I’ve been doing a bunch of puzzles recently. By “puzzles” I mean, in the Mystery Hunt tradition; it’s a weird world that’s not as well defined as “doing jigsaw puzzles” or “doing sudokus”. In this post, I hope to:
- help define what these puzzles are
- highlight/popularize some terms that help communication around puzzles
- give some conceptual handles that will help people describe to themselves what they’re doing when they’re doing puzzles
the schema
Most puzzles involve something like this series of phases:
- The Setup
- The Break-in
- The Work
- Extraction
These are not always once each and linear; usually it’s something like Setup Work Break-in Work Extraction or Setup Break-in Work Extraction Work Extraction. But they tend to go in about this order.
the setup
“Getting stuff into a form you can work with it.” Usually you’ll get a puzzle as a PDF or website, which is usually not a great form: you can’t write on it or collaborate on it. If you’re doing these online, a Google sheet is usually the best form.
Sometimes this step is simple: copy/paste the puzzle into a Google sheet. But often the puzzle doesn’t copy/paste nicely. Or it has graphics or an interactive component, so you have to try to copy the relevant info in, while losing as little as possible. (Maybe this involves clicking around an interactive thing and extracting every bit of info you can.) Or sometimes the puzzle involves an origami 3d shape or something, and the Setup is printing it out, cutting, and taping it into place.
Things to know about the setup:
- it’s usually pretty easy to do, so if you’re braindead or overwhelmed, it’s helpful to your team if you set up some puzzles
- make sure you and a teammate aren’t both setting up the same puzzle. Having a shared Google doc helps; maybe also a chat channel. If you are, clear it up as soon as possible.
- don’t skip it! It can be tempting to just do the puzzle, but with a few exceptions, you probably won’t be able to just breeze through the puzzle, you’ll need another set of eyes, and if it’s not already on a Google sheet, that’s hard.
the break-in
“Figuring out what you need to do in this puzzle.” Like I said, these aren’t as simple as “put the jigsaw pieces together” or “solve the sudoku.” This takes a bit of magical thinking.
Things to know about the break-in:
- you’ll usually know it when you get it. A couple pieces will click together in a way that relates to the theme or flavor text, or something will be in alphabetical order or spell out a message.
- you often can postpone it until after you Do The Work. Sometimes some of the work will be obvious, and some will be non-obvious; just do the obvious part then
- if you haven’t done the break-in, though, be careful about going off too far. If you find yourself making a bunch of leaps and nothing’s clicking, maybe don’t add it to the shared doc so you don’t create too much noise
- this gets easier as you do more puzzles! You see common patterns that usually result in break-ins
- this feels amazing when you get it
the work
“The straightforward part.” After you’ve broken in, you know what to do, now just do it a bunch. “they’re crossword clues but the answers all rhyme with a Sesame Street character” or “it’s a logic puzzle but you have to negate all the odd numbered clues” or something.
Things to know about The Work:
- it’s usually pretty straightforward. If you’re solving with a newbie, try to get them involved here; it’s often easy to make a little progress. Likewise, if your brain is fried, you can often at least Do The Work on some puzzles.
- it does feel pretty good.
- if you have to identify images, you don’t actually have to know what they are; Google image search or Tineye can help you figure out what the picture is of!
- also, for crossword-type clues, google a lot.
- a decent thesaurus can be a big help too.
- if you can reduce the puzzle to a known type, solving it can be automated. For example, cryptograms, common ciphers like caesar, sudoku.
extraction
“Finding the word or short phrase to put into the box.” Often after you break in and do the work, you’ll find yourself staring at a “completed” puzzle. But the answer is almost always a word or short phrase. So you have to figure out how to distill the solved puzzle into a word.
“How to extract” could be a full college course. But some tips, and other thoughts about extraction:
- Make A Big Grid. Once you have a bunch of clues and answers, the extraction is often “sort by X, index by Y”; Google Sheets’s “Create A Filter” feature helps you do this once you have a grid.
- look for indexes. (numbers that could tell you “use the nth letter of each word”)
- check diagonals
- check thematic things (if the puzzle is called Famous Firsts, maybe the first letter of each answer?)
- if you find “answ” or “sol”, the solution might be something like “ANSWER BLUEBERRY” which just means “the answer is blueberry”.
- try to “do it again”; if the puzzle is clues that rhyme with Sesame Street characters, see if this could be cluing a word that rhymes with a Sesame Street character.
- you don’t have to Do all the Work! Usually 80% of the Work is easy, and 20% is hard (last couple clues, obscure references, etc). Skip those and try to find a word that would fit even with a couple missing letters.
- also, try to be a little flexible; you might have gotten some clues wrong along the way so it might not quite work out right.
- use nutrimatic or onelook to find words that fit certain criteria
- it’s usually not that out-there; if you find yourself making more than one or two leaps, scale it back
- more eyes often help here. When you bring in someone else, you have to give them some context (how you solved the puzzle). But don’t tell them all the dead ends you are currently on; you don’t want them to get stuck thinking the same wrong thing as you.
- extracting correctly also feels amazing
an example
This puzzle, Double Jeopardy, from this year’s Puzzle Boat (spoilers for this puzzle btw), works like so:
- the setup: we copied it into a Google sheet. It was a little tricky to format because it didn’t paste straightforwardly, we had to wrangle it into the right rows and columns. (We wanted to keep the format in case it mattered; Jeopardy categories often matter.)
- the work: it’s a lot of clues. We started solving the clues, often by googling (e.g. “Vander’s last name in the anime One Piece” -> DECKEN)
- the break in: we noticed that some of the clues have some part bolded. E.g. “He voices a large feline bus in My Neighbor Totoro”. After noticing that the puzzle is called “Double Jeopardy”, and the flavor text is “Double What?”, we noticed that those bolded parts refer to a common “Double ____” phrase. In the “bus” case, this is “DOUBLE DECKER.” Also, these “double” answers are often one letter off from a regular clue answer (decker/decken).
- the work: now, in addition to answering the questions, we also had to fill in the “double” clues, finding answers that were one off from another answer. We continued doing that until we answered all the Jeopardy questions and all the “double” clues. Each double clue (15 of them) matched to a Jeopardy clue that didn’t have a bold text, so that felt right.
- extraction: now we had a big grid:
bold “double” clue" | “double” answer | Jeopardy answer that contained the bold “double” clue | other Jeopardy answer it matched with | letter that was in “double” answer only | letter that was in Jeopardy clue only |
---|---|---|---|---|---|
bus | DECKER | Naoki Tatsuta | DECKEN | R | N |
We filled this out for all the clues, fiddled with it for a bit, and found that if you sort by the “Jeopardy answer that contained the bold double clue”, and read off the letters that are in the “letter that was in double answer only”, it spells “IT’S MAXIMAL PBIRS.” “Maximal Pbirs” isn’t a thing, but “Maximal Pairs” sounds like a reasonable phrase, so we just figured that we got one clue wrong and submitted Maximal Pairs as the answer.
love the setup and the work
Were you ever in a D&D campaign where you all wanted to be a wizard? My last one was like that. We had a wizard, warlock, bard, druid, and rogue. Especially if you hang out with Smart People, everyone wants to be a wizard. Our DM had to roll us a cleric and a fighter out of kindness so we didn’t get demolished.
Puzzle hunting can be kind of like that. If you remember key parts of a puzzle hunt, you probably remember when you, or someone else, made some impressive logical leap. These are probably break-ins or extractions. These are fun. They’re also difficult. You feel like a wizard!
But you can’t be a wizard without the setup and the work. Sometimes these are even more valuable: the skill required in a puzzle can be just cranking through crossword clues quickly or doing some Google sheets magic to make the puzzle even doable. (there’s even a type of puzzle called a Duck Konundrum that is basically all setup; just really meticulous following-directions.)
So if you’re feeling like a fool because you can’t figure out the break-in, or your brain is fried after a long puzzle, or you’re new and you can’t figure out how you’ll be able to hang with the wizards, do some setup, do some work. Setup and Work are the fighters and clerics of your D&D party; no less valuable than the wizards.
And appreciate others who are doing setup and work. Thank them for it; notice when they do it super well. Others will notice when you do it and will always want you on the team.
blog 2024 2023 2022 2021 2020 2019 2018 2017 2016 2015 2014 2013 2012 2011 2010