Taiyaki Fabulous Museum of Fish is a 2024 homebrew puzzle game consisting of over 200 nonograms to be completed. Set on Fishcross Island, the player must gain the Power of 7 Fish by allowing them to move into the museum, and ultimately coming face-to-face with Chad the octopus. This TAS aims to solve all 202 nonograms in the game, including all the post-game material, as fast as possible, ending on celebrating Chad's Decade Chadiversary.
Game objectives
- Emulator used: Bizhawk 2.11.1
- Aims for 100% completion
Comments
After my previous attempt at TASing a nonogram game, I've sort of been in a mindset of wanting to TAS puzzle games and this game has always been on my radar so I thought I'd have another go at a nonogram game but with a different setting. Compared to Walfie's Nonograms, this was definitely more difficult to TAS because of how the game allows you to input directions and also because there are 52 more puzzles than in that game. I almost don't know why I kept on with this but just submitting the any% run would feel minimum effort and not worthy of submitting so here we are with a 100% run too.
Puzzles
Unlike Walfie, this game doesn't just let you input one direction on one frame, then another the next, then another the next and allow it to work. It took a second to really get into it but by the end it felt a little like understanding what the game accepts as a language, because even though there are some rules, part of identifying the first possible frame to place the next input was being familiar enough with how the cursor acts and the previous directions. Overall, it came down to a couple of fundamentals:
- Consecutive directional inputs of the same direction have to be 8 frames apart at minimum
- A horizontal and then vertical direction or vice versa can be input 1 frame apart. If third different direction is input after, it needs 2 frame between and A held for 2 frames to make sure the cell gets filled in
- You can do diagonal inputs but the next directional input has to be 12 frames later. This can be reduced to 11 frames if the diagonal isn't filling a cell
- Obey the cursor. It can lag behind so you may have to wait, and it may need some extra time for an A press to go through. It may also decide to just not move on or disobey the above rules so always test out every possible directional placement
I did get the gist of it along the way and I was able to zone out and autopilot a lot of the levels which made it all a lot more bearable, almost fun and addictive at times. That last rule was quite annoying though because it meant having to test way more potential input placements than I was expecting there to be and every time I'd change something earlier on then a lot of levels following it would have a couple of desynced inputs requiring me to go through input by input to identify what the single problem was. And it would always be that one or two inputs would decide they needed an extra frame to be registered or A need to held for an extra from for the cell to be shaded. I thankfully didn't have to go back too many times but every time I did was not fun in the slightest. I only found the puzzles in the first bed after finishing the rest of the game so I spent hours after adding them resyncing the run.
Route
The majority of the run could follow the any% route and pick up the additional puzzles along the way. With the time taken to activate warp points not being an issue, there were actual opportunities to use them now and then. There were a couple of cases where backtracking was needed, such as the NPC on the beach who needed the magic jug and the bed in the Bed and Breakfast needing the Lucid Dream. There's not much to comment on since the route is quite straightforward and backtracking is kept to a minimum, with all the backtracking there is benefitting from warping rather than walking across the map between puzzles. What I will say is that this game never seemed to end. Every time I got to the end of a section, I was hoping I'd be in the final stretch but there'd be another 5 minutes of puzzles to work through. Getting to the end was an amazing feeling.
The obvious points of optimisation would be in finding faster solutions. I'm confident with the number of moves made being optimal and I tried my best to create optimal paths that played with the rules listed above but there are certainly better routes out there that do a better job than I could. Computational solutions would be nice but on top of them already being difficult to handle the scale of some puzzles, it would also have to take into account how the game works and that would be a lot to work through.
The run ends on talking to Chad after finishing his final puzzle but I couldn't decide if I wanted to end it after exiting the final nonogram instead because that still technically counts as 100% of the game being completed. It depends on whether talking to Chad is considered reaching the true ending of the game and I'm happy with either outcome.
Challenger: Claiming for judging.