How a 15×15 grid of white squares travels from a constructor's kitchen table to Will Shortz's desk — and eventually to your morning coffee.
Every morning, somewhere between five and ten million people open the New York Times crossword. They approach it the same way they pour their coffee — as part of a ritual so ingrained it barely registers as a choice. But behind that ritual lies an extraordinarily complex editorial machine: a pipeline that transforms a constructor's spark of inspiration into a precisely calibrated daily challenge, graded to the day of the week, vetted by professional test solvers, and polished by a single editor who has now overseen more than 11,000 puzzles.
That editor is Will Shortz. That pipeline has been running since 1993, and very few people outside the crossword community understand how it actually works. In this episode, we pull it apart stage by stage — from the constructor's first blank grid all the way to the moment a solver pencils in that final square and feels the particular satisfaction that has no adequate name.
Understanding the pipeline doesn't just satisfy curiosity. It changes the way you solve. When you know why a Monday clue is phrased one way and a Thursday clue another, when you understand what "fill quality" means and why constructors lose sleep over a single inelegant three-letter answer, the crossword transforms from a test you either pass or fail into a conversation across a page.
The first thing to understand about the NYT crossword is that it is almost entirely created by independent constructors — freelancers who pitch puzzles the same way writers pitch articles. The Times does not employ staff crossword constructors. Will Shortz receives hundreds of submissions every month from a community ranging from seasoned professionals to first-time hopefuls, and he accepts roughly 3 to 5 percent of them.
What makes a submission? A complete crossword puzzle consists of two things: a filled grid (every square containing a letter, with black squares placed symmetrically) and a full set of clues. The grid comes first. The clues, as we'll see, are often substantially rewritten by the editor, so constructors know their clue-writing is more of a starting negotiation than a final product.
Grid construction is an art form with firm technical constraints. Black squares must be placed with 180-degree rotational symmetry — if a black square appears in the upper-left, a corresponding black square must appear in the lower-right. No "unchecked" squares are permitted, meaning every white square must be part of both an Across and a Down answer (no squares that appear in only one word). Minimum word length is typically three letters, and most editors frown on words shorter than four in non-corner positions. The maximum word count is around 78 for themed puzzles, though many constructors aim for 72 or fewer to allow longer, more interesting answers.
For themed puzzles — the vast majority of published weekday and all Sunday puzzles — the constructor must also design a cohesive thematic concept. Theme entries are typically the longest answers in the grid, often sharing a pattern, a hidden word, a wordplay gimmick, or a conceptual link. The reveal — an answer that explains what all the theme entries have in common — is one of the most satisfying moments in solving and one of the hardest moments in construction.
Once a puzzle is submitted, it enters a structured editorial process that most solvers never see. Here is how a typical NYT crossword moves from constructor to publication:
Constructor submits a completed grid plus clues through the NYT's submission portal. Shortz or an associate editor performs an initial review for technical compliance (symmetry, connectivity, minimum word length) and thematic strength. Puzzles that fail basic criteria are declined quickly; strong puzzles move forward for deeper evaluation.
Accepted puzzles are sent to a small group of test solvers — typically Shortz himself plus one or two experienced associates. Test solvers work through the puzzle without any assistance, noting every answer that feels forced, clumsy, or unfair, and every clue that seems ambiguous or inaccurate. Their feedback flags fill quality issues and clue problems before any deeper editorial work begins.
The grid's fill is assessed against quality standards. Answers that rely on obscure vocabulary (crosswordese), contain repeated letter patterns (-ER/-EST stacking), or cross in ways that create unfair solving experiences may require the constructor to revise the grid. In some cases, Shortz will work with a constructor through multiple revision rounds; in other cases, the puzzle is declined and returned with suggestions for resubmission.
This is Will Shortz's most visible and distinctive contribution. Most constructors' clues are substantially rewritten before publication. Shortz adjusts difficulty level for the intended day of the week, adds misdirection where appropriate, updates cultural references to feel contemporary, and ensures factual accuracy. A constructor's straightforward definitional clue for a Thursday puzzle might become a wordplay trap; a Monday clue might be simplified to remove any possible ambiguity.
Every factual claim in every clue is verified by the editorial team. This includes not just obvious facts (the year a film was released, the capital of a country) but also more subtle claims (whether a phrase was actually used in a particular context, whether an attribution is correct). The NYT's crossword has occasionally issued corrections when errors slip through, but the fact-checking process is rigorous by design.
Accepted puzzles are placed into the publication queue, assigned to a specific day of the week based on difficulty. A newly constructed puzzle might wait 6 to 18 months before publication. The queue must be balanced across days and must ensure diversity of constructors, themes, and fill styles. Shortz manages this balance consciously, avoiding runs of similar themes or back-to-back puzzles from the same constructor.
The puzzle is published in the print paper and simultaneously released on the NYT Games app. Every puzzle is archived in the NYT's database, accessible to subscribers through the app's archive feature. Published puzzles are attributed to their constructor by name — one of the crossword world's few visible bylines.
One of the most deliberate features of the NYT crossword is its strict difficulty gradient across weekdays. This isn't just a convention — it's a pedagogical system, a way of teaching new solvers gradually while challenging veterans with increasingly demanding puzzles as the week progresses.
Thursday is the crossword week's wildcard. It is the day Shortz schedules puzzles with unusual mechanical gimmicks — rebus squares (where multiple letters, a number, or a symbol fits in a single square), black-square shapes that form visual patterns, clues with hidden meanings that require solvers to think outside conventional grid interpretation. The Thursday puzzle often produces the loudest complaints and the most enthusiastic praise of any weekday puzzle.
Friday and Saturday are themeless puzzles — also called "freestyle" — which present their own construction challenges. Without theme entries to anchor the long answers, constructors have complete freedom over fill, which means the fill must be exceptional. Fridays and Saturdays typically feature long stacked entries (15-letter answers crossing each other in the same region), and their clues lean heavily on wordplay and cultural knowledge rather than direct definitions.
Sunday's 21×21 format earns its own cultural moment. The Sunday puzzle is the one many casual solvers attempt who wouldn't dare try a Saturday. Its size is larger, but its difficulty sits roughly at Thursday level — complex enough to be satisfying, accessible enough to be completable over a long morning.
The term "fill" refers to the words that fill the grid — all of them, themed and non-themed alike. Fill quality is one of the most contested topics in crossword construction, and it has changed dramatically since Will Shortz took over as editor in 1993. Understanding what Shortz-era NYT crosswords consider high and low quality fill reveals the aesthetic philosophy underlying the puzzle.
| Fill Type | Example | Status Under Shortz | Reason |
|---|---|---|---|
| Classic crosswordese | ESNE, ETUI, ORLE, INIA | Avoid | Obscure words known only from crosswords; solvers must memorize rather than recognize |
| Roman numeral fill | MXII, DCCC | Avoid | No independent cultural meaning; created purely to satisfy grid constraints |
| Partial phrases | A PART, IN ONE | Avoid | Forces solvers to guess which part of a phrase is intended; feels arbitrary |
| Suffix/prefix stacking | Multiple -ER, -EST, -ION words | Avoid | Grid feels mechanical; reveals constructing-by-constraint rather than inspiration |
| Brand names & proper nouns | OREO, UBER, TESLA | Welcome | Common cultural touchstones that solvers encounter daily; feels contemporary |
| Long colloquial phrases | SO LONG STORY SHORT, EPIC FAIL | Welcome | Captures living language; makes experienced solvers smile when they parse it |
| Pop-culture proper nouns | LIZZO, MAYA ANGELOU, BILLIE EILISH | Welcome | Reflects the full breadth of contemporary culture rather than traditional high-culture bias |
| Clever abbreviations | MVP, DNA (clued freshly) | Welcome | Fine when clued imaginatively; problematic when clued literally in older style |
The Shortz era's embrace of brand names, slang, and pop culture references was initially controversial among older solvers who preferred the puzzle's traditional literary and classical emphasis. Over time, it has become the defining characteristic of the modern NYT crossword — a puzzle that reflects the culture of the solvers who will encounter it rather than a narrowly defined highbrow tradition.
Duplicate fill is another quality standard: the same word should never appear twice in a single puzzle, even in different tenses or plural forms. The word ROOT and ROOTS should not coexist. APPLE and APPLE PIE create an uncomfortable near-duplication. Modern construction software flags these automatically, but the constraint forces constructors to think carefully about their entire grid as an integrated composition.
Shortz holds the only known academic degree in enigmatology — the study of puzzles — awarded by Indiana University in 1974. He served as Games magazine editor before taking the NYT position, and he has now shaped the puzzle longer than any previous editor in its history, which dates to 1942.
Shortz succeeded Eugene T. Maleska, whose tenure from 1977 to 1993 was marked by a preference for classical allusions, literary references, and the kind of specialized vocabulary that demanded a particular educational background. When Shortz took over, he brought not just different aesthetic preferences but a different philosophy: the crossword should be fair, contemporary, and accessible to a broad educated audience rather than a narrow scholarly elite.
His innovations were substantial. He standardized the Monday-through-Saturday difficulty gradient. He actively recruited younger constructors, diversifying the puzzle's voices. He expanded the range of acceptable pop-culture references. He introduced the idea of the Thursday gimmick as a regular feature. He began accepting puzzle bylines from people who had never constructed before, creating pathways for new voices into the form.
Not all of Shortz's decisions have been universally praised. Critics argue that the puzzle has occasionally been too pop-culture heavy, that certain topical references age poorly, and that the threshold for what constitutes "fair" obscurity has shifted. The crossword community — now organized around blogs, Discord servers, and competing solve-time leaderboards — debates these questions daily. But Shortz's thirty-plus years have unquestionably transformed the NYT crossword from a niche enthusiast product into a daily cultural institution.
If you were to distill what Shortz and his team look for in a submission, it would come down to a handful of fundamental qualities. These are the checklist items that separate puzzles that earn a closer look from puzzles that are returned with a polite decline.
Black squares must be arranged with 180-degree rotational symmetry. No exceptions in standard puzzle formats.
Every white square must belong to both an Across and a Down answer. Floating or isolated squares are disqualifying.
The theme must not duplicate a concept already published. NYT has 80+ years of archives; Shortz has a strong memory for what's been done.
All theme entries must exemplify the concept equally well. One weak theme entry undermines the whole puzzle.
No crosswordese, no awkward plurals, no Roman numeral fill, no forced partials. Every word should feel intentional, not merely convenient.
Two obscure answers should never cross at an unknowable letter. Every crossing should be solvable through at least one of the two intersecting clues.
All clues must be factually verifiable. Any historical claim, attribution, or statistic in a clue will be checked against primary sources.
The constructor should label their submission with an intended day of the week. The actual scheduling may differ, but the initial pitch signals awareness of difficulty calibration.
The NYT's puzzle ecosystem now extends well beyond the daily crossword. Understanding the crossword pipeline helps illuminate the editorial philosophy across these companion products as well.
The Acrostic (formerly a Sunday puzzle, now published separately) presents a grid with numbered blanks and a set of lettered clue answers. When the clue answers are filled in and transferred to the grid by number, they spell out a passage from a book or other text — and the first letters of all the clue answers, read in sequence, spell out the author's name and work. The Acrostic requires a completely different construction skill than the crossword, centered on finding quotations whose letters interlock cleanly with the grid.
The Spelling Bee, created by Sam Ezersky, presents seven letters arranged in a honeycomb. Solvers must find as many words as possible using only those letters, with the center letter required in every word. A single "pangram" — a word using all seven letters — is embedded in each puzzle. Despite using no grid at all, the Spelling Bee has become enormously popular, reportedly now more played than the daily crossword.
The Connections puzzle, the newest major addition, presents 16 words or phrases and asks solvers to group them into four categories of four, with each category having an increasingly tricky conceptual link. Its rapid rise in popularity — driven in large part by social media sharing — has introduced an entire generation of players to the NYT Games ecosystem who might not have found their way to the crossword first.
Each product has its own editorial staff and its own construction or design pipeline. But they all share the same underlying commitment: puzzles that reward curiosity, that calibrate difficulty honestly, and that treat solvers as intelligent adults rather than consumers to be flattered.
The NYT crossword pipeline is a case study in something much broader than puzzle-making: the mechanics of rigorous editorial curation. The process illustrates several principles that apply across any creative discipline where quality is non-negotiable.
First, standards enable creativity rather than suppressing it. The strict technical requirements of crossword construction — symmetry, no unchecked squares, minimum fill quality — don't limit constructors; they create a shared framework within which genuine creativity can be recognized and appreciated. When the rules are clearly defined, breaking them intentionally (as Shortz occasionally allows for exceptional puzzles) becomes meaningful rather than merely sloppy.
Second, difficulty calibration is a form of respect. The Monday-through-Saturday gradient isn't about making some solvers feel smart and others feel dumb. It's about meeting solvers where they are and providing a growth pathway. A new solver who starts on Mondays and works toward Thursdays is receiving a genuine education in wordplay, cultural knowledge, and mental flexibility — without ever feeling tricked unfairly.
Third, editorial distance produces better work. Most constructors will tell you that Shortz's clue rewrites improve their puzzles. The editor's distance from the construction process allows him to see what clues will land with solvers rather than what the constructor found clever in the moment of creation. This is the same insight that makes any editing relationship valuable — the creator is too close to the work to judge it from the outside.
And finally, the archive matters. Shortz's awareness of what has already been published prevents the repetition that would make the crossword feel stale. Building against an archive — whether of published puzzles, of films, of scientific papers — is what allows a creative tradition to develop rather than just accumulate.
The official NYT submission portal with full specifications for weekday and Sunday puzzle submissions.
Long-form journalism tracing the construction of a single NYT crossword from concept to publication — one of the best available accounts of the process.
The definitive archive and analytics platform for NYT crosswords, with constructor stats, clue history, and fill analysis going back decades.
The primary online community for crossword constructors, featuring forums, resources, and guidance for those building their first grids.