Hall of Mirrors

Written by Jacqui Fashimpaur, Liam Thomas, and Samuel Yeom


We are given a grid full of mirrors with pairs of symbols around the edge. We infer that we must use the mirrors in the grid to reflect one symbol in each pair to the other. Finding these correct paths will help us determine which mirrors are real and which are fake. Real mirrors represent a ninety-degree turn in a reflection's path, while fake mirrors let the "light" pass straight through. If we try to solve this purely as a logic puzzle, we'll find that it's underconstrained.

This is where the feeder answers come in. The symbols around the edge match those associated with entangled puzzles on the map. We realize that we are not just creating paths between symbols, but between pairs of answers. The flavor text says the mirrors "distort everything they reflect": each type of mirror represents a specific word transformation. The entangled answer pairs represent words pre-transformation (Matt’s) and-post transformation (Emma’s), entering and exiting the grid at the points marked with their corresponding symbol. Using a combination of logic and word reasoning, we determine what word transformation each mirror represents, as well as which mirrors must be real for the reflections to work correctly.

Change all vowels
Semantically double
Take opposite
Go two steps back in a sequence
Drop one letter
Take homophone

These are the paths between feeder answers:

And these are the corresponding chains of word transformations:


After this, we recognize that the glowing words “START MAJOR META HELL” are aligned with the grid as well. We use these as inputs for word transformations, bouncing off of the mirrors we’ve already found.


Reading top-to-bottom, we get our final answer: FOURTEEN YEARS BAD LUCK - since breaking a mirror is said to bring seven years of bad luck and both of the twins will need to break one to get home!

Author's Notes

This puzzle was hard to construct! Every new iteration required starting from scratch. We'd begin by coming up with 5-6 possible mirrors that could get us from any remotely reasonable phrase to FOURTEEN YEARS BAD LUCK. Before "START MAJOR META HELL" it was "WILD MYSTERY MIRROR MAZE" and "SOLVE FAKE MIRROR MYSTERY". It was important to us that the extraction transformations be extremely clean, since solvers would have to go through them one-way with no confirmation besides the other words in the answer.

After we had those transformation paths decided, we'd place the mirrors in the grid such that the final extraction bounces all worked out. Then we'd look at what feeder answer paths happened to exist, making minor adjustments and adding a few extra mirrors if needed. Finally we'd brainstorm answer pairs for every slot that had good enough intermediate words at every step.

We did this four times before we landed on the final construction. One transformation path we wished we could use that just didn't fit in this grid is:


Most feeder slots only had one decent pair of answers that we could find, though some had a few options. PIE and RHO used to instead be:


But that had a few issues. Just be glad we didn't go with: