Australian Crosswords

What do you think solvers complain about the most when it comes to cryptic crosswords? Obscure words? They are up there. Imprecise definitions? No-one is happy with these. Extraneous words? Less and less so. What about poor grid design? Almost never (but it should be scrutinised more). No, it’s none of these. In my experience at the top of the list is poor anagram indicators, probably because anagrams are frequently used and setters are looking for something different so as not to be obvious. Usually a setter will use more anagram indicators (in a full or partial mode) than the rest put together, so solvers shouldn’t be surprised when they explore the edges of what’s considered acceptable. However, the same rule applies no matter what anagram indicator is used: it must directly or indirectly point to a mixing up or rearrangement of letters. Some form of contact isn’t good enough, there must be change involved.
If an anagram indicator can only be deduced after the answer has been found and the rest of the wordplay resolved, then it is faulty; if a solver works out that part of the wordplay must be an anagram indicator based on its context and not on its implicit instruction, then it is also faulty.
All Stickler fans will know how much I like to use the anagram device either for the whole answer or, more commonly, as part of the answer. In over 100,000 Stickler clues I used a relatively small list of anagram indicators without ever being accused of repetition or producing puzzles that were too easy due to the number of anagrams I used. It’s also rare for anyone to question my anagram indicators. I put this down to context. Stick an anagram indicator in a clue where it seems to belong and the most common indicator may be glossed over – it’s human nature. So, if there is a secret, it’s this: don’t try and be tricky with an obscure anagram indicator, just use one that’s natural in the surface reading to get solvers believing it has a rightful, and not artificial, place in the clue.

It’s normal for new setters to mimic elements of the crosswords they solve the most, especially those crosswords appearing in major newspapers. They assume this is an acceptable approach as major publications would have style guides, editors, checkers etc to ensure a high standard of work. It’s a fair place to start, but over time setters must come to their own conclusions about what should and should not be used, and not necessarily accept that “because it’s published it must be Ok”. It’s really easy to accept crosswordese (regularly-used crossword elements that outsiders would find hard to understand) as without it setters have less to work with, however, I think it’s important to think through everything you sign your name to as a setter.
Early on I was happy to use anagram indicators like “perhaps” and “possibly” without thinking: I’d seen them used 100s of times and come to accept their roles in clues. However, during cryptic crossword solving classes at my local community college, I was asked how “perhaps” could be seen as an anagram indicator. I had no answer except to say that a lot is implied and solvers knew what it meant when they saw it. Such things frighten new/learning solvers (“what other secret codes do I have to learn?”) as they can’t necessarily rely on a description of an anagram that says “any word or phrase that suggests a rearrangement (of letters)”. The upshot? While I have used them in extreme circumstances, I no longer have them on my anagram list. “out” is out for another reason: you can find a dictionary definition that could make it an anagram indicator, but the average person has trouble thinking of it that way.
Should I reinforce the exclusive nature of cryptic solving by using such devices, potentially locking out new blood? I say no.

Recently I asked if people could spot what I DON’T DO in my crosswords. (My list covers the whole cryptic crossword spectrum, including the entire crossword, grid, types of clues and devices.) One poster suggested that I didn’t seem to use clue cross-referencing, and too some extent that’s true. As with all things regarding my crosswords, I have carefully considered my use of clue cross-referencing.
I don’t set out to create 15×15 regularly-published crosswords with a theme that would involve many clues carrying references to other clues. Why? I don’t like solving that style of crossword myself. If I pick up such a crossword to solve (even by a setter I like) and see heavy cross-referencing, I won’t even start it. (A particular dislike is a clue that refers to another clue that in turn refers to another.) However my regular crosswords may contain either an overt or covert cross-reference if the opportunity arises during the clueing phase – I don’t set out to make it happen.
I have done larger holiday puzzles with themes and those have contained cross-referencing. Larger puzzles can swallow a theme more easily and any cross-referencing doesn’t tend to dominate the overall look of the puzzles. When I do this, I tend to only use one answer in the cross-referencing so the solver doesn’t have keep flitting around the crossword. I have one other “rule” when using cross-referencing: any clue that uses it must still have good surface reading when the answer from referenced clue is substituted in.