Oh, hell. On Friday, I downloaded the just-published solution to the final Listener puzzle of last year (“Coincidence” by Sabre) and noticed that I’d made a small error.

It was a brutally difficult puzzle — easily the hardest of the year, it seemed to me — involving a lot of letter substitutions, and although I made it to the end, answering all the clues and figuring out the two hidden messages (which were quite difficult to work out even once you’d figured out what they had to say), I failed to thoroughly double-check my “bookkeeping” and so missed one place where one of the two letters I had as being involved in the substitution didn’t match one of the two letters involved in spelling out the messages. Particularly stupid because I’d noticed something odd about that portion of the grid and didn’t spend enough time looking at it to realize what was wrong. Argh argh argh.

Americans have a time disadvantage with the Listener puzzle, though. You have two weeks from the time the puzzle appears to the deadline for entries. Entries are accepted only by mail, however, not by email, so if you’re in England you have about a week and a half to solve the puzzle and have it arrive by the deadline, but if you’re in America you have maybe three or four days. The puzzle appears on Friday, so I try to get my entry in the mail on the following Monday if possible. And this was a very difficult puzzle, as I said, and then Friday and Monday were both filled with family stuff for the holidays, so I didn’t finish until early Monday morning and I hurried it into the mail, and I never really had a chance to put the solution aside, sleep on it, and double-check it in the morning with a calm, clear head.

On the other hand, even a late entry counts toward the yearly statistics as long as it is sent before the solution has been published, even though it can’t win the weekly prize. So perhaps this is a lesson that if I smell anything wrong, I should hang on to my entry another day or two till I figure out what’s amiss, even if it means giving up a shot at the weekly prize.

I don’t usually check my solutions, so I have no idea whether my entries were all correct for the year other than this one. It’ll be a sad thing if it turns out that I missed a complete for the year due to one small error on the final puzzle. I’ll find out when the yearly stats are issued. But I’ve made errors on exactly three puzzles in each of the last three years, so the chances are probably not all that great of this being my only mistake all year.

After solving Listener Crossword 4295, “Codebreaker”, a few weeks ago, I took my notes and organized them into a step-by-step solution. Now that the deadline for the prize drawing has passed, I’m posting it on the chance that anyone is still wondering how it could be solved. Look at the PDF file here.

I didn’t finish this Friday’s Listener crossword, “Absolutely Pointless” by Waterloo, until Monday — figuring out the last two clues and filling the last two cells in the grid during my midmorning break — despite having had Friday off work. Partly because I also had a lot of other things I wanted to get done on Friday, but partly just because it was a very tough puzzle.

The idea is that answers can be entered into the grid in any of the eight compass directions, not just across and down. What’s more, an answer can change direction one or more times. The answer can start in any direction, but whenever the letters N, NE, E, SE, S, SW, W, or NW occur in the word, corresponding to the points on a compass, those letters are not entered (making the grid “pointless”) and the rest of the word continues in that direction (which is usually but not necessarily a change in direction). For example, MOONWALK might be entered as

or perhaps as

The instructions, I thought, were much too terse, and I guess I should give a spoiler alert here in case you are working on the puzzle, because I’m about to write about what, as a former puzzle editor, I feel the instructions should have covered. No clue answers will be spilled or hinted at, however.

The very brief instructions left me with the first impression that entries were going to snake through the grid, possibly changing direction with every letter even where there wasn’t a compass direction to fix what that direction was. It wasn’t long before I saw that if that were true, the puzzle was unlikely to have a unique solution (not without a lot more clues), so I decided to solve with the tentative assumption that an entry couldn’t change direction except where it contained a compass direction, figuring that if that wasn’t the case I would bump up against an impossibility soon enough. I never hit a contradiction, so that was the right assumption to make, but there are probably any number of alternate “solutions” possible without that assumption, and as far as I can see they would be valid according to a not unreasonable interpretation of the instructions.

It would have been nice, too, if the instructions had mentioned that each entry contains at least one compass direction and may contain more than one, that not every compass direction causes a change in direction (when the entry is already heading in that direction), and that an entry may therefore change direction once, more than once, or not at all. These are less crucial pieces of information, as you eventually figure out in the solving that these things have to be true or no solution is possible. But it would have made the puzzle easier to get a handle on if they’d been spelled out.

Worst of all, the instructions did not mention that a single entry may go through the same letter in the grid more than once. BORNEO, for example, might be entered as

This was the omission that really annoyed me, because I figured the natural assumption would be that reusing the same letter is not OK, and that therefore they’d surely mention it if it was, and as a result I eventually reached a point where it was impossible to enter all the answers I’d gotten so far, and I had to scrap my first grid and start over with a fresh copy. Bah humbug. I like a tough puzzle, but I’m not so hot on making a puzzle tougher by making the solver guess at what its rules are. And this puzzle was tough enough as it was!

Other than the frustratingly uninformative instructions, though, I thought this was a fine puzzle. I’ve read some criticism online of the fact that it doesn’t end with any kind of surprise — no hidden phrase appearing in the completed grid, no quotation by Lewis Carroll being spelled out somewhere, that sort of thing. But I don’t see why every Listener puzzle needs to be like that. This was a good and enjoyable challenge, if a very tough one.

One of the things most amazing to me about the Listener puzzle is that, even though prizes go only to the first three entries drawn at random each week, every entry is checked, and a full record is kept of how each individual solver does through the year. Given that there must be a thousand or so people who submit at least one entry during the year, this seems both wonderful and awfully obsessive-compulsive to me. And a tribute to the cult-like following that the Listener crossword has acquired over the decades.

Around the end of each March, you can send in a self-addressed stamped envelope and receive a ten-page report about the previous year’s puzzles. This report lists:

• all the year’s puzzles, identified by number, name, and constructor, along with statistics about how many correct and incorrect entries were received for each and how many of those entries were from new solvers
• a list of the most common errors made in each puzzle
• the top 200 or so solvers, along with the numbers of the puzzles (if any) they got wrong
• a table summarizing how many of the people who sent in x errors during the year got y correct (for example, of the seven people who sent in exactly 27 entries during the year, one got 26 correct and two each got 25, 21, and 20 correct, and what possible interest there is in these data I cannot fathom)
• a report of the annual dinner that is held for the constructors (in British terminology, setters) of that year’s Listener puzzles, which many solvers also attend

I made the list of top solvers this year (yay!), getting only three puzzles wrong out of the year’s 53. The first page of the list is taken up by those who made none, one, or two errors, so my three are enough that I only make it to the second page. And not even near the top of the second page! Those who miss the same number of puzzles are further ranked by when they made their errors (the longer your string of correct solutions from the beginning of the year, the better), and as I made my first error on the second puzzle of the year, I am near the bottom of those who made three errors. I haven’t yet counted exactly where I am in the rankings, but it looks like somewhere around 70th or 75th. Not a dazzling performance, but not too bad.

I knew about one of the errors already — while looking at the published solution to one of the puzzles, I spotted that I’d put in the wrong spelling of an archaic word that has several spellings given in Chambers Dictionary, apparently having never taken the time to go back and double-check that my chosen spelling fit the wordplay in the clue (it didn’t quite). As this puzzle was late in the year, my excuse is that I was preoccupied with finishing The Manga Flute. But the other two errors were a surprise, both of them looking like silly copying errors I must have made when preparing my entry.

Into the stapled packet of photocopied pages is folded a single, additional, loose piece of paper. It is a list of the exact errors that you personally made during the year — yes, it really tells you that you got 18 Across wrong on this puzzle and 25 Down wrong on that puzzle and so on. Do you not find it both wonderful and freaky that, in this day and age, there is somebody in England, who cannot possibly be getting more than a modest stipend at best for all this, who week in and week out is actually keeping track of these things for a thousand or so solvers, and will send you your individual stats for the year just for a self-addressed stamped envelope? Overseas solvers like me don’t even have to send stamps, just the self-addressed envelope.

But even this is not the part I find most wonderful and freakiest of all. It is that this list of your own personal errors is neatly written out — yes, I mean by hand — on a sheet of lined paper, the three-hole looseleaf notebook kind. Amazing. Just amazing.

I finished this Friday’s Listener crossword, “Ballad” by Elgin, late Saturday morning after a longer struggle than usual.

It’s a tough puzzle. The across clues aren’t given entry numbers for the grid, and they are listed not in order of their appearance but in the alphabetical order of their answers. Now, this actually helps a lot in solving them: You can figure that a clue, say, about a third of the way down in the list probably starts with a letter around F or G or H. Then as you find some of the answers, that helps narrow down the possibilities for the rest; if one answer is GENERAL and the answer two clues down is GOLDFISH, then you know that the answer to the clue between them starts with a G and the second letter is somewhere between E and O, which probably means GE, GH, GI, GL, GN, or GO. So even without any help from crossing letters, the possibilities for the across clues narrow pretty rapidly, and it wasn’t all that long before I had close to half of them solved.

However, you could solve all the across answers and still not know where to put any of them in the grid. For that, you need help from some down clues. Down clues are ordered normally, so once you solve one you know exactly where in the grid to place it. But you don’t get any help from crossing letters till you’ve placed a few of the across answers, and you can’t place any of the across answers till you’ve solved some of the down answers, so you have to solve at least a handful of the down clues without any help from crossing letters before you can start filling in any of the grid.

But on top of that, the across answers may be entered either left to right or right to left, and there are no vertical bars to show where across entries begin and end, so that has to be deduced as you go along, and there are additional strange things in the instructions about a missing column in the grid and two across answers that have to be entered overlapping and some unknown number of across answers that aren’t entered in the grid at all, and these things also have to be figured out as you go along.

And on top of that, the clues are a lot harder than usual, though inventive and fair, I’d say. So all in all I found filling the grid to be a slo-o-o-o-ow, gradual process.

I didn’t figure out what the theme of the puzzle was until I’d just about finished filling the grid. But it’s a delightful surprise to discover the theme and see how it is worked into the puzzle in several amusing ways, justifying all the odd things about the puzzle. Everything comes together very satisfyingly. All in all, a tough puzzle, but worth the struggle.

There are a few clues where I know from the completed grid that I must have the right answer but I still don’t understand the wordplay. For another clue, I think I understand the wordplay but the clue seems to involve an alternative spelling that, while perfectly familiar to me, is nevertheless not given in Chambers Dictionary, or at least in the iPhone version (the only version that I have the latest edition in). That seems like it shouldn’t be kosher according to the rules of the Listener puzzle, but I’m not really sure.

This morning I finally finished this week’s Listener puzzle, “Sinecure” by Franc. Not a very hard puzzle at all, really, but I got a late start on Friday and then was busy with other things much of Saturday. Still, by bedtime on Saturday I’d solved everything except for unscrambling the ten-letter theme word. Turned out I didn’t have quite the right combination of letters. This morning I spotted where I’d made an unjustified assumption and got the right combination and found the word. I’d never heard of it, but it was a nice surprise to discover that it meant exactly what I’d suspected it had to mean, given the rest of the puzzle. (Turned out that it was a familiar word to Dave, though.)

Woo hoo! My Chambers Dictionary iPhone app is updating free to the new 12th edition! I wasn’t expecting that — if you buy a print dictionary, you don’t get a free update to a new edition, so I was just figuring they’d bring out a separate app for the new edition.

As I cannot even remotely afford to buy the new edition at the moment, and as I don’t seem destined ever to have any of my entries drawn in the weekly competition for the Listener crossword, this is very welcome. Thank you, Chambers people!

Despite my starting late in the day and some of the most impossible looking instructions I’ve ever seen, this week’s Listener crossword, “Confused” by Tea Leaves, turned out to be easier than it looks. After seven or eight clues solved, I started figuring out what was going on and finished before midnight. Very nice puzzle with several sweet surprises along the way.