Listener No 4656: Co-star by Lysander
Posted by Dave Hennings on 14 May 2021
A new setter this week. A straightforward (albeit asymmetrical) grid greeted us, and a nice short preamble. I anticipated an easy solve until I read the first six words: “All clues must be enciphered before entry…”. Lawks! And each letter was to be replaced by either one or two letters, all twenty-six being different.
It was fairly obvious that I wouldn’t be able to determine the coding for each letter until the grid was complete, but one thing that seemed necessary was identifying which letters were to be replaced by two letters — at least, I thought that seemed likely. I listed the alphabet, although I wasn’t too sure how that was going to be used!
Starting on the clues, 1ac Struggler loses races in farm equipment (5) eluded me but was a 5-letter answer fitting a 6-letter slot. (Thank you, Lysander, for not just giving entry lengths against the clues.) 5ac Unlimited mixture with sesame seed (6) came to the rescue with LENTIL and that was a 6-letter entry, so all its letters were encoded by a single letter and I marked them in my list with a single •. Ah yes! Double letters would be marked •• and would eventually give the two letters from which they were to be encoded.
9 Friendly female stirring yolks (6) was also easy with FOLKSY fitting a 10-letter entry. Answers came thick and fast, and after my first pass through the clues, I had all but a dozen resolved. Of these, only six consisted of single-letter encoding: LENTIL, RIEL, LIRAS, TELLER, INSULAR, ADITS.
From this, it could be deduced from BRADS at 1dn that the B was a double letter, as was the X in LEX at 3dn and the C in CIRE at 6. I marked these double letters across the grid lines. It soon became clear that, with these double-letters intersecting single-letters, we would end up with an interim grid with each cell containing one letter from which the final grid could be decoded. Thus the C at 6dn would be encoded in the same way as the N and S of 5ac and 9ac.
After my first pass, all but Q and Z didn’t appear in any answer, but needed to be as we were told “all 26” were encoded differently. These would eventually be given by 14ac Queen has trendy craft that was deceptive (5) for Q-SHIP [Q + ’S + HIP] and 31dn Small business needs European for international second point (3) giving BEZ [BIZ with E for I]. Q-SHIP held me up for some time as I had crossing letters giving it beginning ER…!
So the grid was complete with a single letter in each cell. Where to go from here? A cursory examination of the grid showed that only the letters A, D, E, I, L, N, P, R, S, T, U were in the grid, although I wasn’t sure how important that was. What did seem important was that 21ac PIDSIP, 8dn LIPPIP and 27dn SIILIP had a lot of LIPs, SIPs and PIPs! With a bit of trial and error, these three became DENTED, REDDED and TEERED. REARRESTED at 9ac soon got resolved and everything flowed nicely from there. The final grid contained the same letters in its coded form as given above but with an O and no U.
A careful checking of the answers, entries and final grid made sure that everything was correct. Well it was, once I corrected SARSI at 23ac to SARSA!
Thanks for a remarkable puzzle, Lysander. Where did the idea come from?
Listen With Others 
Alan B said
This was daunting at first but ultimately solvable and rewarding. The whole concept and design were amazing – I find it hard to comprehend how the solution grid could have been created such that it could be enciphered from a grid of mostly shorter words using such an irregular cipher.
I’ve read with great interest three of the four blogs already posted on this puzzle (I have yet to read Steve Tregidgo’s), and naturally my process of solving it was very similar. I used an intermediate grid in which I filled in the answers of the correct length first and then spaced out all other answers roughly at first but more and more precisely as it became evident what the intersecting letters were. I wrote all letters that expanded into two letters in red ink on the relevant cell borders (so that they could easily be seen and then just as easily ignored) and let the pairs of letters blend neatly with the single letters into the pre-encoded entries that filled the grid.
I noted that all 26 letters were represented in my intermediate grid, although all the single letters that now filled it prior to enciphering were from the same set of 11 letters. The enciphered letters were from that same set but for one letter, U, which was encoded to O.
Q-SHIP, appearing as ERSILIP in my intermediate grid, was, not too surprisingly, my last solution.
The endgame was like a Codeword puzzle, but more interesting. I thought of REARRESTED early on from its distinctive pattern of letters (LIELLINSIP before enciphering) but queried it because it gave me the unfamiliar words REDDED, TEERED and DESSE in other entries! Other words, however, like ALTERED, SPIRAL and SPIRIT, also emerged and confirmed my hypothesis. I checked for myself that all the final entries (with exceptions as indicated) were in Chambers: I did this not so much to look for a setter’s error but to ensure that I hadn’t made an error myself, e.g. in an unchecked cell.
I didn’t know that I had the skill to tackle this puzzle, and I think it says a lot for the overall design, the quality of the clues and the precision of the succinct written instructions that I was able to complete it. It was an exceptionally good puzzle.
Huge thanks to Lysander, and thanks also to the bloggers. Steve’s blog also engaged my interest and I plan to visit that next.