Transposition Cipher was inspired by my previous Listener puzzle Impossible Construction. In that puzzle, a central panel in the completed grid contained a word square, will all grid entries outside the central panel being real words. I thought it would be interesting to see what other content could occupy a central region of the grid.

The message TRANSPOSITION CIPHER BY SERPENT has 28 characters which would fit nicely into a 4×7 panel that could be incorporated in a 10×13 grid. I was going to use a standard columnar transposition cipher but then thought it would be more interesting to use the Myzskowski variant with ANAGRAM as a rather natural keyword.

I used the approach of Impossible Construction of jumbling entries that intersected with the central region, with the constraint that the three letters of these jumbles outside that region should form real words. The excellent Qxw software confirmed that a grid-fill was possible. From there, I incrementally refined the grid in order to exclude as many obscurities and inflected word forms as possible.

The ambiguity in the grid-fill was irritating, but I realised the bug could be turned into a feature because it is disambiguated by (what is unambiguously) the correct plaintext.

As always, I must thank my test-solvers, David and Norman, and Shane and Roger at the Listener for their feedback and help in producing the final version of the puzzle. And many thanks to all the solvers who have been kind enough to send comments in with their solutions.

A number of solvers commented on the dimensions of the ciphertext panel being identical to the puzzle number (4747). That’s not something I’d noticed when checking the proof and Roger assures me that this is pure coincidence.

The inspiration for “Compressor” was actually a previous Listener crossword in which a few clue answers were anagrams of place names, these names forming the actual grid entries. I thought I would try something similar using the names of composers. However, it soon became clear that the number of composers whose names could form anagrams were fairly limited. To make it easier for me, and hopefully more difficult for the solver, I thought that if I introduced an additional letter into a composer’s name this may produce a better opportunity to form an anagram, around which a clue could be formed. The actual grid entry would then be the composer’s name formed by making an anagram of the clue answer minus the additional letter. It soon became clear that this could actually work, using the names of many more composers with an additional letter, rather than the composers’ names themselves.

I then set about trying to fill an empty grid using the names of as many composers as I could, and eventually managed to get 15, whose names could all form anagrams of real words by the addition of a single letter (shame that there has never been a composer called “Umbriel” which would have made a nice symmetrical 16th!). The major time-consuming difficulty was then trying to complete the grid, which unfortunately resulted in the necessity to use a few uncommon words (“Eavie”, “Baedi”, “Oromo”). I decided also to make the puzzle a little harder by presenting solvers with a blank grid. However, the inclusion of clues to 2×12 letter words intersecting with the 4×3 letter words, none of which were thematic, would at least give solvers a good start to filling in the grid.

I felt that it was essential to make use of the dropped letters from the thematic clues somehow so that solvers couldn’t just identify the likely composers without properly solving each clue. This led me to the idea of giving an instruction in the non-thematic clues by using a misprint in each of these “normal” clues, whereby the corrected letters in usual grid entry order would produce an instruction to somehow use the discarded letters (see below).

The next difficulty was trying to come up with a suitable title. Clearly “Composers” would have been too much of a giveaway, so I eventually decided to see if there was any possibility that I could treat the title in the same way as the thematic answers became actual entries. In the end I came up with “Compressor” (the “key word”) which by removal of an “R” was an anagram of “Composers”. Using this “R” along with the discarded letters from the thematic clues allowed the formation of two further composers – Schubert and Schumann, which solvers had to identify.

Thanks are due to the Listener editors who improved a number of clues and introduced me to the word “termitaries”! For non-setters of Listener crosswords, it may be worth noting that it usually takes a number of weeks to compile a crossword, and after submission, around 18 months or so before hearing whether or not that submission is accepted as a future Listener puzzle.

KevGar

The overnight ferry crossing to Shetland was, thankfully, a calm one. We spent a little time in Lerwick before setting off on our journey to explore the isles, particularly the spectacular coastal views. We joined the D970 and headed for Northmavine, prepared for the narrow, twisty roads and grateful for the passing-places dotted along our route. In less than an hour we reached Brae where we took a short break. The guidebook promised “a magical landscape” and “stunning cliff scenery” were we to continue northwards, and perhaps later we would see the Northern Lights. So we were back in the car, with cameras at the ready, to confront Mavis Grind, where the Atlantic Ocean could be seen cutting through almost to the North Sea.

We pulled in to admire the little strip of land that joins the Northmavine peninsula to the rest of Shetland Mainland. We talked about the opportunities this geographical feature might offer if it were placed at the heart of a Listener crossword; we pictured a strip of white splitting a grid of blue sea, with the words Atlantic and North Sea on either side. The width of the strip is so narrow – a stone’s throw, shall we say? We carried on imagining …

But enough of this pretence! For me, setting the crossword wasn’t at all like this. It was a different kind of journey. For one thing, I work from home nowadays – at my desk. The landscape I explore is Chambers and the internet. The search for a crossword theme often begins with a play on words. Here, isthmus, a curious word containing four consecutive consonants, was my starting point. It was a problem of pronunciation when I was a pupil in school and a term I learnt in geography alongside oxbow lake, tsunami, and glacier. But now I took it to initiate an investigation and to ask where examples were to be found. Scanning the pages of Google, Mavis Grind (a charwoman in a Dickens novel?) caught my eye and the stories about it were interesting. I looked for ideas for the puzzle and homed in on coincidences: “iSTHmus” and “a stone’S THrow” have common elements; Atlantic and North Sea each have an anagram; Mavis and Grind could easily be clued separately. I also wanted to finish with a puzzle that was compact, as a previous submission had turned out too long, and would not fit into the space provided for the Listener. In the event, I managed to arrange my entries to fit inside a 9×10 grid. Normally, Listener grids contain around 155 cells (rectangular grids in the Listener during 2021 had on average 154.7 cells), so would my 90 cells be acceptable? I needn’t have worried. It was given the go-ahead by the editors who also made several helpful suggestions for which I thank them.

Warning is a favourite poem since I came across it many years ago, a view apparently shared by many as it seems to feature highly in lists of favourite poems. I suspect, like me, many people know the first lines without knowing the title or the poet – it was while looking it up for some reason I can’t now remember (it was in early 2019, and a lot seems to have happened since then) that I discovered via Wikipedia that Jenny Joseph was born in Birmingham, my home town, and that the poem was first published in the Listener. Those connections of course immediately suggested a Listener Crossword, and one that could mark the sixtieth anniversary of publication.

My first task was to find the exact publication date, which wasn’t helped by Wikipedia claiming that it was in 1961. Consulting the index to that year’s Listener magazines produced no result, so I tried 1960 and 1962 and found it under the latter. (I then went back and edited the Wikipedia entry to match, not least because I thought many solvers might check there.) That at least gave me a little more time to compose the puzzle, bearing in mind that date-specific puzzles need to be submitted well in advance.

The idea of having Jenny Joseph in purple with a red hat above was my starting point, with the title also being amongst my earliest thoughts. I produced a grid but wasn’t too inspired by it and ended up working on and submitting a few other ideas, including You Don’t Say which appeared as a Listener in 2020. I came back to Imperial Age in 2021 with the date now getting more pressing and came up with the message — Warning which appeared in The Listener sixty years ago — around that time. It was helpful that I thought most solvers wouldn’t know the poem’s title and so the message would seem to be thoroughly unhelpful until the penny dropped and then a clear confirmation, which I always like to see as a solver. I also decided to go for a letters latent grid to generate the message, mostly because I couldn’t see anything more obviously thematic and it’s a device I hadn’t used before. The grid proved easy to generate thanks to QXW: my respect to setters who produce such grids without electronic assistance.

The submission duly went off and I was pleased to see it accepted for the anniversary date with the editors’ customary improvements and corrections. My thanks to them and to all who have provided feedback — I’m glad that it generally seems to have been enjoyed, obviously helped by the affection many have to Warning and that others were glad to have been reminded of it. So a final thanks to Jenny Joseph for the poem and the inspiration.

From time to time during the 19 years I’ve been setting thematic cryptic crosswords, I’ve asked myself whether I should try creating a numerical puzzle. Apart from the need for a suitable idea for the theme, I had some concerns about ensuring that there would be a unique solution, and that it would be reachable by a sequence of logical steps with a minimum of trial-and-error calculation. My first introduction to numericals was in October 2019 with the Magpie puzzle “HexaFlex”, a base-16 puzzle Tim King (Encota) and I co-set under the pseudonym EP. For that, Tim was the driving force on the numerical side, in a puzzle where digits became letters on the faces of eight connected cubes. This gave me confidence in venturing into the world of numerical puzzles, with my first solo effort, “Pathfinder”, published in the Magpie in September 2022.

Having submitted that one, I immediately started work on another as I had a theme in mind which promised to provide a novel, and hopefully interesting, numerical puzzle. The theme was quantum mechanics, especially the concept of the collapse of superposed states. With its strangeness, for me this branch of physics has always had its own charm. My target was the Listener series.

From the start, the plan was a numerical of the type where integer values have to be assigned to a set of letters. The twist would be that each letter would exist in a superposition of “states” (values) and would “collapse” into one of them, depending on which clue was being “observed”. Moreover, two different letters might assume the same value if they were in different clues. Initial explorations using a superposition of three states per letter showed how intractable such a puzzle would be to set (never mind to solve!), even with some constraint placed on each letter, such as possible values being consecutive. So a rethink was needed.

What I arrived at was two consecutive values for each letter, with across and down clues each using consistent values, with no repetition of values. I felt this to be much fairer to the solver, and the intention wasn’t to set a puzzle of horrendous difficulty. But there’d still be plenty of scope for a solver to make a slip! At that point, the title “Even Odder” suggested itself, given the nature of the number pairs, and the fact that quantum mechanics is just that!

I decided that every clue would be a “formula” using a thematic word. With some help from online resources, I was able to assemble a list of nearly 100 words (excluding plurals etc.) relating to quantum theory and particle physics. At that point, I introduced the feature of digit-triple sums in the grid’s columns leading to some relevant wording below the grid, using A=1, B=2, etc. I thought of quantum physicists’ surnames, but that looked too obvious, and so I decided to use a paraphrased quotation. This gave me a feel for what the grid dimensions would be. The setting process then evolved into a non-linear one, with several requirements having to be met simultaneously. These included how many, and which, letters to use to give me a workable subset from my thematic word list, and that answer lengths were to be no more than 6 digits (to avoid solvers facing overlong calculations) yet maintaining a reasonable overall average. Also, each triplet in the grid would need to add up to a number in the range 1 to 26. My final choices were 16 letters representing 32 values (leaving around 70 thematic words to choose from) in a 10×9 grid with 48 clues, and no entry longer than 5 digits.

So after much head-scratching, and over 30 A4 pages of closely-written notes, I had a completed puzzle which I had some confidence in. Happily, my feelings were confirmed by my two, much appreciated, test-solvers — sincere thanks guys, you know who you are! — and I was able to submit the puzzle to the Listener.

In any numerical, there has to be at least one clue, or a combination of clues, acting as an “entry point” for the solver. I decided to use factorials for this purpose, allowing the answer for 11a to be directly entered in the grid as the only 5-digit factorial (8! = 40320). In conjunction with 12a, it was then possible to discover four letter values. Consideration of 8d, another “factorial clue”, allowed a further six values to be tied down. So ten found, twenty-two to go! The process could be readily continued until all letter assignments had been found, and the paraphrased quotation discovered. Other variations on the entry point exist, of course.

I’m very grateful to the two Listener vetters for their careful appraisal of the puzzle, including some tightening up of the preamble wording. All those solvers who commented on the puzzle, using whatever means, also have my thanks.

Phil Lloyd (Ploy)