There's a tidy story that gets told about the Queens puzzle: an old chess problem, dusted off by a social network, same puzzle, new coat of paint. It's a good story. It's also wrong, because the game you play on your commute and the problem mathematicians have chewed on since 1848 share a name, a piece, and almost nothing else. The split between them is our favorite bit of puzzle history, and it's the whole reason the modern game works.

1848: One Question, No Answer Key

The original puzzle is the eight queens problem, posed by Max Bezzel in 1848: place eight queens on a standard chessboard so that no queen attacks another. A chess queen attacks everything in her row, her column, and both of her diagonals, all the way to the edge of the board. Eight queens, eight rows, eight columns, and a thicket of diagonals to thread. Try it on a real board sometime. Getting six queens placed is easy. Getting to seven feels like progress. The eighth is where the afternoon goes.

The problem spread fast. Carl Friedrich Gauss spent time on it, which is roughly the nineteenth-century version of a puzzle going viral.

And there is real counting to do. The 8×8 board admits exactly 92 solutions. Treat rotations and reflections of the board as the same solution (flip it over, spin it a quarter turn) and those 92 collapse down to just 12 fundamentally distinct arrangements. Ninety-two sounds generous until you remember there are over four billion ways to put eight queens on 64 squares. That ratio is what makes the problem miserable by hand and delicious to a mathematician.

The Math Ate the Puzzle

Once the eight-queens question was settled, the natural move was to generalize: put n queens on an n×n board. This is the N-Queens problem, and it has a personality. A 2×2 or 3×3 board has no solutions at all: the diagonals are too cramped. A 4×4 board has exactly two, and they're mirror images of each other. Then the counts take off: 10 solutions at 5×5, 92 at 8×8, 352 at 9×9, 724 at 10×10, and 14,200 by the time you reach a 12×12 board. Solutions exist for every board size from four up; the barren little 2×2 and 3×3 grids are the only failures.

Here's the strange part of the story, though. For most of the twentieth century, N-Queens didn't survive as a puzzle people played. It survived as homework. It became one of the standard teaching examples in computer science — the problem you solve in week four of an algorithms course to learn backtracking, right after the maze and before the knapsack. Generations of programmers met the queens not on a chessboard but in a for-loop. That's a strange fate for a puzzle: too famous to be forgotten, too solved to be fun. Nobody was doing N-Queens over breakfast.

Then a Job Site Handed Everyone a Board

Skip ahead to the mid-2020s, when LinkedIn — of all places — added a daily game called Queens, and suddenly a queen-placing puzzle was part of the morning routine again. One board per day, on LinkedIn, everyone solving the same one. We're not affiliated with LinkedIn, but let the record show: they put a queen back on the table.

Except look closely at the rules, because this is where the tidy story falls apart. Modern Queens (LinkedIn's version, and the Queens puzzle you can play online here, three new boards a day) asks for one queen in every row, every column, and every color region, with no two queens touching, not even diagonally. Set that next to the classic rules and two things have quietly changed.

First, the colored regions are new. Classic N-Queens has no regions; the board is a bare grid. Second, and this is the one almost nobody notices: modern queens don't attack along diagonals. The only diagonal restriction is touching. Two queens may share a diagonal freely as long as they aren't adjacent.

A concrete case: put a queen at row 2, column 3 and another at row 5, column 6. They sit on the same diagonal, three squares apart. In classic N-Queens that's an illegal position; the first queen attacks the second straight down the diagonal. In modern Queens it's completely fine, because they aren't touching. So a finished modern Queens board is usually not a valid N-Queens solution at all. The two puzzles aren't parent and child. They're cousins who happen to have the same name on the mailbox.

What Modern Queens Actually Is

If the modern game isn't N-Queens, what is it? Strip it to its constraints: exactly one piece per row, one per column, one per region, and no two pieces adjacent in any direction. Any Star Battle player will recognize that instantly: it's a 1-star Star Battle with the regions filled in with color instead of drawn as outlines. Star Battle, the puzzle the NYT publishes as Two Not Touch, uses exactly this rule set, typically with two stars per row, column, and region instead of one. Queens is the one-star case wearing a crown. (Star Battle has its own origin story — competition halls rather than chessboards — which we've told in our history of Star Battle.)

This is why players who finish their daily Queens and want something meatier tend to land on Star Battle: doubling the stars per line and region roughly squares the amount of constraint interaction, and the deductions get correspondingly deeper. We've written a full side-by-side comparison of Queens and Star Battle if you want the differences laid out rule by rule.

Blame Wordle

So what brought the queens back? The format, far more than the math.

One shared board, everyone gets the same one, it resets tomorrow. Wordle proved that formula, and every daily game since has been drafting behind it. We think the revival owes essentially nothing to the mathematics and essentially everything to that ritual.

And notice what the format demanded. The classic problem asks "how many arrangements exist?" On an 8×8 board the answer is 92: a fine answer for a mathematician, a useless one for a daily game. You can't share a moment over a puzzle with 92 right answers. So the modern game inverted the question. Engineer a board with exactly one solution, reachable by pure deduction, and ask everyone to find it. The colored regions are there precisely to pin the solution count down to one. That inversion — from counting solutions to guaranteeing a single one — is the real innovation, and it's why the puzzle had to stop being N-Queens to become popular.

It's also the standard we build to. Each of the three boards we publish daily — easy, medium, hard — has to pass the inverted test before it ships: exactly one solution, and a chain of pure deduction that reaches it with no guessing anywhere. Bezzel handed the world a question in 1848 and the world spent 178 years turning it inside out: same piece, same grid, opposite question. The queens are doing lighter work than they did on Gauss's board. They've never had a bigger audience.