, as I really liked the concept and how it worked. We used Ben’s rings with minimal changes (apart from getting rid of a pair that clued the red ring with “OXBLOOD”), but Ben kindly let me take a second pass at a path that used the rings more extensively and spread out the letters used. I intentionally added some checksum statements that would help confirm the numbers of letters between intersection points, and I hope that helped solvers build the rings more confidently.

April Fool’s: We brainstormed this meta after all other metas were written, and while we tried a few different mechanics, I’m glad we went with the “holiday pranks,” which tied the various rounds together nicely and allowed us to draw solvers’ attention to the art for a while. As for options we didn’t use, if you find yourself trying to write an April Fool meta for your own holiday-themed puzzlehunt, I can confirm that there are a whole lot of pop songs with “fool” in the title, and a whole lot of Shakespeare characters identified as “fools.” FYI.

Metas, yeah! Next time, let’s chat about some individual puzzles.

2019 MIT Mystery Hunt, Part 2: Metapuzzle Development

(This is a recap/review of the 2019 MIT Mystery Hunt, which happened this month. Puzzles and solutions can currently be found here. These posts will contain many spoilers.)

I was spoiled for the last couple of years by having a “buffer week” between Mystery Hunt weekend and the beginning of my semester; sadly, that was a temporary luxury, as this year and next year, my classes begin a couple of days after wrap-up. Interestingly, I usually get a cold right when the semester begins, but this year it came later; so apparently a week after Hunt is actually when my immune system goes on vacation.

When last we spoke, I was lamenting the fix I’d gotten myself into, having committed to overseeing the meta construction/testing process for a Hunt in which metapuzzles had to be linked in a way that made construction/testing a massive pain. Hooray!

Those who know the Mystery Hunt well (and honestly, if you don’t, none of this is going to make any sense anyway) will recall that the strongest Hunt teams finished Setec’s last Hunt, 2017’s Monsters et Manus, in a much shorter time than usual. A lot of hunters, including some on Setec, were perfectly happy with this, since it allowed many more teams to finish the Hunt. But a lot of us, myself included, preferred for the first finish to occur a bit later.

I felt that one cause of the quick finish was how we tested metapuzzles for 2017; since 2011, Mystery Hunt puzzles have generally been written and tested using a handed-down platform called Puzzletron. One of Puzzletron’s best features is that once a puzzle goes into testing, testers can claim a puzzle, solve it (or attempt to solve it), and give feedback without the need for a moderator. In 2017, we set up metapuzzle testing in a similar way, allowing team members to claim metapuzzles to test on their own at their convenience. This allowed us to test metas more efficiently, but it also lent itself to generating easier metas. A lot of testers would finish metas solo, and so much of the feedback was based on how approachable the metapuzzles were by one person working alone. This is very different from how Mystery Hunt teams approach metas, and so when our solo-solvable metas were thrown to actual Hunt teams of rabid wolves, they got torn apart.

As meta coordinator, my plan to combat this was to do all of our meta testing and construction in groups, which is how we built metapuzzles in the 2009 Zyzzlvaria Hunt (scour the solution pages for those and you’ll find a lot of credits for Meta Group A and Meta Group B). So we began by splitting the entire team into three groups, trying to balance experience, interest in testing, and interest in construction. We also tried to ensure that each group had a time of week they could regularly meet on Slack, though groups were encouraged to collaborate regularly between these meetings.

We started by having each group (helpfully named Apple, Orange, and Tomato) generate their own ideal list of A holidays and B holidays, where A holidays were well-known holidays likely to appear in the first part of the Hunt (which would be unlocked earliest and reveal the “midgame” if completed), and B holidays were less obvious holidays that would appear later. We then merged those lists into a larger list for everyone to work with. The list contained all the holidays we used, plus the Fourth of July (A), Leap Day (B), Mardi Gras (B), Talk Like A Pirate Day (B, became an event), and Your Birthday (B, became a cover story for the April Fool in midgame). On our original list, New Year’s was an A holiday and Presidents’ Day a B, and those ended up switching.

We then had a round of construction where teams were asked to focus on writing approachable A-A metas and more challenging B-B metas; I think we allowed A-B meta proposals at this time, but warned they wouldn’t get in unless they were very strong. Groups wrote metas, we tested each with another group, and then each was left as is, scrapped, or revised, before testing with another group. After doing this, I worked together with various editorial types to identify the pairs we were likely to keep, and we designed some potential maps / partial maps.

Then we we went back to the groups and repeated the process, aiming for pairs that were likely to link the partial maps we already had. Some time during the second or third of these cycles, we noticed one of the three groups was less active than the others, so we split them up and redistributed them to the other two. We still wanted to be able to test things more than once, so we had to split each of the (now two) meta groups into smaller groups for testing; as the leader of Meta Group Orange, I found myself sending testing e-mails about Subgroup Mandarin and Subgroup Clementine.

A lot of great metas got generated that were discarded either due to not fitting well in the map, or just not testing the way we wanted. TK shared a diagram during wrap-up (viewable at 6:46) that emphasizes how many metas got proposed, written, and/or tested, and hopefully it communicates how challenging it was to build a set of metas that fit together into an elegant map; in particular, I was very insistent on a symmetrical map so that teams would get a sense of how it was fitting together as they opened new nodes. Off the top of my head, there was a beautiful Valentine’s-Fourth of July meta that got dropped because it was too easy to solve (and then trivially backsolve) with very few answers, and a fun Thanksgiving-Mardi Gras meta that simply didn’t fit the map. In addition, a whole bunch of puzzles that made it into the Hunt were adapted from proposed metas; We See Thee Rise was once Thanksgiving-Arbor, and Chain of Commands, Picture Book, and Schematics (minus the logic puzzles that ended up being the meat of that puzzle) all began life as metas attached to Bloomsday.

This entry’s getting a bit long (and my head is getting increasingly filled with gunk as it gets late at night), so I’ll continue metapuzzle discussion next post with some behind the scenes details on the metas we actually used.


2019 MIT Mystery Hunt, Part 1: Structure and Story

(This is a recap/review of the 2019 MIT Mystery Hunt, which happened this month. Puzzles and solutions can currently be found here. These posts will contain many spoilers.)

How did we get to Molasses Awareness Day? As TK Focht noted at wrap-up, some members of our team (okay, mostly Scott Purdy) have been advocating for a molasses-flood-related theme for years, to the point where a lot of Setec was saying we should wait to try to win in 2019. As it turns out, we got the itch earlier than that. Then we forgot to retire. I seriously did not think we were going to win in 2018.

That said, “molasses flood” itself is not the sort of theme that can sustain 100+ puzzles. (We could have woven it into a Willy Wonka theme, but we rejected Willy Wonka as a theme in 2017 due to at least one team member’s fear of Oompa-Loompas, with me warning Setec that if we didn’t use Wonka as a theme for Hunt, I’d use it for something else. And I did!) So we considered a wide array of other themes with varying levels of molasses content. I think we ended up with about 20 options on our first team-wide preference vote, which occurred before we brought a bunch of new members (I’ll say much more about those new members later on; they saved the Hunt this year and they deserve truckloads of gratitude), which narrowed the field to four clear finalists. I’ll call them A, B, C, and D, without revealing their identities, so that we can repurpose one of them when we run Hunt again in 2083.

One thing that became clear early in the process was that the team was interested in messing with a traditional meta-round structure, likely by putting a bunch of answers in one pot and requiring teams to “sort out” which answers fed into which metas; this is a thing that isn’t obviously solvable, but Foggy Brume (not on Setec) has proven in several Puzzle Boats that it can be done, so we were intrigued. All of A, B, C, and D had some degree of this. C and D were themes I was not jazzed about, so I will not expound further on them. B was my idea and had a multi-phase theme that used “sortable” metas in only one phase, which I figured would be easier to build; I really liked the structure we came up with and was endorsing this as my top choice.

Theme A is the theme you saw this year. It evolved from a TNBC holiday misunderstanding theme that Brian Tivol posed in 2005 (not a typo) when Setec wrote Normalville. He threw it in again this year, and I had the idea that if we wanted sortable metas, we could use a graph in which holidays were vertices but metas were edges, so that teams would have to figure out which puzzles were escaping into which other towns. This also seemed to lend itself well to a molasses flood as an inciting incident; a flood would explain why the walls between towns had been newly breached.

After welcoming our new teammates, we spent some time developing all four finalists in Slack channels and then voted on the final proposals (ranking all four). Theme A won according to pretty much any voting system you could choose to apply to the ranked lists, and I understand it was 1st or 2nd on almost every ranking. My first choice of Theme B, also featuring a molasses motivation, was apparently frequently ranked 1st or 4th. Hashtag polarizing.

It was around this time when I realized I’d shot myself in the foot with my meta structure suggestion for Holidays. In 2017*, I had gotten very frustrated with the pace of meta construction, peaking with me flipping out at Tanis O’Connor (who wasn’t even in charge of metas either, sorry, Tanis) in the middle of a trip to Europe. So when we won again, I thought the best way to keep things running at a pace I was comfortable would be to take charge of metapuzzles.

[*I realize that by “in 2017,” I actually mean “in 2016 for the 2017 Hunt,” but if I worry about that level of detail I’m never going to finish these posts, so any years refer to the associated Hunt.]

But in 2017, we generated a whole bunch of character and quest metas, loaded them into an online system where people could test them on their own time, and eventually picked the ones we liked most. We couldn’t just pick the best ones in this case; we had to pick metas that actually fit together into a graph, since if a holiday only showed up once, its sorting would be trivial. We initially intended to have none of these, but we eventually decided it would be good to have one degree-1 holiday as an “orientation” round, and given our source material, Christmas-Halloween seemed like the natural pair to reveal first.

So anyway, finding a way to build a map full of metas was now my problem, and I have lots to say about it, but that probably fits more into a post about metapuzzles (likely coming next after this one). So let me say a little about story.

Those who have written Hunts with me before will not be surprised to hear that I wrote the kickoff script. I’ve done that for seven out of eight Hunts I’ve worked on, and I think I have a good sense of the sort of wacky self-referential humor that functions well in this setting. I think the audience wants larger-than-life characters that help them enter the world of the Hunt, but they also know that they’re looking at a bunch of puzzle constructors who are puzzle hunters just like they are, not professional actors. I usually know which lines are going to get laughs based on Hunt content or nerd content, and to those of you who came to kickoff, you did not disappoint. Particularly when you laughed at the “pigeonhole principle” line, which helped me win an argument with my wife.

We initially had 26-100 for kickoff due to lack of Kresge, but then Kresge became available when another event canceled. I was actually a little bit opposed to moving, since I thought the bigger fancier space might raise people’s expectations in terms of props/costumes. Several teammates reassured me that Setec has a reputation for low production values (uh, do we?) and I think the finished product was quite good for something we didn’t rehearse in person until the day before. The whole cast was phenomenal, but I want to give an extra shout-out to Steve Peters and Jesse Gelles, who played Santa and Jack respectively and arguably had the heaviest lifting to do character-wise. They showed up for rehearsal with their characters pretty much nailed, and it was great to be able to direct them around landing jokes rather than getting their delivery right in the first place.

It was very unusual for me not to act in the kickoff, but I felt like I’d had enough of a turn in recent years, and lots of people wanted to participate (the part I was most likely to cast myself in, believe it or not, was the Valentine emissary played adorably by Margot Pliska; that character used to be very different, and Margot was a major upgrade). I did throw myself in as the organizational person at the end, where my most important task was to tell people the website URL and keep them busy until 1pm… which, of course, I forgot to do. YOU HAD ONE JOB, DAN.

Those who have written Hunts with me before will likely be more surprised to hear that I did not write the scripts for midgame (the Mayor of Your Birthday Town) or endgame. Those were written by Justin Werfel, who also played the April Fool and did an incredible job with both of those tasks. The evolution of the April Fool as a character was really interesting, in that they were initially conceived as a fairly stereotypical villain, but as we wrote the Hunt, a lot of us noted that we sympathized with the Fool, since they were really only doing what they’re supposed to do. One of my favorite ideas to heighten this sympathy got swept slightly under the rug by the art design once we decided to simplify and represent each town by a single building, with the April Fool living in a rundown shack. In an early conception, every town would have been a collection of buildings, but April Fool’s Day Town would still have been that single little shack with no one else living anywhere nearby.

This is actually the second Setec Hunt in a row where I saw a fairly moving emotional arc develop itself under the story elements, and when we noticed it we committed to making it clearer. In 2017, we designed the midgame and endgame around combining the six characters’ skills, since that’s sort of what you do in a “meta-meta”… This encouraged me to heighten the bickering between MIT characters (and between RPG characters) in the kickoff, and what we ended up with was a fairly sweet story about friends putting aside their differences and working together to save another friend. In 2019, we wrote a whole lot of Hunt before realizing that by making the goal to solve these interholiday problems caused by walls breaking down, maybe our Hunt was a bit more pro-wall than we’d intended? (Ggggggghhhhhh.) We had a really good conversation at our east coast retreat where we conceived the idea that having some interaction between towns, and letting the April Fool do what they do, were both good things in moderation. Justin wrote this really nicely into the endgame skit (which you can see in the video shown at wrap-up), and he and Jesse performed this moment really well. I’m not crying, you’re crying.

I think that sums up most of my thoughts on the skeleton of this thing. next up, I’ll reveal more about how we made the metas happen (which can hopefully help future construction teams decide what to do, what not to do, or maybe both).