A Center Dot book punishes a bad file in a way no sample puzzle ever shows. Your buyer judges puzzle forty-one, when the nine dotted centre cells — one at the heart of each 3x3 box, spread evenly across the whole grid — hold a 4 in the centre of the top-left box and another 4 in the centre of the bottom-right box, because the generator's checker audited rows, columns and boxes and never once gathered the nine marked cells together as a set — or when one box's centre dot quietly vanished in the layout pass, so the region the solution was built on and the region printed on the page are two different sets of nine, and a perfectly correct deduction now looks like a repeat — or when the middle box's dot printed so faint that the buyer never knew the grid's own centre cell carried a second law at all, and solved a harder, different puzzle from the one the key answers. Center Dot Sudoku is classic Sudoku plus one distributed region: a 9x9 grid, digits 1 to 9 once in every row, column and 3x3 box, and the nine centre cells of the nine boxes — the cells at row 2 column 2, row 2 column 5, row 2 column 8, row 5 column 2, row 5 column 5, row 5 column 8, row 8 column 2, row 8 column 5 and row 8 column 8 — which together must also contain the digits 1 to 9 exactly once. Underneath it is a contract in five parts, each part exact. First, the row rule. Second, the column rule. Third, the box rule. Fourth, the Center Dot rule: those nine centre cells are one extra region, a tenth region alongside the nine boxes, and they must hold 1 to 9 exactly once — a full set owed, not merely no repeats. Fifth, closing the contract, given and set integrity with uniqueness: the givens and the dotted set are fixed exactly as published, and exactly one completed grid satisfies every part at once, reachable by deduction alone, with no guessing. A puzzle that breaks any part is not slightly harder. It is false as stated. A good Center Dot book is a verified stack of sets: every dotted nine re-gathered cell by cell against the printed solution, on the final file, puzzle by puzzle.
Four warnings belong at the front, because they are the mistakes this genre makes most often, and all four start with puzzles Center Dot is not. Center Dot is not Asterisk Sudoku, its closest named neighbour, in which the nine extra cells form one shaded star clustered around the middle of the grid: Center Dot's nine sit one to a box, evenly spaced over the whole page, findable by position alone even before any dot is printed, and a checker that gathers a centred star on a Center Dot book has audited nine perfectly respectable cells that are simply the wrong nine, and will pass books whose real centre set repeats a digit. Center Dot is not Girandola Sudoku either, where the nine extra cells form a pinwheel spread across the grid out to its edges: Center Dot's cells never leave the box centres, and a buyer hunting the corners and edges for the region will miss the quiet nine sitting at the heart of every box. Center Dot is not Sudoku X: the grid's centre cell happens to lie on both diagonals, but neither full diagonal is a region here — only the nine centre cells bind, and a solution may repeat a digit along a diagonal without breaking any law this book owes. And Center Dot is not classic Sudoku with decorative dots: the moment a cell is marked as a box centre in this book, it binds as part of the tenth region, and a solution that repeats a digit across the nine marked cells is wrong no matter how clean its rows are.
The book is every centre cell on every puzzle, not the sample puzzle
Every seller checks the sample. The sample is the puzzle on the listing image, the one in the preview, the one whose nine centre cells you gathered while the generator was still open. It will be perfect, because you looked at it. The book is the other ninety-nine puzzles — generated in batch, exported in batch, never re-gathered set by set — and in this genre a batch is exactly where the law quietly dies, because the Center Dot rule is the easiest Sudoku constraint to forget in a checker: it lives in no single row, no single column and no single box, so every ordinary audit passes straight over it. A classic checker ignores the centres entirely and passes everything. A no-repeat checker is subtler and nearly as dangerous in the opposite direction: it confirms the nine cells merely avoid repeats, which any nine scattered cells in a valid classic solution will almost do by accident, and it never asks the harder question this genre actually owes — are all of 1 to 9 present among the nine? A checker that gathers the wrong nine — a centred star, a pinwheel — is the third failure, and the worst, because it produces confident reports about a region the book never promised. The only checker that can catch this genre's failures is one that reads the exact centre set by position, gathers those nine cells, and asks two questions — does any digit appear twice, and is any digit missing? — and if your pipeline never asked both questions set by set, your book's defining rule was enforced by hope.
That is why this genre's reviews, when they turn, turn on trust rather than difficulty. A buyer can forgive a hard puzzle. What they cannot forgive is gathering the nine centre cells, correctly and patiently, one from each box across the whole grid, to prove the last empty centre cell cannot be a 4 — because a 4 already sits in another box's centre, in ink, in the givens — writing in the only digit their logic allows, and discovering the key repeats the 4 among the centres anyway while the missing digit appears in no centre cell at all. One photographed set like that outweighs fifty clean puzzles in a review, and it should: the buyer is right, the file is wrong, and the sample puzzle on your listing, however perfect, is not evidence about puzzle forty-one. So the working rule for everything below is simple. The unit of quality is not the book, and it is not the sample. It is the single centre set, verified on the final exported file, after layout, after any reflow, on the exact bytes the buyer downloads.
The contract, part by part
State the contract exactly, the way your rules page should state it, because every audit that follows is just this paragraph with its sleeves rolled up.
Part one, the row rule: each of the nine rows holds the digits 1 to 9 exactly once. Part two, the column rule: each of the nine columns holds the digits 1 to 9 exactly once. Part three, the box rule: each of the nine 3x3 boxes holds the digits 1 to 9 exactly once. These three are classic Sudoku entire, and they are necessary but nowhere near sufficient — a grid can satisfy all three and still repeat a digit across its nine centres while leaving another digit out of the centre set altogether.
Part four, the Center Dot rule: the nine centre cells — one at the centre of each 3x3 box, at row 2 column 2, row 2 column 5, row 2 column 8, row 5 column 2, row 5 column 5, row 5 column 8, row 8 column 2, row 8 column 5 and row 8 column 8 — taken together, hold the digits 1 to 9 exactly once. Three consequences follow, and your rules page owes the buyer all three in plain words. First, the duty is the full set, not just no-repeat: nine cells holding eight different digits with one repeated and one missing fail this rule twice over, and your page must say the set must be complete, because a buyer fresh from a no-repeat variant will otherwise grade your book by the weaker law — a no-repeat line owes no full set; Center Dot's nine-cell region owes all of it, and confusing the two laws in either direction falsifies a book. Second, the centre set is fixed by position: the region is exactly those nine coordinates — no other cell joins the region because it sits near a box centre, and no centre cell can be dropped from the law because its dot printed faintly. Third, the middle compounds: the grid's own centre cell, row 5 column 5, belongs to its row, its column, its box and the centre region at the same time — four units on one cell, the most restricted cell on the page — and each other centre cell sits at the crossing of its box's middle row and middle column, so a single wrong digit in a centre cell can break several units in one stroke, which is why the audit below reads the centre cells first. And the silence around the law matters as much as the law: a cell that is not a box centre promises nothing beyond its houses, and your page must never let a buyer believe that both full diagonals are policed, or that any marked-looking cell outside the nine coordinates joins the region.
Part five, given and set integrity with uniqueness: the ordinary givens and the dotted set itself are fixed exactly as published — the dots are not decoration, they are part of the puzzle statement, so no cell quietly dotted, undotted or moved between editions — and exactly one completed grid satisfies parts one through four together, reachable by deduction alone. No guessing, no bifurcation, no "the centres also work the other way round." In a distributed-region genre a second solution often hides exactly in the spread: two centre cells in far-apart boxes whose digits can swap without touching any house or the true region's no-repeat surface, which is why uniqueness here must be proven by a solver that enforces the nine centre cells as a region, never assumed from a classic solve.
One more piece of arithmetic belongs in the contract, because the audit leans on it. A complete region of nine cells holding 1 to 9 exactly once excludes each placed digit from the other eight centre cells at once: placing one digit in a centre cell removes that digit from eight cells scattered across eight other boxes at once — a reach no row, column or box can match, because no house touches all nine boxes the way this region does. That reach is the genre's engine. It is also its multiplier of mistakes: an error in the centre set itself — one dot placed on the wrong cell in the export — changes the exclusion map of every puzzle in the book, so the printed puzzle and the solved puzzle silently become two different puzzles that happen to share a grid.
Counting the region itself belongs in the contract too, because sellers get it wrong in print. The Center Dot set is always nine cells — no more, no fewer, and always exactly one per box: an eight-cell marking cannot hold 1 to 9, a ten-cell marking cannot hold them exactly once, and a box contributing two centres while another contributes none has broken the variant's geometry even if the count is nine, so a rules page that cannot point to exactly one dotted centre in each of its nine boxes on its diagram grid is describing a book that cannot satisfy its own cover. Your rules page should therefore never say "the centre cells" as if the buyer could guess which ones: it should show the dotted set on a diagram grid, in the same orientation and marking as the puzzles, name the one-per-box rule in words, and promise that every puzzle in the book uses that same positional set unless a puzzle page says otherwise. A buyer who has to infer the region from the first puzzle's givens has been handed homework where a diagram was owed, and a book whose later puzzles quietly move a dot — a batch generated against a shifted layout and merged in — has broken its contract in a way no single-puzzle check will catch. The audit's set check therefore runs in both directions: file against source, and puzzle against puzzle across the whole book.
The verification audit: eight checks before any file ships
Run all eight on the final file, per puzzle, in this order. The audit assumes nothing from generation; it re-reads the page as the buyer receives it.
House recount. For the printed solution, recount every row, column and box: nine distinct digits, 1 to 9, no repeats, no omissions. This is the classic pass, and it is only the entry fee.
Centre full-set recount. Gather the nine centre cells exactly as positioned — row 2 column 2, row 2 column 5, row 2 column 8, row 5 column 2, row 5 column 5, row 5 column 8, row 8 column 2, row 8 column 5, row 8 column 8 — and confirm they hold 1 to 9 exactly once — no digit twice, no digit missing. Not a spot check, not the middle box only — all nine, every puzzle, corner boxes included. This single check is the genre; a checker that only audits houses has failed to run it, and the difference is precisely the repeated 4 sitting in two box centres at opposite corners of a grid that "looked fine."
Dotted-set audit. Read the dots in the final file against the source puzzle, cell by cell: every centre cell dotted, no extra cell dotted, no dot placed one square off after a layout pass, and exactly one dot per box. In this genre the dotted set is law, so a marking defect is a rules defect: a centre cell that lost its dot in the PDF is a constraint the buyer cannot see and cannot be blamed for misreading.
Wrong-nine sweep. Confirm the checker and the rules page never gathered a different nine: not a centred star cluster, not the pinwheel cells of Girandola, not both full diagonals, not any nine convenient cells. Re-read the audit configuration and the page copy, and strike any test that grades a set the book never promised. This is the genre's signature false pass, and a book "verified" under it has not been verified at all.
Given-agreement sweep. Diff the printed givens against the source puzzle cell by cell, and confirm every given that sits in a centre cell agrees with the centre set as printed — a given moved one cell by a reflow can land outside the centre the solution was built against, and no house check will ever notice.
Uniqueness under the real rules. Re-solve each puzzle with a solver that enforces rows, columns, boxes and the nine centre cells as a full region, and confirm exactly one solution, reachable by deduction without guessing. A classic-only solver is not evidence here: it will happily certify puzzles that have a second solution hiding in the spread.
Stray-law sweep. Confirm nothing unmarked is being enforced and nothing marked is being ignored: no solution may be failed for repeating a digit along a full diagonal the key notes invented, and no centre cell may be excused as "only a guide." The file's dots and the file's law must be the same set, in both directions.
Printed-proof solve. Solve one puzzle per difficulty band from the printed file itself — paper or the delivered PDF at reading size — gathering the nine centre cells by eye as a buyer would. Dots too faint to spot at reading size, a centre dot that reads as a printing speck, and digits that vanish against the dot are failures this check exists to catch, because the buyer will catch them in a review instead.
A book that passes all eight earns its rules page. A book that skips check two has not been audited at all, whatever else was run — it has only been proven to be a classic Sudoku book with dots drawn on it.
Difficulty is how hard the centres work, not how few givens you printed
Given count is the laziest difficulty metric in Sudoku, and in Center Dot it is close to meaningless. Two puzzles with identical given counts can sit at opposite ends of honest difficulty, because what makes a Center Dot puzzle hard is not how many cells start empty but how much work the distributed region does: how often the decisive deduction is a centre exclusion reaching across the grid rather than a house exclusion next door, and how early the solve is forced to gather nine cells that never share a house with each other. A puzzle whose centres never decide anything is a classic puzzle wearing the dots as costume — sellable, perhaps, but not the book the cover describes, and buyers who came for the variant can tell by puzzle ten. A puzzle whose every step runs through the grid's centre cell is a bottleneck, not a challenge: it reads as a trick repeated until it opens.
Grade instead on four things you can measure. First, centre-decision share: in a solver's step log, what fraction of placements were forced by the centre region rather than by a row, column or box. Second, middle load: how early and how often the solve is forced through row 5 column 5, where four units meet and the pattern's pressure concentrates. Third, cross-grid reach: how often a decisive centre deduction runs between far-apart boxes — a corner box's centre excluding a digit in the opposite corner's centre — instead of staying inside neighbouring boxes where ordinary logic would have found it anyway. Fourth, classic depth underneath: the ordinary solving techniques the puzzle still demands once the region has done its work, so an "expert" label is not carried by one distributed gimmick over an otherwise trivial grid.
Then hold the listing honest against the grades. The sample puzzle on the listing should sit at the book's true median — not the easiest puzzle, polished, and not the hardest, weaponised. State the region in the listing in plain words: that the centre cell of each 3x3 box together form one extra region that must hold 1 to 9 exactly once. Buyers of variant Sudoku shop by law, not by artwork; the seller whose listing states the law exactly is the seller whose reviews talk about difficulty as a preference instead of as a surprise. And never grade with a no-repeat assumption hiding in the tooling: a difficulty engine that treats the centres as merely no-repeat will call puzzles solvable that are not, and unsolvable-looking puzzles that are merely waiting for the missing-digit deduction, and the grades will drift from the buyer's experience in exactly the direction that produces "not as described."
Print honesty: the law is printed as nine small dots, the first thing a bad export kills
In other variants the law is printed as shading or lines; in Center Dot it is printed as nine small dots, one at the heart of each box, and small marks have their own failure mode: a dot is easy to drop, easy to shrink below visibility in a print export, and easy to mistake for a speck, a stray pixel or a blemish on the page — and every one of those failures deletes law from the page while leaving the puzzle's difficulty fully intact. The buyer is then solving a harder, different puzzle from the one the key answers, and no errata note that says "the centres are the usual ones" fixes a page where the buyer cannot tell which nine cells are marked — even though, in this genre alone, a careful buyer could in principle reconstruct the set from position, because the centres never move.
So the print rules are mechanical. Print every dot at a size and weight that survive a bad photocopy: test the actual delivered file on the worst output you can find, not the best screen you own. Keep the dots individually findable — exactly nine, one per box, each readable as a deliberate mark at the centre of its cell, never as noise, and a buyer should be able to count exactly nine without a magnifier: if they count eight or ten, the page has failed before the puzzle has begun. Keep digits legible beside and over the dots: a given printed over a heavy dot that swallows its strokes turns reading the given into guessing it, and the grid's centre cell — the cell the whole solve leans on — is the worst place in the book for a digit to go faint. And give the buyer a fallback: a rules page that describes the dotted set precisely enough — a diagram grid with the nine centres shown, and the coordinates stated in words — that a buyer whose print faded can reconstruct the law without writing to you. The set is positional and fixed, so say so and show it; if a puzzle's dotted set ever differs, flag it on that puzzle's page, never silently.
The solution key carries the same duty. Print the key with the same nine cells dotted at the same strength as the puzzle page, large enough that a buyer disputing the region can gather the nine key cells themselves and count 1 to 9 across them. A key that drops the dots — common in batch exports, where the key is rendered from the solution grid alone — makes every centre dispute unanswerable from the book itself, and unanswerable disputes become reviews. The genre runs on gathered sets; the whole book, givens, dots and key, must be gatherable end to end.
Licensing, provenance, and whose variant this is
Center Dot Sudoku is a modern variant in the public Sudoku tradition: the rules themselves — classic houses plus the nine box-centre cells holding 1 to 9 as one extra region — are a published convention used by setters, apps and championships, not a private invention you license or a brand you borrow. The name is the marking itself — a dot at each centre — shared variant culture, not a trademark you are borrowing. What you may not borrow is other people's puzzles, other people's authored sets, and other people's artwork. Generate your own puzzles, mark your own grids, and keep the generation record: the software or method, the centre set used, the audit output per puzzle. That record is your provenance if a marketplace ever asks, and it is your errata map when something slips through.
State your licence terms for the buyer plainly, the way the honest sellers in this niche do: personal use, print for your own solving, no redistribution of the file. If you allow classroom or club printing, say how many copies and for whom; silence on the point does not protect you, it just guarantees the keenest buyers — teachers and club organisers, often your best repeat customers — will either ask and wait, or walk. And keep one rule for yourself above the licence text: never lift puzzles from a free puzzle site, a newspaper, or another seller's book to fill a slow volume. The variant community is small, centre patterns and given patterns are fingerprints, and a single recognised puzzle ends the trust the audit was built to create.
Corrected-file delivery is the promise that closes the loop. When an error does slip through, the fix is not a quiet re-upload that leaves early buyers holding a false book: it is a corrected edition, a note saying which puzzle changed and why, and the new file pushed to every buyer of the old one. I run my own delivery for books like these through FileFreight, which exists for exactly this unglamorous job — flat-rate delivery of the file, license keys where a book needs them, and a corrected edition reaching the people who already paid, not only the people who buy tomorrow. Whatever tool you use, hold it to that standard: if your delivery cannot re-serve a corrected file to past buyers, your errata policy is a sentence you cannot keep.
The launch checklist
- The rules page states the contract in five parts, says the centre cell of each 3x3 box forms one extra region with the other eight, states that region must hold 1 to 9 exactly once, and shows the exact dotted set on a diagram grid.
- Every puzzle was audited on the final exported file: houses recounted, the nine centre cells gathered by position and confirmed complete 1 to 9, and the dots diffed against the source, cell by cell, exactly one per box.
- Uniqueness was proven by a solver that enforces the centre region, and no puzzle in the book needs guessing.
- Difficulty grades come from how hard the centres work — decision share, middle load, cross-grid reach, classic depth — not from given count, and the listing sample sits at the book's true median.
- The delivered file was print-tested on poor output: exactly nine dots are countable at reading size, digits stay legible over the dots, and the key prints the same nine cells dotted.
- Puzzles, grids and artwork are your own generation, with the generation record kept; licence terms state personal use and any classroom or club allowance in plain words.
- A corrected edition can be re-served to every past buyer, with a note naming the puzzle that changed — and the delivery tool you use was held to that standard before the first sale, not after the first complaint.
Top comments (0)