I can date the genesis of Monky Puzzle precisely. It was the day this Futility Closet post was published.

It was clear to me that Cistercian Numerals were both obscure and interesting, which is an ideal combination for Magpie crosswords. Also, they could be made out of crossword bars. I just had to work out how to make a puzzle around this.

When I noticed that 1+2+6 = 9 worked as a crossword bar superposition, I knew that was what I should use.

Because of how the numeral system works, this can be done with four different four-digit numbers adding to 9999. My original idea was to clue a carte blanche grid that would reveal the four four-digit numbers in the four quadrants, and require solvers to do the calculation and then introduce new bars, making new real words, in the centre. This had a load of two-letter answers and some tricky constraints for the central long answers that had to be divided into sub-words by the new bars. I couldn’t make it work.

My next idea was that to make the final grid emerge late in the solving process, I could clue the whole thing as jumbles, with the promise of an all-real-word final reveal. I did a bit of consulting about whether it can ever really be fun for solvers to enter all answers as jumbles. It seems to give the setter a rather unfair advantage and the solver a lot of tedious cold-solving. Encouraged by the response that it is OK if it’s appropriately handled, I looked for ways to give solvers the best chance I could.

Everyone loves a bit of symmetry, and I was able to make the starting and finishing grid symmetrical. Adding the rule that no bar appears in both grids gave some useful help, and actually made grid creation more interesting, so I was pleased with that idea. It was helpful that the number of down clues for the original grid matched the number of non-thematic across answers in the final grid. So I could include some vital extra assistance by adding final-grid clues as extra words. I was lucky to be able to do this with just one word per clue. It was worth making some of the clues a bit loose to avoid having to say ‘one or two words’ in the preamble, which introduces the sort of ambiguity that I don’t love when solving. Giving clues for the final grid meant that I didn’t have to worry about whether the original grid was underdetermined, which it probably was.

The title was so obviously right for the puzzle that I had to use it. The fact that it was useful for finding CISTERCIAN was a bonus. I did my best to make the clues reasonably interesting while keeping them cold-solvable. I avoided plain anagrams, because I don’t like to be given an anagram when the answer is jumbled – it undermines solving the clue.

The part of the preamble about ‘parts of circles’ was added only at the last minute to confirm ARCS. I was simply trying to clear up loose ends, and those letters were otherwise left hanging. I wish they were something like MONK but filling the grid was difficult enough (for my computer) without extra constraints. I used Quinapalus’s Qxw which has a ‘free light’ feature that allowed me to define the constraints for both grids at once. My first attempt contained no Ws (required for the Treasure Hunt), so I just kept trying with W in different locations until something worked. I did my best to avoid plurals, but in the end the grid is necessarily full of low-value Scrabble letters, and I had to settle for what I could get.

I know that Cistercian numerals aren’t really a cipher, but CISTERCIAN CIPHER was much easier to put into the grid, and it’s a common enough name for the system that I felt I would get away with it, which I think I did.

Because I felt that I hadn’t really succeeded in my goal of requiring solvers to do the Cistercian addition to prove correctness (it was possible to get it right just by labelling the central shape), and I’d made an all-jumbles entry requirement, I was prepared for solvers to be underwhelmed. I was pleasantly surprised by the reaction. It turns out that solvers just need symmetry, interesting bar patterns, and a weird number system and they are happy.

Posted on Wednesday, January 14th, 2026 at 12:00 pm under Setting the scene.
Leave a response