Beginner to Advanced · 6×6, 8×8 & 10×10 Strategies
The patterns that actually save time on Mini Sudoku, worked through on real 6×6 boards. Pencil-mark discipline, hidden singles, pointing pairs, and a short practice session you can keep up.
Beginner
The fundamentals. Get comfortable on 6×6 first, then 8×8 and 10×10 follow naturally.
Every Mini Sudoku deduction is the same move repeated: find the cell where the rules have already made the choice for you. On a 6×6 board each cell belongs to a row of six, a column of six, and a 2×3 region of six, and those three groups overlap heavily. Five distinct digits between them is all it takes to force the sixth. The two diagrams below are real positions from puzzles in our library — the digits are exactly as they appear on the board, and each one links through so you can finish it yourself.
Beginner #1
Naked singles: a cell where only one digit fits.
It is the cheapest win on the board, and on 6×6 it comes up constantly: the cell only has to see five distinct digits across its row, its column and its region before the sixth is forced. The first diagram below is exactly that, with all three groups shaded so you can watch them add up.
Beginner #2
Use pencil marks. Update them every time you place a value.
Notes cost seconds and save minutes. The habit that matters is the update, not the marking: the moment you commit a digit, strike it from the notes in that row, that column and that region. Stale pencil marks are worse than none at all, because you will trust them.
Beginner #3
Look at each region for the digit that's missing — it's often easier than scanning the whole board.
Six regions, six digits each. Asking which of 1–6 is missing from a region narrows the hunt to two or three cells instead of thirty-six. Cross-hatching, in the second diagram below, is the same instinct pointed at a digit rather than at a region.
Beginner #4
After every placement, re-check the row and column for new singles.
Placements arrive in chains. A digit you just wrote removes a candidate from up to nine other cells, and one of those is very often forced as a result. Solvers stall by re-scanning the whole grid; the next move is nearly always next door to the last one.
Look at the shaded cell in the bottom-left corner. Its row already holds 5, 4 and 3. Its column adds a 1. Its region — the block of six cells to its right and above — adds a 6. That is 1, 3, 4, 5 and 6 accounted for, so the cell can only be a 2.
Nothing clever happened here: the three shaded groups simply cover five of the six digits between them, and the only work was reading all three instead of one. On this particular board it is also the only cell where that works — everywhere else you have to go looking.
The region in the middle of the right-hand side is completely empty, and it needs a 5. Rather than test each of its six cells, follow the 5s that are already on the board. The 5 in row 4 sweeps across the region's bottom row and kills all three cells. The 5 in column 5 comes down and kills one more. The 5 in the bottom-right corner comes down column 6 and kills another.
Five of six cells struck out, one left standing — the shaded cell takes the 5. Note what candidate-counting would have told you here: nothing. That cell still has five possible digits, so it is invisible to a naked-single search. Scanning by digit is the only cheap way to find it.
Intermediate
Hidden singles, pairs, and box-line reduction. The next layer once the basics feel automatic.
Intermediate play is where pencil marks start paying for themselves. Once every empty cell in a unit carries its candidate list, you can stop asking “what fits here?” and start asking “where can this digit still live?” — a different question with different answers. Most of the moves below are eliminations rather than placements; they clear the fog so the next single becomes obvious.
Intermediate #1
Hidden singles: a digit that can only go in one cell within a row, column, or region.
A hidden single hides behind a busy cell. The cell may still show three or four candidates, so counting candidates will never surface it. You have to come at it from the other end: pick a digit, ask where it can still go inside this row, column or region, and count the answers. One answer means you can write it in.
Intermediate #2
Naked pairs: when two cells share the same two candidates, those digits can't go anywhere else in the unit.
Two cells, the same two candidates, one unit. Between them they have already claimed both digits, whichever way round it falls, so every other cell in that unit can drop both. On a six-cell unit that often collapses the rest of the row in a single pass.
Intermediate #3
Compare row and column patterns to find contradictions early.
Rows and columns cross, and the crossing is where contradictions show up first. If a digit is confined to two cells in a row and both of those cells sit in the same region, you have learned something about the region too. Reading both directions at once is what separates a two-minute solve from a five-minute one.
Intermediate #4
Box-line reduction: if a digit in a region is confined to one row or column, remove it from the rest of that row or column.
Box-line reduction is the workhorse of the intermediate set. It never places a digit by itself — it only deletes candidates — but those deletions are what make the next hidden single visible. Treat it as setup, not as a finishing move.
The shaded column has four open cells. Pencil them out and you get 2 or 3 at the top, then 2, 4 or 6, then 2 or 6, then 3 or 6. Not one of those is down to a single candidate, so a naked-single scan walks straight past this column.
Now flip the question. The column still needs a 4, and the 4 shows up in exactly one of those four lists — the highlighted cell. It does not matter that the cell has two other candidates: if 4 has nowhere else to go in the column, the cell is a 4. That is a hidden single, and finding it is worth the thirty seconds the pencil marks cost.
Speed
A short, repeatable session for getting noticeably faster within a week or two.
Speed on Mini Sudoku is almost entirely pattern recognition, and pattern recognition responds to frequency rather than duration. Ten focused minutes a day will move your times further in a fortnight than a two-hour session at the weekend. Here is a session short enough that you will actually keep doing it.
Speed #1
Use 6×6 as a warm-up before timed runs on bigger grids.
A 6×6 warm-up puts the shapes back in your head before the clock matters. Two minutes on a small grid is enough to remember what a naked pair looks like, and it costs you nothing if the timed run then goes badly.
Speed #2
Place all candidates for one digit at a time. It's faster than scanning cell by cell.
Scanning by digit beats scanning by cell because you only hold one number in your head at a time. Sweep 1 through 6 across all six regions, place everything that is forced, then sweep again. Most 6×6 boards give up two or three digits on the first pass.
Speed #3
Play daily. Pattern recognition gets noticeably faster within a week.
Recognition, not reasoning, is the thing that gets faster. The logic never changes; what changes is how long it takes you to notice that a row already holds five of its six digits.
Advanced
These come up on harder boards. Add them once the fundamentals feel automatic.
A 6×6 grid is small enough that the advanced patterns rarely all appear on one board. What you do need is the confidence to keep going when the easy placements dry up — which usually means trusting a chain of ordinary deductions rather than hunting for something exotic. The worked solve at the end of this section is a good example: it starts with one non-obvious move and then coasts.
When a candidate only appears in one row or column within a region, eliminate it from the rest of that row or column to force progress.
The region has not told you which of its cells takes the digit, only that the digit stays on that line. That is enough. Anywhere else on the line is now dead, and on a 6×6 grid a line only has six cells to begin with.
If a pair (or trio) of cells inside a region share the same candidate, remove that candidate from other cells in the intersecting row or column.
This is the same idea read from the other side. Two cells in one region both wanting a 3, and both sitting in row 5, means row 5 gets its 3 from inside that region — so the other four cells of row 5 can drop it.
Assign two colours to a candidate pair across the grid. If one colour causes a contradiction, eliminate it to solve the pattern quickly.
Reserve this for the handful of hard boards that resist everything else. Chain the two-cell choices for a single digit, colour them alternately, and look for two cells of the same colour that can see each other. That colour is wrong, and the other one is your answer.
This board has fourteen clues and no naked singles at all. Every empty cell has at least two candidates, so the usual opening move is not available. That is the moment most solvers reach for a guess.
Instead, take the top row. It is missing 4, 5 and 6, and the 6 sitting at the start of row 2 covers both of the row's left-hand cells at once — one through the shared column, one through the shared region. So the 6 goes in the last cell of row 1. That single placement finishes column 6 down to one gap, which forces a 4, and the 4 in turn unpicks row 3 completely: 6, then 1, then 2.
Four moves in and a whole row is done. Everything left on this board falls out to plain singles from here — no colouring, no guessing, just the chain reaction that one careful first move set off.
Every board on this site ships with the same four aids. None of them solve anything for you, and all of them are optional — but the first two change how quickly you can work through the techniques above.