A short, slightly chaotic history of the crossword

The first crossword puzzle was a mistake. Or an accident. Or, depending on who you ask, a desperate piece of holiday filler. What it definitely was not, in 1913, was the start of the most successful new-media format of the entire 20th century. But that’s what it became.

December 1913: The hole in the magazine

Arthur Wynne was a 42-year-old Liverpudlian journalist working for the New York World. The World, in 1913, was running a Sunday magazine called Fun, and Wynne was responsible for filling its puzzle page each week with riddles, anagrams, and word squares. The week of December 21st, 1913, he needed something new.

What he came up with was a diamond-shaped grid of 32 numbered squares, each accompanied by a clue. He called it a “Word-Cross.” It ran on December 21st under the headline “FUN’S Word-Cross Puzzle.” Two weeks later, a typesetter accidentally swapped the words around — and the format had its real name: “Cross-Word.”

The first puzzle’s difficulty was, by modern standards, almost insulting. The clues included “What we all should be (KIND)” and “The close of day (EVE).” But it sold newspapers. The mailbag the following week was unusually heavy. The World ran another one. Then another. Within a year, the crossword was a regular Sunday feature.

1924: The book that changed everything

For ten years, the crossword was a curiosity — a thing the New York World did, that other newspapers occasionally copied, that didn’t really exist outside the Sunday supplement. Then, in April 1924, two enterprising young publishers named Richard Simon and Max Schuster founded a new company and made a bet: that someone might pay 1.35 dollars for a book of crossword puzzles.

They called it The Cross Word Puzzle Book, included a pencil with each copy, and printed an initial run of 3,600 books. It sold out in three weeks. The next print run sold 40,000. By the end of 1924, Simon & Schuster had sold over 750,000 copies of crossword puzzle books and the entire United States was, briefly, completely insane about crosswords.

The Boston Public Library temporarily banned crossword reference books because they were tying up the dictionaries. Pittsburgh trains posted dictionaries in the dining cars. The Baltimore & Ohio Railroad outfitted a few cars with crossword-puzzle stations. Universities added crossword tournaments. There were Broadway shows about crosswords. There were, briefly, crossword fashion accessories.

1942: The Times resists, then surrenders

The New York Times, characteristically, was late to the form. The paper publicly disdained crosswords as “a primitive sort of mental exercise” well into the 1930s, and refused to run one as a matter of editorial principle. In 1924, an editorial in the Times specifically singled out the crossword as “a sinful waste in the utterly futile finding of words the letters of which will fit into a prearranged pattern.”

Then came Pearl Harbor. With the United States at war and newspaper readers in obvious need of distraction, the Times reconsidered. On February 15th, 1942, the paper ran its first Sunday crossword. The editor was a 44-year-old Connecticut woman named Margaret Petherbridge Farrar, who had spent the previous 16 years editing crosswords for Simon & Schuster.

Farrar’s contribution to the form is so significant that she gets her own profile, but the short version is: she invented almost every modern crossword convention. Symmetrical grids. The Monday-easy-to-Saturday-hard difficulty curve. The rule against unkeyed letters. The rule against two-letter answers. The convention that themes go on Sunday. The basic editorial principle that a clue can be hard but must be fair. She did all of this, and she did it before the puzzle had a permanent editor at almost any major American newspaper.

The American grid vs. the British grid

Almost as soon as the crossword crossed the Atlantic to Britain in 1922, the form forked. American grids developed toward maximum interconnection — every letter checked by both an across and a down clue — and toward straightforward definitional clues. British grids went the other way. They allowed unkeyed letters. They embraced wordplay over definition. By the 1930s, the British puzzle had become something genuinely different: the cryptic.

The cryptic is, as a friend of mine once put it, “a crossword that wants to fight you.” Each clue contains both a definition and a piece of wordplay (an anagram, a hidden word, a charade), and the solver’s job is to figure out where one ends and the other begins. The form was popularized by a series of constructors writing under pseudonyms in The Times of London, The Guardian, and The Observer.

For most of the 20th century, the two traditions barely spoke to each other. Americans considered British cryptics impossibly cruel; Britons considered American crosswords trivially easy. Both views are slightly fair. Both are slightly wrong. We have a beginner’s field guide for Americans curious about the cryptic form if you want to investigate.

1993: Will Shortz and the modern era

For 50 years after Margaret Farrar took the editor’s chair at the New York Times, the puzzle was edited by a small succession of capable but conservative editors who maintained the form roughly as Farrar had left it. The puzzles were good. They were also predictable. By the late 1980s, a generation of younger constructors — raised on cryptic puzzles, on rebus tricks, on themed Sundays that bent the rules — was getting frustrated.

In 1993, the Times hired Will Shortz, a 40-year-old former editor of Games magazine, to take over the crossword. Shortz was the first crossword editor in American history to hold a college degree in enigmatology — a degree he had personally invented at Indiana University, because no other university offered one.

What Shortz did to the New York Times Crossword over the next 30 years has been written about extensively elsewhere. The short version: he introduced rebus squares, which had been nearly absent from the American grid. He encouraged constructors to break the symmetry rule when a theme demanded it. He published the first wave of women, queer, and non-white constructors that the form had ever really seen. He turned the Sunday puzzle from a serious-but-staid weekly fixture into the weekly cultural event it is now.

He also presided, in 2014, over the launch of the NYT Mini — the 5×5 daily that almost certainly did more than any other single product to bring the crossword to people under 35.

The mobile decade: 2014–2024

The crossword, as a format, was born on paper, lived on paper for 100 years, and then — in the span of about a decade — moved almost entirely onto phones. The transition was led, somewhat unexpectedly, by themed crosswords from outside the major newspapers. CodyCross, from Brazilian developer Fanatee, launched in 2017 and proved that themed grids worked beautifully on a portrait phone screen. 7 Little Words, from Maine-based Blue Ox Family Games, had been on phones since 2011 and had quietly built one of the most durable daily-puzzle communities on the App Store.

The NYT Mini, the WSJ Daily, and the LA Times Crossword all moved to mobile in roughly the same window. By 2020, the NYT Games app had over 5 million subscribers. By 2025, it had 11 million. By 2026, the New York Times reported that Games revenue had surpassed Cooking and was second only to News in the company’s digital portfolio.

2022: Wordle

The single most important thing to happen to the crossword family in the 21st century did not, technically, happen to the crossword. It happened to a tiny word game built by a software engineer named Josh Wardle for his partner.

Wardle launched Wordle to the public in October 2021 with no marketing, no plans, and no monetization. By December, three million people were playing it daily. In January 2022, the New York Times acquired it for an undisclosed seven-figure sum. Within six months, Wordle had brought the format — the daily, finishable, share-on-social-media word puzzle — to a population of users who had never previously played a crossword in their lives.

That population, in 2026, is now the largest single audience for word puzzles in human history. Many of them have never solved a crossword. Many of them now do.

Where we are now

The crossword in 2026 is in the strangest position of its 113-year history. The format is more popular than it has ever been — by a factor of probably ten. The construction community is more diverse and inventive than it has ever been. The puzzles are, by most quality metrics, as good as they’ve ever been or better.

And yet the form has fragmented spectacularly. There are now at least a dozen distinct daily word puzzles competing for the same morning attention. There are now serious arguments inside the construction community about whether the NYT Crossword is still the form’s flagship or whether the title has quietly passed to Connections, or to Wordle, or to one of the cryptic Sunday puzzles. There are now, for the first time in the form’s history, more daily puzzle solvers under 30 than over 60.

What hasn’t changed is the basic shape of the experience. Open it. Play it. It ends. Tomorrow there’s another one. That, somehow, has been the same for 113 years.

Arthur Wynne would, I think, be quietly amazed. Margaret Farrar would barely be surprised. She knew, as early as 1924, what she was building.