by Chris Yao
How to Solve
There are 6 possible combinations of entrance & exit locations for the grid (and its recursive instances): N+S, E+W, N+E, N+W, S+E, S+W. Logically working out each of these (a common way to start them is, after picking the entrance/exit gates, to observe the row labeled "7" forcing the path through the upper left area, and then find limited possibilities for the leftmost column and bottommost row) without considering their embedded grids results in:
N+S: no solution E+W: no solution N+E: 3 solutions (that use the same squares in different orders) N+W: 1 solution that goes through no embedded grids S+E: 1 solution that goes through an embedded grid as N+W S+W: 1 solution that goes through an embedded grid as S+E
When factoring in embedded grid possibilities, 2 of the 3 N+E solutions are eliminated because they depend on passing through an embedded grid as N+S or E+W. That leaves one N+E solution that goes through embedded grids as S+W and S+E. By elimination, this N+E solution also must be used for the root grid.
The result goes through the grid solutions (ordered by when they are entered): N+E, S+E, N+W, S+W, S+E, N+W. Shown here:
Note that grids 5 and 6 are just the reverse paths of grids 2 and 3. The above grids also have some steps highlighted corresponding to the exit instructions. When these highlights are assembled on the grid (and colored as white), it results in:
This represents the Pirate Flag emoji: 🏴☠️. (Note that some platforms, that do not support showing this as a single emoji, might instead display it as the Waving Black Flag emoji followed by the Skull and Crossbones emoji.)
The Skull and Crossbones emoji by itself is also a reasonable interpretation, and submitting it returned a hint that such did not have the appropriate (black) background. During the hunt, pretty much every solving team submitted the Skull and Crossbones emoji before submitting the correct Pirate Flag emoji.
The original idea from brainstorming was to incorporate fractal maze mechanics into a grid logic puzzle, with 2-3 grids that recursively contained themselves and one another. Each grid would support ~2 different internal paths to be used in the final full path.
It took a fair amount of effort to hash out the first grid to support two decent exit sets/paths simultaneously. Once done though, it felt like that grid formed a nice network of separate mini-logic areas, so rather than moving on to creating the second grid, I checked out what would happen if I tried to solve the finished grid using another, randomly chosen, exit set. That random trial nearly succeeded in solving to a unique path! It was close enough that several tweaks allowed for the integration of the third solvable exit set, without needing nearly as much effort as it took for the second set. At that point, the clear route was to abandon the notion of multiple grids, and flush out the rest of the paths within this single grid. Each additional exit set took less and less effort to integrate support for, all the way through to the point where each of the 6 possible exit sets were hand-solvable simultaneously. This setup provided a lot of flexibility for designing the final fractal structure!
The spots to place the recursive studio rooms were then determined such that the full path would go through 4 of the 6 exit sets (some of them multiple times), which seemed like a good amount. For the 2 exit sets that weren't part of the final path, it was strongly considered to set them up to be effective dead-ends in terms of the fractal structure, by having their local solutions recurse into themselves or into each other. However, I wanted to avoid leading people into solving grids all the way to completion only to find that they ultimately weren't used, so I ended up adding some walls such that those 2 exit sets would reach logical dead-ends rather than solvable paths.