A short, slightly chaotic history of the crossword

The crossword is the product of an editorial mistake. On Sunday, December 21st, 1913, the editor of the New York World’s Sunday magazine — a man named Arthur Wynne — had a small hole on page 26 to fill. He filled it with a diamond-shaped grid he’d sketched the previous afternoon, called it a “Word-Cross,” and ran it as a one-week novelty.

The next week, his postbag was full of complaints. Not because readers hated it. Because they wanted another one. Wynne, who had been planning to never run a Word-Cross again, ran another one. The format took off so fast that within a few weeks the printer had transposed the words “Word” and “Cross,” producing the now-permanent name crossword, and Wynne — a Liverpudlian who’d emigrated to America in 1905 — found himself the unwilling architect of an entire genre.

That, in 200 words, is most of how this happened. The rest is detail.

The 1924 panic

By the early 1920s, the crossword was a national phenomenon. American newspapers had begun running them daily. Then Simon & Schuster did something that, in retrospect, was either a stroke of genius or a desperate gamble: in 1924 they published The Cross Word Puzzle Book, a paperback collection of 50 grids with a tiny pencil attached.

The book sold over 400,000 copies in its first year. The publishing house, then less than a year old, was financially saved by it. (Simon & Schuster, eventually owners of an enormous chunk of American literary history, was technically founded to publish crossword puzzles.) Within months, every major paper in the country was running daily crosswords. Most still are.

The 1924 boom also produced the genre’s first moral panic. Newspaper columnists wrote serious pieces — not satirically — about whether “the crossword craze” was destroying productivity, ruining marriages, and corrupting the youth. The Times of London ran an editorial calling the crossword “a menace because it is making devastating inroads on the working hours of every rank of society.” This was published in 1924. It is approximately the same article every news outlet later wrote about television, video games, and Wordle.

Margaret Farrar invents the modern crossword

Wynne built the form. Margaret Farrar built the rules.

Farrar, a former secretary at the New York World who had begun editing crosswords in the early 1920s, did three things that defined the format we now play. First, she standardized the grid: 15×15 squares, rotational symmetry, no two-letter words, every white square a part of both an across and a down word. Second, she wrote the first crossword “style guide” — a working set of rules constructors should follow when writing clues, including the now-foundational principle that a clue and its answer must always agree on part of speech. Third, in 1942, she persuaded the New York Times — which had spent two decades publicly disdaining the format as beneath its dignity — to launch a Sunday crossword.

The first NYT Sunday crossword ran on February 15th, 1942. Farrar edited it. She continued to edit it for the next 27 years. Her successor, Will Weng, ran the puzzle for nine years. Eugene T. Maleska followed for 16. Will Shortz — who is, as of this writing, on year 33 of editing the NYT crossword — has now held the job longer than any of his predecessors. We wrote a longer profile of Margaret Farrar here, because almost nobody outside the puzzle world knows her name.

The British divergence

Meanwhile, across the Atlantic, the crossword was becoming a different animal entirely. The British, who had imported the form via the Sunday Express in 1924, found the American style — straightforward definitions — dull. Within a few years, British puzzle editors had developed the cryptic crossword: clues that combine a definition and a piece of wordplay, usually disguised as a coherent English sentence.

The cryptic was popularized by a Cambridge classics graduate writing under the pen name “Torquemada” (after the Spanish inquisitor) who set puzzles for the Observer from 1926 until his death in 1939. Torquemada’s cryptics were so famously hard that the paper published “Torquemada answer columns” the following week, where readers could write in to argue about whether a particular clue had been fair. (Most weren’t.)

By the 1950s, every major British newspaper had a daily cryptic. The Times of London had one. The Guardian had one (eventually with the legendary “Araucaria” as its lead setter). The Telegraph had one. The form has not changed in any structural way since. Our beginner’s field guide to cryptics covers how to actually read one.

The dictionary problem

For most of the crossword’s 20th-century existence, every grid contained a small recurring vocabulary that exists almost nowhere else in English. Three-letter Greek goddesses. Four-letter Egyptian sun gods. The river ARNO. The poet POE. A specific kind of small Italian sandwich case (ETUI). Crosswordese.

This vocabulary was, for decades, a feature, not a bug. Constructors needed short common letter combinations to make the grid work, and editors built up a shared mental library of “grid glue”: words that fit specific letter patterns and could be reasonably clued for a general reader.

By the 1990s, crosswordese had become a problem. The puzzle vocabulary felt stale. Constructors were leaning on the same hundred words. Will Shortz, on taking over the NYT in 1993, made it a stated policy to reduce crosswordese, push for fresher fill, and aggressively age out clues that referenced obscure 19th-century opera characters. He largely succeeded — the modern NYT is dramatically less crosswordese-heavy than the 1985 version. But the genre still has a working sub-vocabulary of about fifty words you’ll see in nearly every grid. Our working glossary covers the worst offenders.

The digital fall, and the digital rise

For a stretch in the late 1990s and early 2000s, the crossword looked like it might genuinely die. Newspaper circulation collapsed. The audience for the print Sunday crossword was visibly aging. The first digital crossword apps — brittle, ugly, full of typos — were widely panned. Several major papers quietly dropped their daily puzzle.

Two things saved it. The first was the New York Times’ launch of NYT Crosswords as a paid digital product in 2009, which proved that a daily puzzle behind a paywall could not only survive but thrive. The second was the iPad, which turned out to be — in retrospect, almost obviously — the perfect crossword surface. Big enough to show a real grid. Small enough to hold over breakfast. Easy enough for older solvers to use without help.

By 2015 the NYT Games product was profitable. By 2018 it was a major engine of subscription growth for the paper as a whole. By 2022, after the acquisition of Wordle for what was variously reported as “seven figures” and “a lot,” the NYT Games division had become — and remains — one of the largest standalone games products in the world by daily user count.

The themed-puzzle revolution

In parallel with the NYT’s digital rise, a different revolution was happening on phones. Brazilian developer Fanatee launched CodyCross in 2017 as a themed crossword wrapped in a science-fiction frame. It was an unusual product: free to play, generous with hints, structured around hundreds of themed packs. By 2020 it had passed 100 million downloads. By 2025 it had passed 200 million. It is now, by any measure, the most-downloaded crossword product in human history. Our 2026 review covers why it works.

CodyCross was not alone. 7 Little Words, from Maine-based Blue Ox Family Games, launched in 2011 and has been quietly running ever since — no microtransactions, no ads in the paid tier, no streak gamification. Themed Crossword Solver, Daily Themed Crossword, and a long tail of regional dailies have all carved out durable audiences.

The Wordle moment, and after

Then, in late 2021, a software engineer in Wales named Josh Wardle published Wordle as a free web game for his partner. By January 2022 it had millions of daily players. By February the New York Times had bought it. By summer 2022 it had spawned Connections, Strands, and the entire current ecosystem of NYT word puzzles.

The Wordle moment was, in retrospect, the resolution of a long crossword question. For decades, the genre’s critics had argued that the daily crossword was too hard, too long, too intimidating, and too connected to print culture to appeal to anyone under 40. Wordle answered all four objections at once. Five letters. Two minutes. No paywall. A new puzzle every day. The format that emerged in its wake — daily, brief, free, social — has become the default shape of the modern word puzzle.

It is also — and this is the surprising statistic from the long feature we ran on this last month — the most popular morning activity for Americans aged 18–29. The crossword is 113 years old, was nearly killed by digital, was nearly killed again by mobile, and currently dominates Gen Z’s morning attention.

What happens next

The honest answer is that no one knows. The genre has now died and revived three times in 113 years, and each revival has been triggered by a different technological shift: the daily newspaper, the paperback book, the iPad, the social-share screenshot. There is no obvious reason to think this pattern is going to stop.

What seems likely: more themed verticals (CodyCross-style, but in different cultural niches). More cross-language constructors. More LLM-assisted clue databases (we covered the early experiments here). More new daily formats from new players — Apple, LinkedIn, Spotify, all of which have launched or are reportedly developing daily-puzzle products.

What seems unlikely: that the basic shape of the experience changes. Wynne’s 1913 grid was rectangular, definition-driven, and ended in a small puzzle-completion moment. So is today’s NYT Mini, today’s CodyCross, today’s Wordle. The format works because the format works. It always has.

That’s the chaotic 113-year story. The next chapter is being written tomorrow morning, by whoever pulls out their phone in bed and taps a yellow square.