Puzzle Cottage
The mathematical prehistory: Euler and Latin Squares (1783)
Sudoku’s ancestor isn’t a puzzle at all — it’s a piece of pure mathematics. In 1783, Swiss mathematician Leonhard Euler studied what he called “Latin Squares”: an n×n grid filled with n different symbols, where each symbol appears exactly once in each row and once in each column.
Latin Squares are everywhere in math — combinatorics, statistical experimental design (especially agriculture, which is where they got formalized), error-correcting codes. They’re also exactly the row/column constraint of Sudoku. What Euler didn’t add was the third constraint that makes Sudoku Sudoku: the requirement that each 3×3 sub-grid also contain every digit exactly once.
Euler couldn’t have invented modern Sudoku. He was studying combinatorial structures, not designing entertainment. But every modern Sudoku grid is a Latin Square — one with extra constraints.
The 19th-century French puzzle precursors
Several 19th-century French newspapers ran number puzzles that are sometimes called “proto-Sudoku.”
- Le Siècle (1892) — published a 9×9 magic-square puzzle where rows, columns, and diagonals all summed to a fixed number. This is closer to a magic square than to Sudoku — it lacks the no-repeat constraint.
- La France (1895) — published a similar puzzle but with the explicit no-repeat-in-row-column constraint. This is the closest 19th-century approximation to Sudoku, but it still didn’t add the 3×3 box constraint.
Why didn’t the French puzzles catch on? Probably the same reason most newspaper puzzles don’t: they ran for a few years, the columnist who ran them moved on, and they were forgotten. The puzzle space had to wait nearly a century for someone to add the 3×3 boxes and create the modern format.
1979: Howard Garns and “Number Place”
The actual inventor of modern Sudoku was Howard Garns, a 74-year-old retired American architect from Indianapolis. In 1979, working freelance for Dell Magazines, he designed a puzzle called “Number Place.” It first appeared in Dell Pencil Puzzles & Word Games magazine that year.
Number Place is identical to modern Sudoku in every meaningful way:
- 9×9 grid divided into nine 3×3 boxes
- Some cells filled in as clues
- Goal: fill the rest so every row, column, and box contains 1–9 exactly once
- Unique solution by construction
Garns died in 1989, six years before the puzzle he created became a global phenomenon. He never saw it called “Sudoku,” never knew the format would dominate newspaper puzzle pages worldwide. Dell Magazines didn’t credit him in the original publication — the byline was just “Number Place.” His authorship was traced retroactively in the 2000s when the puzzle’s history became newsworthy.
1984–1986: Maki Kaji, Nikoli, and the rebranding
In 1984, the Japanese puzzle magazine Nikoli — founded by Maki Kaji — began running Number Place. Kaji had spotted the puzzle in an American magazine while traveling and saw potential. Nikoli’s contribution was twofold:
- The new name. Kaji branded the puzzle “Sūji wa dokushin ni kagiru” — literally “the digits must be single (unmarried)” — later contracted to “Sudoku” (数独). The “single” refers to each digit appearing only once per row, column, and box.
- The aesthetic refinement. Nikoli established what became modern Sudoku conventions: rotational symmetry of given clues, no more than 30 clues for puzzles to feel elegant, careful difficulty curation. American Number Place puzzles weren’t held to these standards.
Kaji didn’t patent or trademark “Sudoku” outside Japan, which mattered enormously later: the rest of the world could use the name freely. Maki Kaji died in 2021, age 69 — by then his name was associated with the global Sudoku phenomenon, even though he didn’t invent the puzzle.
2004: Wayne Gould and the global explosion
For 20 years (1984–2004), Sudoku was a Japanese puzzle — popular in Japan but unknown elsewhere. The breakthrough came from an unlikely source: Wayne Gould, a retired Hong Kong judge from New Zealand who had spent six years writing a computer program to generate Sudoku puzzles.
In late 2004, Gould walked into the offices of The Times of London with a laptop full of computer-generated Sudoku and pitched the editor on running them daily. The Times agreed. The first daily Sudoku appeared on November 12, 2004.
Within weeks:
- The Daily Mail began running daily Sudoku.
- The New York Post picked it up in April 2005.
- By summer 2005, Sudoku was in every major English-language newspaper worldwide.
- Sudoku books topped bestseller lists in the UK, US, Australia, and many other markets.
Gould’s real innovation wasn’t the puzzle — that was Garns’s — or the name — that was Kaji’s. It was the computer generation. Hand-curated Japanese Sudoku from Nikoli couldn’t scale to daily-newspaper publication globally. Gould’s software could produce unlimited unique-solution puzzles at any difficulty. That’s what made daily-paper Sudoku economically viable.
Timeline: from Euler to today
Is Sudoku Japanese? The complicated answer
The puzzle was invented in America. The name is Japanese. The aesthetic conventions (clue symmetry, difficulty curation) are Japanese. The global popularization came from a New Zealander living in Hong Kong publishing in a London newspaper. Sudoku is a peculiarly multinational invention.
Most popular accounts call it Japanese because the name is Japanese and Nikoli’s aesthetic refinements gave it the polish that made it appealing globally. That’s defensible — without Nikoli, Number Place would have remained an obscure Dell Magazines puzzle. But it wasn’t invented in Japan, and crediting Japan for the invention erases Howard Garns.
The fairest summary: American invention, Japanese refinement and rebrand, Hong Kong-via-Britain global syndication.
How many Sudoku puzzles are there?
In 2005, Bertram Felgenhauer and Frazer Jarvis used computer enumeration to count valid 9×9 Sudoku grids. The answer: 6,670,903,752,021,072,936,960 — about 6.67 sextillion.
That number is for valid grids (every row/column/box satisfies the rules). When you account for symmetry — rotations, reflections, swapping any two digits, swapping band-rows or stack-columns — the number of “essentially different” Sudoku grids drops to about 5,472,730,538 — about 5.5 billion.
For puzzle generation, the numbers are different again. A puzzle is a partially-filled grid with a unique completion. The minimum number of clues needed for a unique-solution Sudoku puzzle is 17 — proven by McGuire, Tugemann, and Civario in 2014 via exhaustive computer search.
The pronunciation question
In Japanese, Sudoku is pronounced soo-DOH-koo (each syllable equally weighted, no emphasis). In American English, it’s usually pronounced soo-DOH-koo with the second syllable stressed, often losing the Japanese rhythm. Both are accepted.
The bigger pronunciation curiosity: many English speakers say “sah-DOH-koo” (replacing the “oo” in the first syllable with a schwa). This is wrong in Japanese but extremely common in casual American usage.
Where to play Sudoku now
Free daily Sudoku at Puzzle Cottage — algorithmically generated with seeded backtracking (same approach Wayne Gould used in 2004, modernized), three difficulty levels, every past puzzle free to replay forever. No subscription, no sign-up. See also our free Sudoku alternative page for the full overview, or Sudoku tips when stuck for strategy.