I became hooked on Tetris in a French ski resort in about 1989, and later obtained a copy for my now obsolete PC. So Tetris came to mind when I set out to become a Listener Setter. I submitted an earlier version of this puzzle in 2004, but failed to understand some of the constraints then applicable, and the puzzle was rapidly and rightly rejected as being too large.
A few years later, following retirement, there was time to try again. The original idea seemed sound, but this time I would meticulously follow the very helpful guidance published on the Listener Crossword site.
The concept was to have normal across clues and to have Tetris shapes to interlock with them instead of having down clues. I had not come across any other use of Tetris in The Listener, but it was possible that I had missed something. I have since seen a comment that the idea was first aired long ago, and I would be interested to know when and how.
Given the ostensible theme of Tetris, was there some other hidden theme that could be found? The anagram of SITTER didn’t spark any ideas, nor the thought that TETRIS plus HELEN produces THE LISTENER. TRISTE did, however, ring a bell, not just from French but from Don Quixote – for some reason I had come across “the knight of the sorrowful countenance” as being the standard translation of “el caballero de la triste figura”. Wonderful – although I’m not familiar with Spanish cryptic crossword anagram indicators (if they exist), I felt that “triste figura” could surely be stretched to suggest an anagram of “triste”. (Whole treatises have been written on the meaning of “triste figura” – I am merely supplementing them.) So that gave me the cryptic theme. The Spanish phrase is helpfully in the Oxford Dictionary of Quotations and, not being in English, made it less likely that the puzzle’s mysteries would be quickly exposed.
So Don Quixote and the phrase seemed suitable for the perimeter, but it was hard to find the precise words to fill the perimeter exactly. The puzzle’s grid had to be 14 by 12, so there had to be 48 letters. Various combinations of La Mancha, Cervantes, etc failed to work, but fortunately Sancho Panza did and, miraculously, he was the supposed author of the sobriquet. Don Quixote and Sancho Panza could be satisfactorily positioned across the top and bottom, with the phrase as the left and right columns.
There are seven possible Tetris shapes, and they suggest letters, which led to the concept that the shape could be described by its letter as a misprint in the clue. Six shapes were obvious letters – O, I, T, L, S and Z, but the remaining one was a problem. I decided it should be “r”, albeit in lower case. It was only during the vetting process that I discovered that that shape is usually (where else?) represented by “J”; whilst this may have made clueing a little more difficult (the five “Z”s posed significant problems anyway), it would have been a more elegant choice.
A perfect construction would have an equal number of each shape – hence the choice of a 14 by 12 grid, which could accommodate six of each shape. I failed to achieve this, but at least managed a reasonable balance: six of O, T and r, seven of I and L, and five of S and Z.
Forty-two 4-letter words would be excessive (and certainly not the answer to life, the universe and everything), so I decided to pair shapes to produce a number of 8-letter words. Ideally the join should have been in the middle of the word so that the eight letters could flow from first to last. This proved to be too demanding a constraint, so I accepted 8-letter words with the two shapes touching at any point. I was pleased to be able to fit in thirteen 8-letter words; unfortunately only five join at the middle of the word.
To reinforce the theme I decided to place the phrase “TILTING AT WINDMILLS” in a centralised position, and crossing to suggest the shape of a windmill. (Tetris shapes can be rotated as they fall.) That produced a further severe constraint when filling the grid. The across clues had to comprise ten letters, which could be one or two words. I had hoped to have at least four of these as 10-letter words, but this proved too demanding, particularly as “tilting at windmills” only left me three rows which did not have two fixed letters from that phrase. And those three rows were near the top and bottom where X, Q and J (among others) had fixed positions. I ended with just two 10-letter words. With the “double” shape clues, however, the average clue length was 5.65 letters, which seemed entirely acceptable.
Filling the grid proved extremely time-consuming. Although there appears to be lots of flexibility, with the shape clues going in either direction, the constraints proved very difficult to manage. The perimeter and central phrase determined 66 of the 168 cells (almost 40%). Additional 10-letter words had to be sacrificed, sometimes in the interests of extra 8-letter words. Perhaps some neat programming might have helped, but that was beyond me.
If one allows “checking” to include the perimeter contents and the central phrase, then I would claim that every cell is checked, and indeed eighteen are double checked. I am proud of that.
The concept of a pangram never crossed my mind – in fact the puzzle falls short only by a “v”.
As a novice at clueing, I found it quite difficult, and I recognise that some of the surface readings are not as felicitous as I would have wished. The 8-letter shape clues had to have two misprints, not unfairly located – all bar one having one in the definition and one in the wordplay. That raised some challenges – e.g. ENORMOUS has to have two Zs as misprints, though I was very happy with “Outsid(Z)e type h(Z)ero in counter claim”. Most of the across clues comprised two successive clues, which again complicated the production of satisfactory surface readings. I received a good deal of assistance from the editors in tidying up many of the clues to ensure that they met strict Listener standards and avoided deprecated constructions, but there remained a number where their slightly reluctant tolerance was appreciated.
The across clues could have been clued separately, but that would suggest the need for bars and extra numbers in the grid. Bars were needed to show the shapes, so it was undesirable to have them in the rows. There was no symmetry in the across clues, so I decided to have them as one or two clues run together, leaving the solver to deduce the lengths and break in the two-word rows. The published solution shows bars to clarify the shapes, but the requirement to show bars in submissions was dropped during the editing process. Whilst this simplified submissions and checking, it had the disadvantage that it was possible to reach a solution without having solved all the shape clues.
A nice suggestion was made that I might have required the Tetris shapes to be shown in their traditional colours. This would have produced a pretty end result. Of course, the original version of Tetris was monochrome, and there would also have been an issue about how to mark “tilting at windmills” – perhaps drawing straight lines through it would have made the point well.
The puzzle was undoubtedly tough, but it was hard to judge just how difficult it would be. Clues could have been easier, and it was always an option to give word lengths. Space was a problem, so it helped a little to combine some across clues and omit clue lengths.
I originally submitted the puzzle in February 2013, when there was a substantial backlog of puzzles to be assessed. As a new setter I could expect no special consideration, so I was not surprised to hear little more than an acknowledgment for a long time. In December 2013, having been prompted by details of the Setters’ Dinner, I looked up Tetris on Wikipedia and discovered that it was originally released on 6 June 1984. The thirtieth anniversary was coming up (albeit with potential competition from the seventieth anniversary of D-Day), and I drew this to the editors’ attention. I suspect that that prompted them to have a look at the puzzle, and as a result, after a lot of assistance and some brilliant ideas from them, the puzzle was eventually cleared for publication.
I must also thank my long-standing friend Bundle (who has not set for some years) for his considerable help, initially as a guinea pig and later in improving the grid and the clues. I have been fascinated and greatly gratified to read comments about Tetris (mostly extremely kind), both those forwarded by our wonderful statistician and those on another website.