I have always enjoyed a game of Boggle. In my family I am banned from using a pencil while playing, because the manic scratching is off-putting to the other players. At a friend’s birthday party a couple of years ago, he asked me to bring a Boggle set and then forced me into taking on all comers throughout the night (I didn’t really need much forcing). Then, when Ned (friend of the Magpie) introduced a few of us to the Squaredle app – basically a Boggle game that has somehow avoided being taken down for copyright violation – it struck me that it would make a fun theme for a puzzle.

The big surprise was that Boggle hadn’t been used for a theme before, given how often a crossword grid has become a game board of some variety in the endgame. I always enjoy it when that happens (even if chess is perhaps a little overrepresented in such puzzles), so I decided to set to work on a Boggle puzzle, and hope that Ned hadn’t been similarly inspired – although perhaps a group of setters all writing a crossword with the same overall theme would be a fun challenge someday?

As Boggle is a 4×4 game, splitting up a 12×12 grid into sixteen 3×3 squares seemed like the obvious choice. So far as I can recall, the idea of using the mean letter from each square came pretty early on, too, even though it seemed like a daunting task even when I first considered it. Nothing ventured, nothing gained, though, so I set about seeing if it were possible.

First, I used Chambers dictionary to produce all the ‘common’ 16 letter words, and looked through the list for something with a reasonable set of letters – most importantly, no As and nothing too close to the end of the alphabet – and, ideally, some tangential connection to Boggle or the word ‘boggle’. There weren’t too many feasible options, but INCOMPREHENSIBLE fulfilled the first condition (at least, so I thought) and just about the second – as in, ‘the mind boggles’. There was a third condition, though: the word would have to be possible to create with an actual set of Boggle dice. That’s when I discovered that the letter distribution on the dice was completely changed in 1987, and I had to rule out some words because they didn’t work with the new set – but happily INCOMPREHENSIBLE worked in both the new and the old versions.

There is, no doubt, a smarter way to create the grid than to do it slowly in Excel, with formulae calculating the mean and conditional formatting letting me know when it came out as desired, but that was the method I adopted. The B seemed like it would be the most difficult of the dice to do, so I started with it in the top left hand corner and worked from there – alas, my first attempt fell down with only 5 out of 16 dice successfully completed. This was going to be tough. I won’t go into details on all the other failed attempts (which achieved 10, 10, 11, 10, 7, 9, 12, 6, 8, 10, 13, 12, 14, 14, 15 dice respectively), but they taught me two things: firstly, R & S were going to be every bit as hard as B, due to the unhelpful set of letters towards the end of the alphabet; and, secondly, I should avoid putting two difficult dice next to each other orthogonally. That’s because it’s basically impossible to get a set of words overhanging, say, both C and S dice, given the constraints placed on both of them. But H or L? Easy street.

The secret, in the end – other than perseverance – was to have at least one 3×3 square with lots of options coming into it: the P square had RA? entering it, giving me plenty of choices for the final letter that would allow me to balance it against other entries. The utter joy when I finally found a grid that worked was one of the greatest feelings I’ve had as a setter, and I’m fortunate that ‘THE THE’ struck me at the right time, given that it’s not in any of the wordlists I was using and it’s the only entry that would have made the final grid work. In the screenshot below, seeing the all-yellow conditional formatting – telling me that everything had worked – still brings me a warm feeling.

I found this the most difficult grid construction I’ve ever put together, so I hope solvers will forgive me for the prefix, abbreviation, and imperfect unching. At least I didn’t sacrifice grid symmetry, a cardinal sin in my eyes (and a conversation for another day).

It was clear that I’d need quite a long message to tell the solver everything they needed to do, so I chose a gimmick that gave two letters per clue, and the idea of shaking up words seemed nicely thematic, as well as being a bit different. Thank goodness there’s a word that begins with X and ends with L, even if it sticks out like a sore thumb. Finding the words to be jumbled meant a lot of searching through Chambers for words beginning and ending with the letters I needed, ideally words that weren’t too obscure, and then testing them one by one to see if they had any anagrams that wouldn’t leap out too badly in a clue.

In the original submission the message used ‘mean average’, a phrase I frequently used in my university days, but Jason suggested that this could be confusing when all it needed was ‘mean’, so I did some tweaking. This meant I lost my favourite jumble, a switch from CHARISMA to ARCHAISM, but also meant I switched out one or two of the weaker clues, so I was happy to do it.

Finally, spelling out BOGGLE in the grid seemed like an efficient way to confirm what was going on, and I was fortunate that it appeared in grid order already. I had intended to have each letter appear on a separate die, but it turned out I used the wrong L of RALLYE – I can’t remember if that was a deliberate decision to give myself more letters to use in the message (because the L that I ended up using was just in an across entry, whereas the other was also in a down entry) or simply an error. Let’s pretend it was the former.

Thanks as always to the editors (what a great bunch) for their work, to Ned for the inspiration and then test-solving, to my other test solvers, and to the people who sent feedback. I’m glad you liked it.

Posted on Monday, February 3rd, 2025 at 11:56 am under Setting the scene.
Leave a response