Mystery Romance by Samuel
As I may have mentioned before in my 2006 Listener solving weblog, my brother Philip is very into classical music. Not well-known classical music, rather English composers who will be unheard of by most people. He actually runs a company called Chosen Arts that aims to promote such composers. He has published books on such composers as Gerald Finzi, has been an active member of societies including the Havergal Brian society (of which Dimitry is a senior member), and so on. However, his most loved composer is Edward Elgar, and it was Philip who first suggested a puzzle based on Elgar, sometime back in the summer of 2006.
I’m not an expert on Elgar, but everyone has heard of the Engima Variations, and it seemed a good crosswordy sort of theme – Enigma Variations suggesting anagrams of ENIGMA, and the obvious link to the EV in the Sunday Telegraph. It seemed such an obvious theme, however, that it surely had to have been tackled before. A trawl through the EV and Listener archives, as well as through my box of old Magpies revealed a surprising fact – the Enigma Variations didn’t appear to have been used as a theme in a few years, if not longer, so it was fair game. 2007 would be the 150th anniversary of his birth, so it was probably best to get a move on.
Unlike some puzzles, this didn’t arrive as a fully-formed idea, but needed some thought. So, as normal, it went into the ‘ideas for puzzles’ part of my brain – the part which I normally examine on long car journeys to help pass the time. In this case, the gestation period wasn’t very long – in fact the following Friday, when looking through ODQ, I came across the surprising fact that the dedication to the Enigma Variations was to be found in the puzzle:
To my friends pictured within
The mental leap to deciding that the puzzle could give the quotation, and have to contain all of the ‘friends pictured within’ was not a big one, so the next step was to sit down and work out how to fit them all into the grid.
The biggest problem seemed to be that lots were given simply by initials. This rather scotched the idea of having them all found in the grid (I had some vague idea of having all of the friends in the grid, in the shape of EV, or something similar. The next idea was to have all of the ‘initial’ friends as initial letters of clues. This would depend upon the number of letters in the friends, however. Helpfully, this added up to 27. It would have been good if this could have been 36, but unfortunately not.
This left NIMROD, YSOBEL, TROYTE, DORABELLA and the unknown * * * to sort out. I fancied having these as four unclued entries, which would leave the * * * to fit in somewhere. But where? And what could I do to ensure that people had actually located all of the 13 friends?
This provided a poser, but I sat down with Sympathy to come up with a grid that would contain the four names above. The other idea in the meantime was how to indicate the ENIGMA VARIATIONS. I probably didn’t have to – the quotation and the names would be enough – but it would be nice to have them shown somehow. The idea then came to me to have the anagrams of enigma omitted from answers before entry. This may be more challenging to the solver than just having answers eg containing anagrams of enigma, as it would mean that not all entries could be made normally. Which was nice. It also meant that there was probably some more scope with the entries, as with this method some pretty long words could be fitted in to a 12×12 grid.
At that point, I used TEA to create a list of words containing anagrams of ENIGMA, of which there weren’t that many. Another long car journey home from Surrey, and I figured that leaving one of the friends to be written beneath the grid (or found in the grid) would be the way to ensure that solvers had found all of the friends. And what about * * *? Well this was meant to represent an unknown lover of Elgar, so why not make this into the title? The title of the puzzle would thus be * * *, and I would have to put something into the preamble to make the solver believe that they had to find the correct title of the puzzle.
So, it was time to take stock. I had a 25 letter quotation to generate, anagrams of ENIGMA to be removed from answers before entry, 27 clues to start with set letters (although there was some scope for changing the order of the friends to fit in with the clues), 3 unclued entries in the grid … was there enough thematic stuff in there?
I thought yes, so it was time to tidy up and generate a fill.
A 36 clue 12×12 grid would be nice and tidy, so with the 3 unclued entries that gave 33 clues to write. I had 25 letters to generate for the quotation, so that left 8 clues not needed – could I get 8 words with anagrams of ENIGMA removed into the grid? I then had 27 clues to start with set letters, leaving 6 spare. Hmm, what to do with those 6? I needed to use them. It was then that I realised that ENIGMA contained 6 letters, and that 27 + 6 = 33. Brilliant!
The final piece in the jigsaw – what to use as the gimmick to generate the 25 letters for the quotation? I couldn’t think of anything new that was vaguely thematic, so as I hadn’t done superfluous letters for a while, I thought I would use this.
So, ideas all in place… and then a week’s thinking time to make sure that I couldn’t come up with anything better. A week turned out to be a month, nothing better came to me, so it was time to fill a grid.
I don’t know how other setters handle this sort of thing, but I used Sympathy. I started with the same grid that I had used for a previous Samuel puzzle, and started by adapting it slightly to fit in the three unclued entries, TROYTE, YSOBEL and NIMROD. To check that a fill was still possible after placing these in the grid, I then used Autofill to check that all was okay. I was only interested in positioning of the three answers that gave more than 10 fills at this point, as slotting in 8 entries where anagrams of ENIGMA had been removed would soon put a limit on the possibilities.
So began a long slog, that took about two months overall – during which time I almost gave up. Some puzzles have a lot of thematic material in the grid, some have none – and to come up with a grid and a fill with 3 unclued entries, and 8 other entries under massive constraint (a total of 11 of the 36 lights) was very challenging indeed. It was easy to get 5 or 6 answers omitting enigma variations into the grid, but then it all went pear-shaped.
In the end I was so desperate that I almost gave up. However finally, one glorious Sunday, I finally came up with a grid and a fill that did work! There were some horrible words, though. LEAMINGTON SPA was a nice one that contained an anagram of ENIGMA, but very hard to clue I would have thought. LIGAMENTAL would be hard to get a good definition for, LETTERMAN was a surname of a chatshow host, and TYPHOEAN looked plain horrible to clue. Oh, and there was one name, ETHEL, which broke a Samuel rule about having given names as answers. But mostly okay, though, complete with one word that I had wanted to write a clue for ever since I started out as a setter – YO-YO. Why? For no other reason that I remembered an old Hale and Pace sketch where one of them says ‘Why, oh why, oh…’ and the other says ‘spells Yo-yo’. I’m not a fan of homophone clues, but this would be an exception.
So, some four months after the puzzle’s inception (and it now being almost Christmas) it was time to write clues.
Some clues came easily (TRON for scales was the first, but never quite worked out how to punctuate this correctly), some not so easily (ABUTTER, for instance, which had a much harder definition than first appeared). What soon became obvious that I was generating a lot of clues with film/TV references. The De Niro reference for CINEMA-GOER was the first, closely followed by HOTSHOE (part of camera shoot…), and so on. 23A, TRIGEMINAL, stole the use of Tracy EMIN from a clue in a Listener (but no more than that), and eventually the puzzle was complete. At some point during this time, the * * * of the title got replaced with Mystery Romance, which was a representation of the ‘unknown lover’ referenced by the stars, and hopefully less obvious to any opera buffs solving.
The puzzle then went for test-solving, then shortly afterwards to The Magpie. Mark Goodliffe did a superb job to solve it only a few days after submission, and we liaised on amending a few clues – notably a change to YO-YO, and the introduction of a pre-nooky judo costume!
The final change (a clue change for CATALASE) happened on 1st February; it was then time to wait until publication. And, at that point, there seemed little else to say on the puzzle.
Until, four months to the day after I deemed the puzzle complete, a Listener appeared called ‘Initials’ by Mango. Aaarrrrgggghhh! I had already lost out on a Magpie puzzle because of duplication of theme, but this was duplication of the theme in spades. A very anxious three week wait followed, but on a couple of days after the publication of the solution to Initials, I received an e-mail from Mark saying that the Mango puzzle was sufficiently different to mine that it wouldn’t affect publication. Whether this has affected solvers’ experiences will be seen in time, but hopefully it’s caused no problem.