As the captain of a team of superb puzzle-solvers who theoretically do not care whether or not they win the annual MIT Mystery Hunt as long as everybody has a good time — but who, let’s face it, kinda do want to win one of these days — it is difficult not to be of two very conflicting minds about the chart you see here.
The chart was shown at this year’s wrap-up by Setec Astronomy, the team that had spent the previous year creating nearly 200 incredibly challenging puzzles and activities, which they had parceled out to thousands of solvers in Boston and around the world over the course of the weekend. Setec had earned this responsibility by winning the 2018 Mystery Hunt; the winner of their Hunt will, naturally, go on to create a Hunt of their own for 2020.
When the chart appeared on the screen, nobody on stage said anything at all. They simply displayed it and stood back. Within moments, its meaning became clear, and the groans began. Palindrome (my team) and the eventual winners, Left Out, had been neck and neck through the first thirty hours of the Hunt. And then, lo and behold, Palindrome started to pull away! We eventually would reach a thirteen puzzle lead over our closest competition… at which point we hit what I shall now call the Palindrome Plateau.
We’ve plateaued before, heaven knows, but rarely this dramatically. (Or anyway, I’ve never seen it so dramatically depicted.) Ten hours — ten solid hours — of being stuck on every single puzzle we had left. And in that same ten-hour period, Left Out never faltered — they maintained their steady pace, overcame us, and seized the day.
It’s tough to be 100% blasé about that.
And yet, my main takeaway from this year’s Hunt was how much fun I had. The structure of it was ingenious. Every Hunt has a theme, and this year we were treated to a trip through the Holiday Forest, sort of like the one Jack Skellington wanders into in A Nightmare Before Christmas. Now, normally, a Hunt is divided into specific sections, each with its own set of puzzles. Solve the puzzles, and you get a set of answers. Tie those answers together in some magical way, and you’ll have solved a metapuzzle — a major leap forward in your team’s progress. In a typical year, given a holiday-themed Hunt, you’d expect to have a bunch of Christmas puzzles and then the Christmas metapuzzle, and then Halloween puzzles followed by a Halloween metapuzzle, and so on. But the Setec folks are far more clever than that. Yes, we had Christmas puzzles and Halloween puzzles, and puzzles for plenty of other holidays besides — but to solve one of the important metapuzzles, you had to take some answers from one round and some answers from another, and you didn’t know which answers you needed and which ones you had to reserve for later. For die-hard puzzlers who have seen their share of metapuzzles, this was a delightfully crafty curveball, one that raised the trickiness factor just enough without pushing things into the red zone of impossible.
Except I didn’t solve any of the metapuzzles this year. And in fact… as I look through the long, long list of puzzles served up in the 2019 Hunt, it would seem that I didn’t solve any of those, either. Oh, I kept a hand in, don’t get me wrong. I entered information on spreadsheets. I researched things. I nailed down a few cryptic clues here and there, and my brother Dan and I even solved an entire variety crossword and discovered its secret message, which proved valuable to solving the larger puzzle that it was a part of. But at no point was I ever the person who stood up and said “The answer to this puzzle is WHATEVER!”
So why on earth do I go to this event where the puzzles are so hard that some years, like this year, I can float through the entire weekend without solving any of them?
The team, folks. That’s why. The team. I enjoy solving puzzles all by myself, heaven knows, and I enjoy going to events where I’ll co-solve puzzles with three or four friends, but there is nothing like being a part of a Mystery Hunt team. It’s the difference between a solo violin player, a lovely string quartet… and a symphony orchestra so brilliant, their music blows you out of the room.
I’m sure most solid Mystery Hunt teams feel like that to some degree, but I can only talk about my own. Palindrome is an ever-evolving collective of teachers, engineers, programmers, scientists, writers, and who-knows-what-else. We get people who join the team for a single year, just for the experience of it, and we have folks (like me) who have been on the team for decades. As you might expect of people drawn to a gigantic puzzle-solving competition, many of them are smart, and some of them are brilliant. We set up shop online and in a couple of MIT classrooms, and together we assemble a human machine that, it sometimes seems, can solve just about anything. I love being a gear in that machine.
You might think that if you cram dozens of people into a small classroom in the dead of winter and give them puzzles in languages they don’t speak, they might, somewhere around 3 or 4 in the morning, get a little cranky. This doesn’t seem to happen on Palindrome. I’m not even sure why. It’s not like we have a screening process. We’re an open team. Anybody can join us. (Unless we get too big. Then, alas, we have to shut the door.) But somehow each year we wind up with a group that strikes just the right balance of competitive and carefree. We work hard, and God knows we have periods of burning frustration, but it never stops being fun.
And so if you gave me advance warning that in 2020, Palindrome will once again screw up a massive lead and drop down to second place, it wouldn’t stop me for a moment from packing my bags next January.
On the other hand: But seriously, what the hell happened to us? A ten-hour period with no correct solutions? I knew we were stuck on things, of course, but I didn’t realize how long we spent spinning our wheels. Time gets kinda slippery at the Mystery Hunt.
Looking at the puzzles we had open during that time, our weaknesses become clear. They are very human weaknesses. To wit:
Information Overload. Several puzzles required a key insight before you could really start rolling on them. Standardized Mess is a good example. We had that baby open and unsolved for sixteen hours, and there is simply no excuse for it. It’s a fake multiple-choice exam, followed by the seven Scantron forms of various test-takers. Their grades appear on the forms, but for several students, their real grades don’t match the percentage shown. Why? And what sort of hint is tucked into the (as we call it) “flavortext” — that bit about teachers needing to grade these exams quickly. What does that mean?
We filled many different worksheets in a Google Spreadsheet with gobs of information. The true answers; each student’s false answer; if you assign a 0 to each wrong answer and 1 to each right answer, what that would look like in binary, or in hex, or in ASCII.
There was a note on each Scantron form, and the first letters of these notes spelled the word COMBINE. So let’s combine the forms! Let’s combine the grades! Let’s combine the answers! Let’s combine things no reasonable person would expect us to combine!
After sixteen hours, our spreadsheet for this puzzle was an imposing wall of data. Because entering data feels useful: You’re doing something, not just staring, and besides, if you do the right thing — if you cross-reference two pieces of information in just the right way — you never know, the answer might leap out at you!
And yet what was needed here was a fairly simple insight: Why did the puzzlemaker use these questions to generate the right and wrong answers we used to fill our spreadsheets? The light bulb, at long last, flickered to life. Go take a look at the 21 questions and see if you can spot the common element. Go on, I’ll wait. Believe it or not, it’s not really all that hard.
Did you try? Okay. It turns out that every question — every single question on the test — makes a reference to the phrase “upside-down,” or to something famously known for being upside-down. (Houdini, bats, Tobey Maguire in his first Spider-Man movie.) And now it all becomes clear: The teachers graded some of these tests a little too quickly, feeding them into the machine the wrong way. They’re upside-down. If you orient them correctly, and then lay the seven forms on top of one another, the darkened bubbles form the letters in the word FLUNKED, and that’s the answer to this puzzle. We aimed a fire hose of information at our faces, and that kept us from seeing it.
Miscommunication. Palindrome is a large team. In 2017, we peaked at a rather insane 150 people. This year, we had about 100 — not quite as staggering, but still pretty sizable.
In a group that large, it’s pretty easy to miss it when somebody says the exact right thing.
One of the metapuzzles I alluded to earlier combined answers from the Indian holiday of Holi and the Boston holiday of Patriot’s Day. The instructions for the metapuzzle were obscure: We were told we had to start on “the RED route,” without being told what that meant. Okay, then, what are some possibilities? The Red Line of Boston’s T made a lot of sense, but that went nowhere when you tried to apply the puzzle’s other instructions. So maybe another subway system, somewhere else? Tokyo? New York? Maybe a board game like Ticket To Ride?
We discussed this on and off for hours. Along the way, several people said the right answer — Olympic rings — without being heard. It’s not like they knew their suggestion to be correct and trumpeted it with euphoria. No, it was simply a suggestion: Something to consider. We didn’t, until we did, and then got to hear the agony of several teammates smacking their foreheads and saying “I suggested that hours ago!”
Institutional Memory, or Lack Thereof. An excellent point one of my teammates made in our post-mortem is that a small group of solvers could easily waste their time going down the same wrongheaded avenue on a stuck puzzle as a previous group of solvers, because nowhere was the basic idea behind their approach recorded, and the result of that approach revealed. (That data was probably somewhere on the puzzle’s Google spreadsheet, but without a more coherent organizational system, good luck finding it — see Information Overload, above.) One of the key items on my checklist for next year some kind of brainstorming space for each puzzle, so that solvers can see what’s already been tried.
So, yeah. I’d say we learned some valuable lessons this year — the hard way, which is how most valuable lessons are learned. I shake my head in disbelief every time I look at that chart. But the Palindrome Plateau doesn’t shave more than an atom or two off of the sheer enjoyment of Mystery Hunt weekend. Furthermore, our descent from the lead means that Left Out, a stellar team featuring puzzlemakers I admire very much, will be serving up next year’s Hunt. I can’t wait to see what they have for us.