A comment from a solver prompted me to reply with some observations on writing clues for the monthly Azed clue writing contest; a further comment persuaded me that it would make sense to create a more permanent collection of tips which would encourage interaction and could be refined over time.
When I started submitting clues for the competitions, I thought I had a pretty good idea of how to write a clue, and I was terribly disappointed that my first few clues (all excellent, I felt) met with no success. Reading the monthly Azed slip, looking at the successful competition clues, and analysing Azed’s own clues in more detail, I realized that my clues weren’t actually that great, and many of them were unsound. Then I got my first VHC, and over the next few years I enjoyed a reasonable degree of success; although there have been occasions when I’ve asked myself ‘why didn’t he select my clue?’, in general when my offerings haven’t met with favour I’ve known in my heart that they didn’t deserve to.
How do you maximize your chances of getting a VHC or even a prize? Keep looking to improve your submissions and learn from your mistakes – the tips that follow may help a little. Do feel free to add your observations, to ask questions, and to disagree with me – I will be disappointed if the content of this page remains static. I will also be happy to give my honest opinion on clues that readers have submitted (or, indeed, not submitted) – either post them here in the comments or, if you would prefer that they were not seen by a wider audience, mail them to me using the address on the Contact page (only after the competition has closed, please). I might include such clues on this page if they illustrate a point, but I promise not to attribute them to an author.
Apart from point 1, these are not hard and fast rules, but they are based on my personal experience, both good and bad, of submitting competition clues.
1. Soundness of clues is absolutely crucial – Azed will not knowingly award a prize or a VHC to an unsound clue. Read through your clue as the wordplay is intended to work (rather than the surface reading) and make sure that it is grammatically correct when interpreted in that way.
Bad example: “I am embraced by crazy spinster” for MAID. As soon as you start to read the wordplay, you will see that it would have to start “I is…” in order to be sound. The easy fix is to use the future tense, ie “I’ll be embraced by crazy spinster”.
2. Writing a competition clue is quite different from writing a clue for a crossword, because Azed is going to get 100+ clues for the same word, and if there is (for instance) an obvious anagram then he is quickly going to get fed up with seeing it used time and again (as it undoubtedly will be). Also, while brevity might be prized in crossword clues, a two-word double definition clue is very likely to be used by other competitors, while relatively wordy clues still have a good chance of success (see tip 11), . Try to come up with an original idea, because that will get his attention.
From the slip for comp 1752: “MIDWINTER produced a nicely varied crop of entries. Sadly, the highly appropriate anagram of ‘wind’, ‘rime’ and ‘t’ (usually for ‘time’) proved so popular that none who used it managed higher than an HC (hence the extra-long list). It always seems unfair when this happens, but I don’t think there’s much I can do about it. There are always other approaches to try if you suspect that your idea, however neat, may have occurred to lots of others.”
3. Avoid the obvious definitions where possible. A look through the &lit archive will show that many of the successful entries employ definitions which are cryptic or, at least, unusual (along the lines of Azed’s use of ‘snob’ to indicate a shoemaker).
Good example: R J Hooper’s “It’s looking like a wet weekend with chill to follow” for HANGDOG [HANG + DOG]
4. A bit of deception is good (eg a noun masquerading as a different part of speech in the surface reading), and a well-disguised break between definition and wordplay is another plus.
Good example: L M Inman’s “Lying parallel, chief cause of congestion?” for DOUBLE-PARKING [DOUBLE PAR KING, ‘lying’ = participle verb/adjective, ‘chief’ = adjective/noun]
5. Plenty of successful clues, particularly &lit ones, use single letter indicators (eg ‘start of…’), but if you are going to employ this device then (i) make sure that it is sound – “Bull’s Head” is fine for B but “Beachy Head” isn’t, and (ii) make sure that you get full value from it, as in, say, ‘too close for comfort’ for OVERT.
Good (actually, very good indeed) example: R J Heald’s “Take the lead in Cinderella, playing girl who works in rags” for HACKETTE [(TAKE THE C(inderella)*. This ticks many boxes, in particular point 3]
From the slip for comp 1676: “We all, myself as much as anyone, include references to letters in initial, medial, final, alternate, etc positions, and these are well-established conventions, but I do regard them as relatively weak and to be resorted to when nothing else seems acceptable.
A look at some of the other examples on this page may suggest that Azed’s view on this point has softened, but you will also see that in all instances the letter selection elements are seamlessly incorporated into the clue as a whole.
6. Avoid noun anagram indicators on their own, so “Eton mess” is not acceptable for NOTE, although “Truss in a mess” would be ok for RUSTS.
7. Remember that definitions by example (eg ‘setter’ for DOG or ‘Trotter’ for DEL) must always be indicated by a qualifier such as ‘perhaps’ or a question mark.
Good example: Dr I S Fletcher’s “Torte without filling can upset Bunter, perhaps” for VALET [(T(ort)E LAV)<, Lord Peter Wimsey’s valet – “Bunter?” would also be acceptable]
8. If the competition clue word is an across word, it must be clued as such (so ‘rising’ as a reversal indicator would be no good), similarly if it is a down word then ‘from the east’ wouldn’t work for reversal. If the competition word doesn’t appear in the grid (this may happen in certain ‘specials’), Azed is happy for it to be clued as either an across or a down entry.
9. Try not to make clues too easy.
From the slip for comp 1908: “Perhaps the biggest problem of all was to come up with a clue which didn’t scream the answer out loud – even some of those quoted [as VHCs] came close to this. I have no objection to easy clues, but do beware of those which present the solver with no challenge at all.”
In practice, this only applies to simple clue types, such as ‘hiddens’ and ‘take the first letters’. Azed doesn’t seem to mind if the answer is obvious when the clue is more involved.
Example: Dr S J Shaw’s “Smash US TV series heaped woe for such housewives” for DESPERATE [comp anag &lit, (US TV SERIES HEAPED WOE)* = (DESPERATE HOUSEWIVES)*, the answer here being more obvious than its cryptic origin]
10. Very short clues can be successful, but they often fall foul of tip 2. In recent times, particularly, Azed has often favoured clues of considerable length, a quick check on the archive showing that prizewinning clues of 10 words or more are surprisingly common even for target words of four or five letters. Azed likes clues which tell a story, and the number of words (within reason, 15 probably being the upper limit) is not an issue as long as they all have a part to play.
Example (good or bad? – you decide): D F Manley’s “What hand is raised for – needing pee, about to burst, wanting the loo ultimately?” for UPBEAT [(PEE ABOUT – (th)E (lo)O)*]
11. You will see that many of the prizewinning clues are &lit (all-in-one) clues. Some words lend themselves better to this type of clue than others, and there is no doubt that a good &lit clue will beat a similar clue which includes a separate definition. You will also see that often the ‘definition’ is fairly loose – a degree of additional latitude is given to this sort of clue. But producing a good (and sound) &lit is far from easy, and you will find that the published &lit clues have usually been produced by experienced competitors – that is no reason not to attempt one, but here you need to read your clue carefully and ask yourself whether as a whole it is valid both as a ‘definition’ of the solution and a wordplay that leads to it.
Great example: N C Dexter’s “Item gran arranged family slides in?” for MAGIC LANTERN [CLAN in (ITEM GRAN)*, &lit]
12. Likewise compound anagrams (particularly those which are also &lit) often feature among the winners. Probably the main reason for this is the considerable range of additional treatments afforded by this type of clue (greatly increasing the chance of a unique clue as suggested by tip 2), but again these are not simple to write. Remember that the wordplay needs to indicate in some way that the solution (in a comp anag &lit often represented by ‘this’, ‘what’ etc) plus some other letters can be rearranged to form another collection of letters.
Good example: C G Millin’s “Nasty innuendos could be this – also noun” for SNIDE [comp anag &lit, INNUENDOS* = (SNIDE + NOUN)* – note the use of ‘nasty’ and ‘could’]
13. Clues which relate to a recent event (a scandal, an election, or both, say) are often successful. Azed does not care whether the surface reading of the clue will make any sense in a year’s time. Also, Azed is no prude, and while Viz-style obscenity – or anything likely to give offence – should be avoided, a bit of naughtiness in a clue certainly won’t damage its chances.
Example: M Hodgkin’s “Bojo’s beloved going nuts about bride-chamber at Number Ten – outdated stuff!” for BOMBAST [BO(jo) + (bride-cham)B(er)) in MAST]
Example: J C Leyland’s “How intimate embraces lead to rakish Don Giovanni maybe getting end away?” for IMPROPERLY [(R(akish) OPER(a)) in IMPLY, &lit, note the ‘maybe’ for the definition by example of OPERA]
14. As witnessed by Azed’s own clues and many successful entries, humour is a prized commodity. This may or may not be combined with topicality or a spot of mild smut.
Good example: N G Shippobotham’s “Trimmed pet’s hair, wanting a cool cat” for HIPSTER [((PETS HAIR) – A)*]
15. Clues which refer to Azed puzzles and competitions often do well, and not just in puzzles celebrating an Azed landmark.
One of many examples: M Barley’s “Entries to Azed get brain working a good deal” for BARGAIN [(A(zed) G(et) BRAIN)*]
16. Be wary about using letters from the NATO phonetic alphabet (eg ‘Mike’ for M) – Azed has written in the past about his unfamiliarity with them, and while, as correspondent Azedophile has pointed out, Bravo, Charlie, Echo, Golf, Papa, Romeo, Tango and Victor have all appeared in published clues there is evidence that he has rejected other letters from the alphabet in the past, so I would advise sticking to those eight
17. Azed quite often gives reasons in the slip as to why he’s rejected or marked down particular clues – this can give a good idea of other things to avoid.
From the slip for comp 1385, PANTRIES: “Four examples of unsoundness from quite experienced competitors: 1) ‘Botched repaints? Sounds like Ladas.’ This is an instance of ‘a clue to a clue’, as Ximenes would have called it, with no actual definition part at all. (The fact that it’s not difficult to solve is irrelevant.) 2) ‘Twist – near pits – presses for food.’ Nice idea, with its misleading reference to Oliver T., but the first dash spoils it for me. For the anagram to work you have to assume it isn’t there, but it is, and you can’t just wish it away. 3) ‘Where, each being shelved, peach tins are arranged.’ Again a promising idea, with the nice double meaning of ‘shelved’, but for it to work grammatically as an ‘& lit.’ it needs another ‘are’ or an ‘is’ after ‘are’. 4) ‘Cake goes in these.’ A similar attempt at an ‘& lit.’, similarly flawed but much more easily salvaged. The syntax doesn’t work in the cryptic reading, though ‘You’ll find cake goes in these’ would be fine and entirely acceptable.”
178 If you plan to use a particular indicator, say ‘riotous’ to indicate an anagram, but are not sure whether it is likely to be accepted, go to andlit.org.uk and do a keyword search (in this example, for ‘riotous’) on Azed clues. If the indicator has been used in successful clues for the purpose to which you intend to put it, then all is well; if there is no precedent, the fact that you were already in enough doubt to check probably suggests that you should try something else.
19. One final point based on bitter personal experience – make sure there are no typos in the clue you submit. There are few things worse than having your clue returned to you by Azed with a mistyped word circled in red pen!
Really helpful article, thanks. I’ve submitted 7 or 8 clues over the past 18 months or so, and have been very happy to receive a couple of VHCs. But I find it almost impossible to predict which ones will do well. The one I liked best, which Azed evidently didn’t, was for PEANUTS:
With unbridled flair, showman finally dropped underpants performing comic strip (7) (UNDERPANTS – DRN)*
I suspect this failed on account of the deletion of the final letters. They can be deleted in the order I generated them but maybe the presence of the comma was the problem and/or the absence of an “and”? I’d be interested in your opinion.
Hi Simon, and welcome to the blog. I’m afraid that your comment got flagged as spam (I’ll try to find out why) and I’ve only just found it – sorry. You are now on the ‘approved list’, so any future postings should appear on the site immediately.
Regarding your PEANUTS clue, I don’t see anything that would have disqualified it out of hand. When non-consecutive letters are to be deleted from anagram fodder, it’s usual to provide a second anagram indicator, but Mark Barley’s VHC clue for CABRETTA, ‘Arrange to rent a cab, saving on shoe-leather’ [(TO RENT A CAB – ON)*] got through. In the cryptic reading, the placement of the comma is not ideal – it would sit better between ‘dropped’ and ‘underpants’ – but again I don’t think it’s a showstopper. I would say, though, that Azed does on occasion allow things in clues from established competitors that strike me as questionable, which is why I advise relatively new entrants to make their clues ‘ultrasound’ wherever possible.
I think that with just one judge there are going to be occasions when a clue just doesn’t appeal to them for whatever reason, although I’d have to say that there have been one or two clues of mine where I would have liked to ask Azed “What didn’t you like about it?”
Congratulations on getting another VHC for your GASTROSOPHER clue.
Update: I found the issue in the spam filter – it has now been dealt with.
Many thanks for the hints and tips. I aspire to sending in an entry once a month, but I’m not there yet. I am in the “keen but inept “ section of this class. Must try harder. Keep at it, that is you and me both!
I think that most successful entrants would admit that they were pretty inept when they started out, even if they didn’t realize it at the time. The most important thing by far is to read your proposed clue through and check in particular that:
– The definition can reasonably indicate the solution (eg if the solution is an adverb, a definition that indicates an adjective is no good)
– The cryptic reading of the wordplay is grammatically sound (avoid classic traps such as ‘I am entertained by fellow’ for MAIN, where ‘I is’ would be required. “I’ll be entertained by fellow” is fine)
– The elements of the wordplay are valid (eg ‘May first’ cannot indicated the first letter of ‘May’. ‘First of May ‘is good)
– The elements of the wordplay can combine to form the solution (eg “certain about guy’s backing” can’t give SURNAME, because the whole thing ‘is backing’. ‘Certain about guy coming back’ works for SURNAME)
– There is nothing in the clue that shouldn’t be there (eg ‘the’ in ‘bent the rules’ for LURES must be removed, as it isn’t part of the anagram fodder)
If you can regularly turn out sound clues, it won’t be long before you start getting some success.
I saw a ‘Beachy Head’ for B in the Sunday Times clue writing contest recently which had me wondering if it was acceptable after all (if not to Azed). Could ‘river head’ (the source of a river) provide cover for using ‘head’ this way? And ‘rivet head’ (the head of a rivet)? Both are in Chambers.
Hi Chris, welcome to the blog, and thanks for your comment
First of all, my apologies – both your comments got picked up as spam, and I only found them when I was doing a final check prior to clearing out the bin. It’s the first time that’s happened for a legitimate comment, so I’ll try to find out what set the spam filter off. You’re now on the ‘approved’ list anyway.
There is, of course, no right answer to the question. I think that the most important factors are the expectations and reactions of the audience – regular solvers of certain puzzles, such as the Guardian back-pagers, will be on the lookout for anything that might suggest the first/last/middle letter of a word, even perhaps part of a word, as in ‘lionheart’ for IO. They are probably not over-fussed about the precise grammatical accuracy of the clue. I may not like this, but if the mass of solvers consider it fair then – frankly – it’s none of my business.
Azed and I are in the ‘sticklers for accuracy’ camp, therefore since ‘Beachy Head’ is certainly not the ‘head of Beachy’ we would both consider that unacceptable. ‘River head’ and ‘rivet head’, however, occupy an area of a greyer shade. A ‘rivet head’ is the head of a rivet, but ‘rivet’ here is a noun which is being used attributively – a ‘rivet head’ and a “rivet’s head” are not quite the same thing, but does that matter? As with ‘Beachy Head’, I would reject ‘Eton mess’ out of hand for TONE (it isn’t a mess of Eton), but what about ‘flower arrangement’ for FOWLER – a ‘flower arrangement’ is definitely an arrangement of flowers, though normally not just a single bloom. ‘Mango pickle’ for AMONG has perhaps even more going for it.
The one thing I can say for sure is that Azed competitors should avoid noun indicators such as these without prepositions or apostrophes when writing their clues (and when submitting puzzles for the Listener crossword, but even then they’ll just get changed, it won’t cause the whole puzzle to be rejected). I reckon that when it comes to blocked puzzles you would probably get away with most, if not all, of the examples above.
Thanks for your fascinating insights into Azed’s thinking regarding clue competitions. I’m not sure that I agree with you about your point 6 however – I see no real evidence of his aversion to letters from the NATO phonetic alphabet. Working from memory and scanning the Azed archive, I was able to adduce the following examples: B=Bravo in a VHC for NOVE(M)BER 1277 (as well in fact as O=Oscar), C=Charlie in a VHC for CHALET 1286, E=Echo in a second prize for PARTERRE 1771, G=Golf in a VHC for FAIRY GOLD 1871, P=Papa in a VHC for MONTE DI PIETA 2438, R=Romeo in a VHC for HEART 1797 and for PROPOSAL 2738, T=Tango in a VHC for HACKETTE 2014 and V=Victor for PUBLIC SERVANT 2356. While the NATO phonetic letters are not listed alongside abbreviations in Chambers (which is perhaps slightly curious), reference is made to their use as code words in international radio communication eg the listing for Bravo cites its use as a code word for the letter b. As citation in Chambers is often the main or only yardstick by which Azed judges the acceptability of words, it would seem illogical for him to reject their use in his clue competitions. I think from a clue writing standpoint it’s definitely useful to have them available as potential clue components (it’s curious that we’ve not yet seen F=FOXTROT, K=KILO or U=UNIFORM which might each come in handy on different occasions!)
Thanks, Azedophile – just the sort of feedback I was looking for. My observation was clearly too bald a statement, however I think that one should still be wary. In the slip for 1680, he wrote:
Interestingly, a number of clues submitted used w = whisk(e)y, I assume from the NATO phonetic alphabet. This alphabet is not given in Chambers and I can’t find it in other dictionaries of comparable size, so I’m not too happy about allowing it (as distinct from, say, the IVR abbreviations, which are all in C). Eyebrows would, I’m sure, be raised if I accepted, for example, L for Lima, M for Mike or P for Papa.
The next slip included:
As several of you pointed out, I was mistaken in saying last month that the NATO phonetic alphabet is not given in Chambers. I didn’t think to look up the words which are used to represent their initial letters, which is where they are all shown. It still seems odd to me that the lexicographers chose to deal with them in this way without at the same time including them as abbreviations at the single-letter entries. But I must clearly overcome my reluctance to use them myself and allow you to use them likewise.
I’m pretty sure that he is not familiar with them all, which may explain why, say, ‘uniform’ has not appeared in a published clue. I strongly suspect that ‘Mike’ may also be an issue, which is why ‘Mike Pence’ has never been seen for MP. I have updated the observation accordingly (and dropped it down the list).
Many thanks for the analysis and comments. Very useful. I’ve only been solving AZED puzzles for about 18 months now and have recently had a first HC in the competitions. The point about the definite article was one I hadn’t picked up and will probably sink my latest entry too, as well as the fact that it’s pretty long! Thanks again. Mike
There are times when the definite article is acceptable in a wordplay element, in particular where it is linked directly to a noun in the cryptic reading, so ‘the end of March’ is fine for H, but ‘the Wild West’ is no good for STEW. When in doubt, though, leave it out!
There is no issue per se with long clues – and a number of recent prize winners have been decidedly wordy – as long as all the words play at least some part and a ‘thread’ runs through the whole thing. However, I would advise relative newcomers to keep clues succinct – to put it bluntly, your name isn’t going to catch Azed’s eye, so your clue needs to. For examples of how to make every word in a clue count, go to andlit.org.uk, select ‘Browse Archive’, ‘Find a competitor by surname’, choose H and then R.J. (Richard) Heald – you will also see how well his clues address the points on my list, particularly the first three.
An HC (as well as your PERDU clue, which had a lot more good about it than bad) clearly indicates that you’re on the right track, and getting a clue into the published list is just a matter of time – I can assure you that it’s well worth the wait!
Many thanks for this informative article. I would welcome any feedback on my recent clue for PERDU and why it may have fallen short!
9 Initially, Rishi’s surrounded by the bumbling, mostly duped, old guard, with little hope of success (5)
Definition: old (obsolete) guard, with little hope of success
R (Initially, Rishi) inside (is surrounded by) anagram (bumbling) of DUPE (mostly duped)
Hi Mike, and thanks for submitting your clue for public scrutiny!
I like the definition ‘old guard’, a nice bit of misdirection, and the comma after ‘guard’ means that the ‘old’ can also qualify your second obsolete definition, ‘with little hope of success’.
The biggest problem I see here is the second wordplay element, ‘the bumbling, mostly duped’. Firstly, the word ‘the’ is redundant – if you look at the successful clues, you will see that the definite article only appears when it has a role in the wordplay (eg ‘the French’ for LA). The comma between ‘bumbling and ‘mostly’ is of course required by the surface reading, but punctuation is equally important when looking at the cryptic interpretation – this could legitimately indicate an anagram of THE, but not of DUPE(d). Applying reduction and anagram indicators to the same word can be tricky, although ‘[surrounded by] mostly duped, bumbling [old guard]’ would be fine here. That said, the elements of the wordplay are a little too honest, with the surface and cryptic readings being very similar. I would try to avoid the last letter removal if at all possible, while adding a little more deviousness, with something like ‘[Rishi’s] brought in to dupe bumbling [old guard]’. The ‘old’ also makes me think that ‘doddering’ might chime slightly better than ‘bumbling’.
The second definition is a nice idea, but I’m not sure whether the additional five words justify their inclusion (that is very much a personal view, though). I realize that I’ve changed the sense of the surface reading, but I would have shaped these basic ingredients into something along the lines of:
Rishi initially is brought in to dupe doddering old guard
This is absolutely sound, and the last five words could be reinstated (although he probably has every chance of achieving his revised target 🙂 ). In general, though, I would advise keeping clues relatively short and sweet, at least until you are on Azed’s ‘radar’.