Listen With Others

I find that one of the joys of teaching children chess is that the children often provide me with inspiration. Thinking how to explain something in a way that connects with them often results in new ways of looking at things. Similarly in the context of my inquisitive daughter. How do you explain how an “&lit.” clue works to a 7-year-old with a lack of vocabulary? The clue “Racetrack is the place to see this going round repeatedly (3)” was the result in my case.

A further question about how setters get ideas for and set a Listener puzzle led me to pick up her younger brother’s picture book and flick through. I stopped at the page with the starfish and the idea of “fish star” suggested itself to me – “PISCES in 5 arms radiating out from centre” according to my first written note. This end game survived throughout the numerous iterations of the puzzle idea while other features came and went. Some of these features were specific nods to the inspiration. References to the “star fish” Peach from Disney’s Finding Nemo featured in some of them, including in the test-solver version which had an additional message to find “Colour Peach” (thereby constraining all clues, but a step which probably would not add enough to the puzzle). A hidden reference to a character in one of my daughter’s favourite books was disrupted by a small change in clue order. A mention of my daughter in the final grid remains.

The grid went through many fundamental rewrites. The first attempt was a conventional one using squares, but after a few iterations, I quickly decided the theme would be enriched enormously by using a pentagonal grid. I printed a grid from the learning resources website by Jo Edkins at http://www.gwydir.demon.co.uk/jo/tess/pent.htm and drew in constraints with a central P, 5 Is and 100 cells in a diamond shape for a grid fill by hand.

The next stage was to produce a grid fill. I had a wide range of thematic ideas, which were mostly scoped out in the many hours sitting waiting as a chess parent while my children played in junior chess tournaments. One for which I produced an almost complete grid fill was the idea of 5 thematic groups linking to make words or phrases STAR – FISH – DIVE – BOMB – SHELL (…STAR etc). This was more time consuming as it was important that while producing the fill I made sure there were enough non-thematic words available to provide checking and a unique grid solution. I decided that I could do better, instead implying synonyms for starfish. My next completed grid fill had 5 of each STARS, ASTEROIDS, FISH, synonyms for HAND (implying five fingers) and shades of RED (herrings). I had constrained myself at this point to use only entries in Chambers. I was still not happy with this, however, as the items in the sets were somewhat arbitrary. Instead I decided that I should use very specific sets. After some research, and trial and error as to what would fit, I finalised the list as 5 brightest stars, 5 brightest asteroids, 5 fish most commonly eaten with chips (in UK), 5 fingers on a hand and 5 seas of the “port of 5 seas” (Moscow). I found that to squeeze all of the items in from these very specific sets required a slight increase in grid size.

My next step was to prove that it had a unique solution which could be determined from the entry answers and to make sure that those entries were adequately checked by other entries. At this point I had in mind that only the first squares of entries would be numbered. I stalled somewhat in this. I then put it aside altogether as I could not see a way forward for checking the solution, nor could I face doing yet another grid fill to avoid the ambiguity I spotted in one of the PISCES. It was when the first of the excellent SQUARE ROUTES® puzzles was published that I thought there may be a way forward, and I shared the basic concept with the authors of that puzzle, hoping that an extension to their setting software might do the trick. The answer was it would unfortunately require far too much computing time to extend from 5×5 square to over 100 cells using pentagons, but I was encouraged to continue looking for another way to make it work.

The idea of using initial and final squares as enumerations came to me, which allowed my daughter and I to check what words were needed in addition to the thematic ones to make sure that the grid was uniquely fillable. I dropped several words at this point to leave just 36 entries (all but 2 required to make it unambiguous). While checking that this worked, the idea of how to resolve the ambiguity in PISCES also occurred to me although it involved a very significant constraint on the clueing, it followed the thematic theme of 5 and led in a natural way to the ordering of the groups to indicate the synonyms.

Finally (almost) came the clue writing. As a solver, I find clues which are thematically related to the puzzle itself more satisfying. As a setter, I enjoy writing thematically constrained clues, but it can be hard work. Producing thematic clues using definitions which did not give the middle game away, having a specific first letter and a specific fifth letter in the fifth word was probably the biggest challenge I’ve had in any crossword to date. Clues such as the &lit one for CANOPUS (“Receptacle with truncated animal cover over?”) took many rewrites to get right. Even then purists would argue that it should be “and” not “with” in the wordplay reading (I vacillated between the two), but I think it can be justified by considering “cover” to be imperative. The clue to AZOV (“Inland Eastern European sea, fifth’s covered all over in water fern”) was another tricky one, and while the final version is sound, I would have preferred to have found a simpler treatment. I’d also have written clues to some of the non-thematic entries quite differently if I’d not had the constraint of an additional hidden message there in the original version.

Kudos to my test solvers, who are cited in the grid, for solving the puzzle in its trickier form before clues had been refined and more cross-checking words added. The Listener Crossword editors made the whole puzzle a lot more accessible to the typical solver by adding these additional words, as well as changing the second message which originally read CHOOSE THE RIGHT ARM POSITION. A few clues were removed, others trimmed and new ones introduced. Apart from the thematic clues, the only constraint now was that all clues should have at least five letters in the fifth word.

And the mistake in the clue numbering? I would have asked my daughter to check the proofs for me, if she had not been away at university. This puzzle is dedicated to my inquisitive daughter, now 20.