Hi, that's some high praise, thanks!

I was afraid the difficulty would be out of whack: we finished writing the cases late in the night just before the deadline, so we had no time for human testing.

I would like to return to the game after the jam ends; I will definitely need some distance to see problems. I actually thought it might be too easy, because, you know, the cases I wrote were clear to me ;-). The first one (the mansion) was meant to be mostly an extended tutorial, so if you want to try again, I would recommend that one. I wrote the last one (power plant)  with a specific goal to be hard, even making sure there's a red herring solution that's gramatically correct.

We had some ideas how to make the game more playable. One of the more important was to split the word list to several categories (verbs, nouns, adverbs) and allow selection of the right one only: this should cut the confusion and make the UI more pleasant. I am also considering getting rid of the 84x48px resolution, just to allow using a more legible font. We aimed for maximum information density, but the trade-off is we just didn't have enough pixels to draw nice glyphs.

I am really happy to hear the descriptions are good. We both are playing pen-and-paper RPGs as game masters, it seems the years did pay off :-).

If you want spoilers, the GitHub repository linked at the game page contains complete source code, including the solutions. Open the 

the-grind-2025/scenario/<name>.py

directory and search for DeductionScreen - each one defines a list of words (in the correct order) that are the solution.