When you stall: a repeatable order of attack
A half finished grid that has stopped moving is a structural problem, not a vocabulary one. Here is the checklist that unblocks it.
Every solver hits the same wall. Two thirds of the grid is filled, one corner is blank, and rereading the same four clues produces nothing. A stall is rarely a vocabulary problem. It is usually a structural one, and structural problems respond to a fixed procedure rather than to more staring.
What a stall actually is
A blocked corner has one of three causes. Either you are missing a piece of knowledge, or you have read a clue in the wrong sense, or you have written a wrong answer that is now poisoning every crossing. The third cause is by far the most common and the hardest to see, because the wrong answer usually felt obvious when you wrote it.
The procedure below works through the cheap causes first. Run it in order. Skipping to the interesting step is how solvers lose twenty minutes.
- Reread every unsolved clue in the corner, slowly, out loud if you can. Roughly one in four stalls ends here.
- Write out the letter pattern of each empty slot, for example
_ A _ E _, instead of holding it in your head. - Attack the shortest slot with the most crossing letters, not the clue you find most appealing.
- Test the standard endings and beginnings against each pattern.
- Doubt your most confident answer in that corner and remove it.
- Reread the clue in a different register: proper noun, abbreviation, phrase, pun.
- Put the puzzle down for an hour.
Write the pattern down
Holding _ A _ E _ in working memory while also generating candidates is more than most people can do. Written down, the same pattern becomes searchable: BAKER, WAGER, TAKEN, LATER, PAPER, CAMEO. Reading that list against the clue is a different mental task from inventing it, and it is much easier.
Do the same for the crossing entries. If a down answer you are sure of supplies the A, note which across entries depend on it. Those dependencies are what you will audit in step five.
Mine the endings and beginnings
English answers cluster at their edges. A slot ending in _ _ _ E D or _ _ _ E R is common; one ending _ _ _ I N G is common when the clue is in the present participle; _ _ _ E S T follows a superlative clue; _ _ _ L Y follows an adverb clue. At the front, RE, UN, IN, DIS and OVER cover a large share of six and seven letter answers whose clue contains the words again, not or too much.
These are not guesses. They are consequences of the grammar rules covered in reading the clue, and they turn an open slot into a short list. Very short slots have their own shortlist, the recurring fill covered earlier in this course, and a blocked three or four letter entry is very often one of those.
A stalled corner almost always contains one answer you are certain about. Certainty is the symptom, not the defence.
Doubt your best answer
When the cheap steps fail, assume you are wrong somewhere and go looking. The suspect is not the answer you hesitated over. It is the one you wrote instantly, because that is the one you have stopped checking.
Audit it properly. Does it match the clue in tense and number? Does it fit the enumeration exactly? Is there a second word with the same meaning and the same length, which is often the real trap: STARE and GLARE, ANGRY and IRATE, QUIET and STILL, SMALL and SHORT. Erase the suspect, leave the square blank, and solve the crossing entries without it. If they come, you found the blockage. The full diagnostic is set out in pencil discipline, and the wider working order in solve the crossings first.
Change the register
If the corner is still dead, the clue is probably not being read the way it was written. Run through the alternatives deliberately. Is the answer a proper noun, given that the capital letter at the start of a clue proves nothing? Is it an abbreviation, signalled by a shortened word in the clue? Is it a multi word phrase rather than a single word? Is the question mark at the end telling you the definition is a pun rather than a synonym? Terms such as register, fill and enumeration are defined in the glossary.
When to stop, and what to do next
Setting a puzzle aside for an hour is a technique, not a surrender. The effect is well documented in solving columns: the answer that would not come at eleven arrives in thirty seconds at noon, because the wrong reading has faded. The Wordplay column discusses this and the general anatomy of hard clues most weekdays.
When you finally check a finished puzzle, read the answers you missed rather than just marking them. Seeing how a clue was meant to be parsed is the part that transfers to tomorrow. Published grids such as the completed New York Times answer sheets are useful for that review, and the official archive at the New York Times crossword section lets you replay the same puzzle later from scratch.
Abbreviations: how a clue tells you the answer is shortened. The flags that tell you to write three letters instead of a word.
One clue technique, every Sunday
A short email with one device, three worked examples, and a puzzle to try.