Nonogram Techniques: From Overlap Counting to Never Guessing
Here's a move you can make in a nonogram you've never seen before, with zero deduction: the row is 10 wide and its clue is a single 7. Push the run as far left as it can sit — cells 1 through 7. Now as far right — cells 4 through 10. Compare. Both extremes, and every placement in between, cover cells 4, 5, 6 and 7. Wherever that run of seven eventually lands, those four cells are inside it. Fill them before you've read anything else on the grid.
That's overlap counting, the first rung on a short ladder of techniques that solves every proper nonogram in existence. Proper is worth stressing. These puzzles ship with a unique solution reachable by logic alone — a standard the format has carried since it came out of Japan in the late 1980s, where the designer Non Ishida turned skyscraper window-grid art into a puzzle and the genre ended up carrying her name. No rung below ever requires luck. That's the contract.
Rung 1: overlap counting
The arithmetic behind the opening example: double the clue, subtract the line length, and whatever's left is the number of guaranteed cells in the middle. A 7 in a 10-wide line leaves 4. An 8 leaves 6. A 6 leaves 2. A 5 leaves nothing — the two extreme placements just miss each other, and that line keeps its secrets for now.
Multi-clue lines work the same way once you pack them. For clues 4 and 3 in 10 cells, shove everything left (cells 1–4 and 6–8), then everything right (3–6 and 8–10). The first run always covers cells 3 and 4; the second always covers cell 8. Three fills off a completely cold line.
Sweep the entire grid this way before doing anything else, biggest clues first. Any clue longer than half its line pays out immediately, and those opening fills are fuel for every rung below.
Rung 2: edge anchoring
Borders squeeze runs, and squeezed runs leak information. If the first cell of a clue-4 row turns out filled, the run has nowhere to hide: cells 1 through 4 fill and cell 5 is definitely blank. Even one step of contact helps — a filled cell in column 2 of a clue-5 row forces cells 2 through 5, because every legal placement of a five that touches column 2 covers them.
Edges close doors too. When the run nearest a border is finished, cap it — the cell beyond each end of any completed run is empty by definition. And a gap that's too small for anything is dead space: a 2-wide pocket against the wall in a row whose smallest remaining clue is 3 will never hold a thing. Kill it now, on purpose.
Rung 3: the X does half the work
Fills get the glory; blanks win the puzzle. Every certainty about emptiness deserves a mark — flip on Mark mode and plant an X — because an X isn't bookkeeping, it's ammunition. A clue of 5 in a 10-wide row gives no overlap at all. But drop an X in cell 3, learned from some column, and the run is suddenly confined to the seven right-hand cells, where overlap counting pays out three fills. One blank bought three squares.
The rhythm to build: finish a clue, X the rest of its line immediately; place an X, re-run the overlap arithmetic on the shorter segments it just created. Segments are the quiet secret of the whole format — every X splits a line into smaller lines, and small lines give up their cells fast.
Rung 4: cross-reference relentlessly
No line solves alone. Every fill you make in a row lands in some column and tightens it; every column X shortens a row somewhere. So don't grind one line to exhaustion — take the certain deductions, then follow the cells you just changed into their crossing lines while they're fresh. Solving a nonogram feels less like reading rows top to bottom and more like chasing a spark around the grid.
When you're choosing where to look next, the priority order is: lines that just received new information, lines that are nearly complete, lines with huge clues. A fresh X in a nearly-solved column beats an untouched empty row every time. If you've read our sudoku techniques piece, this is the same muscle — every placement is information somewhere else. The pace is lopsided by design: the first tenth of the grid takes half your time, and the last half falls in a rush.
Rung 5: never guess. Actually never
Every rung above produces certainties. A guess produces a lie that looks like one. Wrong fills don't fail loudly — the grid accepts them, and they sit there quietly poisoning every deduction that later touches their row or column. Ten minutes on you hit a contradiction, and by then forty marks descend from the bad one with no way to tell which forty. The logic chain isn't decoration; it's the only thing separating knowledge from noise, and one guess snaps it.
Stalled means exactly one thing: a deduction is still sitting somewhere, usually an overlap opened up by your latest X. Go find it. This is what separates the nonogram family from a game like Minesweeper, where the odd forced guess is genuinely part of the deal. Here the no-guessing contract is absolute — the same one Slitherlink makes about its single loop and Hidato makes about its number chain. The information is always sufficient. Trust it, mark your blanks like they're worth points, and the picture will walk out of the grid on its own.