That sounds like a fascinating experiment. I can picture each clue as a node, each hypothesis as a branch, and the final reveal as the resolution node. If you can capture the subtlety of the red herrings and the interplay of motive and opportunity, it would be a great way to see how narrative logic translates into computational logic. Just be careful not to strip the atmosphere—those classic whodunits thrive on the tension between what’s obvious and what’s hidden. Good luck with the coding!