What Makes a Monday Crossword a Monday

A Monday crossword and a Saturday crossword are the same size. Same 15x15 grid, same rules, often a similar word count. One takes four minutes and one takes forty. The difference is not in the grid at all, and that surprises people who assume difficulty is a structural property.

I keep a spreadsheet of around four hundred published daily grids with their clues broken into categories. What follows is what falls out of it, with the caveat that this is one reader's dataset and not a peer-reviewed study.

The clue does almost all of the work

Take the entry OPAL. On a Monday it is clued as a gemstone. On a Saturday it might be clued through an obscure birthstone month, or a place name, or a piece of wordplay that never mentions gems at all.

Same four letters. Same crossings. The difficulty moved entirely into the clue, which is why constructors say the grid is fixed by the time difficulty is set. In the sample, clue directness - whether the clue gives a straightforward definition of the answer's primary sense - separates weekday tiers more cleanly than any structural measure.

Difficulty is not built into the grid. It is written into the clues, on the last day of the job.

Four dials that actually move

Definition directness. Early week clues define the answer's most common meaning. Late week clues define a secondary sense, or define the answer only obliquely.

Wordplay density. The proportion of clues doing something other than defining - puns, question marks, misdirection - rises steadily through the week. Monday grids in the sample average close to zero; Saturday grids are full of them.

Proper noun load. This one is contentious, because a proper noun is only hard if you do not happen to know it. But late-week puzzles lean on names from narrower domains, which raises variance more than average difficulty.

Entry familiarity. Late-week grids include more phrases, more multi-word entries and more contemporary slang. These are not obscure, but they are harder to retrieve from a partial pattern than a single dictionary word.

What does not change much

Word count barely moves for themed puzzles, because the theme dictates it. Block count barely moves. Average entry length moves a little, mostly because themeless late-week grids run longer stacks.

This is the useful finding for constructors: you cannot make a puzzle easier by adding black squares, and you cannot make it harder by removing them. You are adjusting texture, not difficulty. The corner design affects how a puzzle feels to start; it does not set how long the solve takes.

The Thursday exception

Thursday breaks the curve, and it does so on purpose. It is not simply harder than Wednesday - it is a different kind of task, usually because it contains a trick: a rebus, a grid that reads in an unexpected direction, or a rule that changes partway through.

In solve-time data Thursday shows a much wider spread than its neighbours. Solvers who spot the gimmick early finish faster than on a Wednesday; solvers who do not are still there an hour later. Average difficulty is a poor description of a bimodal day.

How to use this as a solver

If you are stuck on a late-week grid, the most productive assumption is that your reading of a clue is the intended misdirection, not the answer. Re-read the clue looking for a second part of speech: a word you took as a noun that could be a verb, a phrase that could be literal instead of figurative.

And if you are trying to get better generally, solve early-week puzzles for speed and late-week ones for practice. They train different things. Building the habit is easier when the puzzles come to you - a daily roundup of the major grids is a low-friction way to do it, and if you want to compare how one paper's difficulty curve differs from another's, the Wall Street Journal archive runs on a different weekly rhythm to the Times.

The reference literature is thinner than you would expect for a century-old format, but the general crossword entry and the grid-by-grid statistics at XWord Info are the two starting points worth knowing about, and between them they will let you check most of what I have said here against the record.