Original title: Non-attacking rook placements on crossword grids
Joel Brewster Lewis, Robert Won · September 2026
Think of a crossword puzzle as a board of white squares for words and black squares that divide them. The authors wanted to know how to place chess rooks on the white squares so they do not attack each other. In their setup, a complete placement means placing rooks so that every across word and every down word contains exactly one rook. The researchers analyzed how many valid arrangements can exist on different boards. They proved an upper limit on the total number of placements for any general grid. Next, they looked at sparse grids, where black squares do not touch edge-to-edge. Finally, they proved that specific boards called permutation grids always have at least one valid rook placement.
Imagine you're assigning seats to guests at a banquet with tables arranged in rows and columns, but some tables are broken and can't be used. You need exactly one guest at each table in every row and section, but the broken spots make it hard to avoid empty or overcrowded areas. In this paper, the banquet floor plan is the crossword grid, and finding a seating where each row and column has exactly one guest is like a non-attacking rook placement.
This work helps mathematicians understand patterns in crossword grids and relate them to matrix properties, advancing combinatorial theory. It opens new questions and potential applications in puzzle design and algorithm development.
The work is theoretical and does not address practical algorithms for finding rook placements. Results are limited to specific grid types like sparse and permutation grids. Many questions remain open, as indicated by the conjectures posed.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Non-attacking rook placements on crossword grids — Joel Brewster Lewis, Robert Won
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.