MIT Mystery Hunt 2026 just happened and I led the design for one of the early dimension rounds: the Land of No Name, which we internally just referred to as the "alphabet round". This post will go into some behind the scenes details for the round's design. Since it would be tricky to coordinate writing and testsolving for the round, I put together a round design document for authors and editors to refer to. Here are the unedited contents of our design document. Difficulty ratings were out of 6, where a difficulty of 1 denoted fish puzzles or easier and a difficulty of 6 denoted puzzles that were too hard, even for Mystery Hunt. As a general rule of thumb, most of our dimension round puzzles aimed to be difficulty 4, with some exceptions made for difficulty 5 puzzles. # Alphabet Round Structure ## Premise This round is a dimensional rift where the alphabet is missing! As you solve puzzles, you recover letters in the alphabet and you can see those letters in all the other puzzles now. There are exactly 26 puzzles, one for each letter of the alphabet. ## Testsolving Guidelines For your assigned answer, look at Alphabet Round Difficulty Guidance (see table below) and see whether or not the letters required for your puzzle is “most”, “some”, or “none”. Conduct at least 1 testsolve of unlocked letters on the low-end or high-end of your letter bracket based on the table below. The puzzle should be solvable on the low-end of your puzzle bracket and you should use the high-end testsolve to get a sense of how much easier the puzzle gets with letters unlocked. Also, make sure not to include the letter your puzzle unlocks in the testsolve | Letter Bracket | Letters Unlocked (low-end) | Letters Unlocked (high-end) | | -------------- | -------------------------- | --------------------------- | | most | BDFGHIJKLMOQTUVWYZ | <all letters> | | some | GHIJMOQZ | BDFGHIJKLMOQTUVWYZ | | none | <no letters> | GHIJMOQZ | ## Details 1. **Do the missing letters affect the puzzle titles?** Probably yes. I think it’s more interesting if the titles are also obscured. 2. **Can solvers see which letter is unlocked from unsolved puzzles?** Probably not. Thematically, I think it’s more appropriate not to see a letter before unlocking. 3. **Are the letters unlocked and feeder answers tied together?** Yes, when writing puzzles for this round, your puzzle will be assigned a specific answer as well as a specific letter it unlocks in the round. That said, you don’t need to worry about which letter your puzzle unlocks when writing. 4. **Are non-alphabetical letters affected?** No. Allowing emojis and numbers seems like it would be keeping in theme, since we had paintings and stuff before language. 5. **Do the missing letters affect answer submission?** No. There are a couple possible implementations where answer submission is affected, but with the gimmick + scale of this round, it’ll lead to a better solving experience if solvers can just submit answers normally. 6. **What difficulty should feeders be for this round?** We’re aiming for a spread of difficulties, but as a rule of thumb, we’d like to target no letters puzzles to be on the harder end (3-4 difficulty), some letters missing puzzles to be on the medium end (2-3 difficulty), and nearly all letters puzzles to be on the easy end (1-2 difficulty). 7. **What do puzzles look like when letters are missing?** Letters that are missing are replaced with question marks or a custom scrawl image we put into the font for this round. This way, solvers still can get enumerations. Puzzles with interactivity, images, or audio will need to get cleared with editors and tech to figure out how it will incorporate missing letters. 8. **What is the unlock structure like?** The current idea is that puzzles unlock at the same time and it would be part of the challenge in figuring out which puzzles are actually solvable, though this is subject to change. 9. **Do feeders need to be written with the round gimmick in mind?** Puzzle authors don’t need to keep the round gimmick in mind, but editors should consult to estimate if the puzzle is solvable with roughly no letters (e.g. purely visual puzzles), some letters missing (e.g. crossword clues), or nearly all letters (e.g. a dropquote puzzle). Editors will work to try to have a good balance of the 3 classes. 10. **How will testsolving work?** Initial testsolves should be done with all the letters provided, but it doesn’t count as a clean testsolve. After initial testsolving, based on the previous classification, the two clean testsolves will be done with either 0 letters, a random selection of ~8ish letters, or a random selection of ~20ish letters. A big testsolve will also be needed for the round to help tune unlocks and individual puzzles. 11. **How much tech work is required for this?** A decent amount. Basically, the puzzle contents need to have all the text replaced server-side depending on the team’s current solve state. Probably we’ll also want an LRU cache of the replaced text, keyed on the solve state bitmask? If there are puzzles that want to incorporate images and audio with text in them, then we’ll need to figure those out ad hoc. Likely, images should try to use SVGs so there’s a possibility of rewriting them before serving. # Alphabet Round Difficulty Guidance ***Note**: The "Can Omit" column corresponded to whether or not we thought some team could solve the meta with just that letter missing.* | Unlocked Letter | Can Omit | Letters Required <br>(no = 3-4, some = 2-3, most = 1-2) | Additional Notes | | --------------- | -------- | -------------------------------------------------------- | -------------------- | | A | | most | Ensure difficulty 1 | | B | | some | | | C | | most | | | D | | some | | | E | | most | Ensure difficulty 1 | | F | | some | | | G | yes | no | | | H | yes | no | | | I | | no | Ensure difficulty 3- | | J | yes | no | | | K | | most | | | L | | some | | | M | yes | no | | | N | | most | | | O | | no | Ensure difficulty 3- | | P | | most | | | Q | yes | no | | | R | | most | | | S | | most | | | T | | some | | | U | | some | | | V | | some | | | W | | some | | | X | | most | | | Y | | some | | | Z | yes | no | | # Post Mystery Hunt Design Thoughts I think there is an extremely large design space available to puzzle hunts and hope that this round managed to explore an interesting subset of that space. When fleshing out the details for this round, the question at the foremost of my mind was "What do folks like about puzzle hunt-style puzzles and how does this round contribute to that fun in an additive way?". Certainly the round idea is compelling because it's never been done before, but it may have not been done before because it wouldn't work well in practice. It was difficult to gauge whether or not the structure would test well prior to having all our answers assigned and puzzles written, so I put some amount of thought into the difficulty curve, the letter unlock mapping, and whether or not all puzzles should be unlocked at once. I was also involved as a round-level editor to help strike a balance in puzzle content and make puzzles approachable at their intended position in the round. Based on how we saw the round played during Mystery Hunt, I think we managed to achieve something that was uniquely fun. With that knowledge, I want to end on two high-level reasons for **why** I think solvers found the round gimmick fun: * **Shortcutting**: There is a long history of solvers delighting in solving meta with partial information and then backsolving feeder answers. On a smaller scale, many solvers also enjoy being able to Wheel of Fortune the final answer or cluephrase from having only forward-solved a portion of the puzzle's content. I think at its core, this round builds on that kind of enjoyment and pushes it to a structural level. * **Progress**: I think solvers undeniably enjoy making forward progress in the hunt. Typically, this is measured in puzzle solves for a round, but our decision to release all 26 puzzles at once allows solvers to see another metric of progress: namely the solvability of the remaining puzzles. And just as unlocking new puzzles in a round can provide additional momentum to a solving team, I believe the incremental "unlocking" nature of the round together with our intended difficulty curve helped establish solving momentum.