Nonograms (also called Picross, Griddlers, or Hanjie) are logic puzzles where you fill in a grid to reveal a pixel art image using row and column number clues. The overlap method is the single most important technique — compare where a clue's leftmost and rightmost possible placements intersect, and mark the overlapping cells as filled. Combine that with edge logic and X-marking elimination, and you can solve any properly designed nonogram without guessing. Start with 5x5 grids, graduate to 10x10 within a week, and tackle 15x15 or 20x20 once the three core techniques feel automatic. Picross S (Nintendo, $9.99) and Nonograms Katana (free, Android/iOS) are the two apps worth starting with.
What Are Nonograms?
A nonogram is a grid puzzle. Typically square, though rectangular variants exist, where you fill in cells to create a hidden picture. The numbers along each row and column tell you how many consecutive filled cells appear in that line, and in what order. A row labeled 3 1 means there's a group of three filled cells, then at least one empty cell, then a single filled cell, somewhere in that row.
The finished result is always a pixel art image: a heart, a cat, a coffee cup, a spaceship. That's part of the appeal. Unlike Sudoku, where the completed grid is just numbers, a solved nonogram gives you something to look at. The Japanese puzzle magazine Nikoli popularized the format in the late 1980s, and Nintendo brought it mainstream with their Picross series starting in 1995 on the Game Boy. The puzzle goes by different names depending on where you encounter it. Picross (Nintendo's trademark), Griddlers (in the UK), Hanjie (from the Chinese name), Paint by Numbers, or just Nonogram.
What makes nonograms different from most puzzle games is the guarantee: every properly constructed nonogram has exactly one solution, and that solution is reachable through logic alone. No trial-and-error needed. No lucky guesses. If you understand three techniques. Overlap, edge logic, and elimination. You can solve any standard nonogram, period. The grid size just changes how long it takes.
If you enjoy logic puzzles in general, you might also want to check out our roundup of offline puzzle games that work without wifi, nonogram apps are well-represented there since they don't require a server connection.
How to Read Row and Column Clues
Every row has a set of numbers on its left side. Every column has a set of numbers above it. These numbers are your only input. Understanding them precisely is the difference between solving puzzles and staring at blank grids.
Single number: A row in a 10-cell grid labeled 4 means exactly four consecutive cells are filled somewhere in that row. The remaining six cells are empty. You don't yet know where the group of four sits. It could start at position 1, 2, 3, 4, 5, 6, or 7.
Multiple numbers: A row labeled 2 3 1 means three separate groups appear from left to right, a group of two, a group of three, and a single cell. Each group must have at least one empty cell separating it from the next group. So the minimum space this clue needs is 2 + 1 + 3 + 1 + 1 = 8 cells. In a 10-cell row, that leaves only 2 cells of slack.
The zero clue: Some apps show a 0 for rows or columns that are entirely empty. Others leave the clue area blank. Either way, this means every cell in that line should be marked empty (X). These free lines are your first move. Mark them immediately.
Full-line clue: If a row in a 10-cell grid is labeled 10, the entire row is filled. If it's labeled 4 5, that requires 4 + 1 + 5 = 10 cells, which means the placement is fully determined, no slack at all. Fill the whole row: four filled, one empty, five filled.
The critical concept here is minimum length. For any clue, add up all the numbers plus one gap between each pair of numbers. A clue of 3 2 4 requires at minimum 3 + 1 + 2 + 1 + 4 = 11 cells. If the row is 15 cells wide, you have 4 cells of slack. If the row is 11 cells wide, the placement is unique. This minimum length calculation drives everything that follows.
The Overlap Method (Core Technique)
This is the technique that unlocks nonograms. Once you understand overlap, you'll go from confused to confident in about 20 minutes of practice.
The idea: take a clue, imagine sliding it as far left as possible, then imagine sliding it as far right as possible. Any cells that are filled in both positions must be filled regardless of where the actual placement ends up. Those overlapping cells are guaranteed.
Example: A 10-cell row with clue 7. Push the block of seven all the way left: cells 1-7 are filled. Push it all the way right: cells 4-10 are filled. The overlap is cells 4-7. Those four cells are definitely filled, no matter what. Mark them. You haven't solved the row, but you've established four certain cells that will help you when you cross-reference with column clues.
The math behind it: For a single clue of size n in a row of length L, the overlap produces n - (L - n) = 2n - L guaranteed cells. If this number is zero or negative, the overlap method gives you nothing for that clue. The block has too much room to slide around. A clue of 3 in a 10-cell row: 2(3) - 10 = -4. No overlap. A clue of 8 in a 10-cell row: 2(8) - 10 = 6. Six cells are guaranteed.
With multiple numbers: The same logic applies to each group independently, but you need to account for the minimum positions of all groups. For a clue of 3 4 in a 10-cell row: push everything left (cells 1-3 filled, cell 4 empty, cells 5-8 filled). Push everything right (cells 1-3? No. Cells 1-3 empty, cells 4-6 filled, cell 7 empty, cells 8-10 filled... wait). Let me be precise. Leftmost: 3-block at 1-3, gap at 4, 4-block at 5-8. Rightmost: 3-block at 3-5, gap at 6, 4-block at 7-10. Overlap of first group: cell 3. Overlap of second group: cells 7-8. Mark those three cells.
When you first start solving, run the overlap method on every row and every column. On a 15x15 puzzle, this initial pass typically fills 30-50% of certain cells. Those cells then constrain other rows and columns, creating a cascade of deductions.
Edge Logic and Gap Analysis
Edge logic comes into play once you have partial information. Some cells filled, some marked empty, from earlier overlap passes or cross-referencing.
Edge expansion: Suppose a 10-cell row has clue 3, and you already know cell 2 is filled (from a column clue). The group of three must include cell 2. That means the group can only span cells 1-3, 2-4, or possibly other positions containing cell 2. But now consider: if cell 1 is also filled, the group must start at cell 1 (since cells 1 and 2 are filled and must belong to the same group of 3), so cell 3 is also filled. That's edge expansion, a known cell at the edge of the grid extends the group in the only possible direction.
Gap analysis: If you have a row with clue 5 and cells 3 and 7 are both marked as X (empty), the group of five must fit entirely within cells 1-2 (too short. Only 2 cells), cells 4-6 (too short — only 3 cells), or cells 8-10 (too short. Only 3 cells). None of those segments are long enough. That means one of your assumptions was wrong, or more likely, you'll reach this deduction from the other direction and realize that cells 3 or 7 can't be marked X because the clue wouldn't fit. Gap analysis is about checking whether remaining empty segments can actually accommodate the clue groups.
Pinching from both ends: When X marks accumulate on both sides of a row, they shrink the effective row length, which makes the overlap method more powerful. A clue of 4 in a 10-cell row gives no overlap initially (2(4) - 10 = -2). But if cells 1-2 and 9-10 are marked X, the effective length drops to 6 cells, and the overlap becomes 2(4) - 6 = 2 guaranteed cells. This is why solving nonograms accelerates. Each piece of information makes every other deduction stronger.
Split groups near edges: In a 10-cell row with clue 1 6, the minimum length is 1 + 1 + 6 = 8. Push left: cell 1 filled, cell 2 empty, cells 3-8 filled. Push right: cell 3 filled, cell 4 empty, cells 5-10 filled. The first group's overlap: cell 1 only in leftmost, cell 3 only in rightmost. No overlap. The second group's overlap: cells 5-8 (four cells). But there's an edge deduction too: if from a column clue you learn cell 1 is filled, then the first group must sit at cell 1, which means cell 2 is definitely empty, which constrains the second group further.
When to Mark X's (Elimination)
Marking cells as empty is just as important as filling them. Every X narrows down possibilities and feeds back into the overlap and edge logic techniques. Some solvers neglect X-marking and then wonder why they're stuck.
Completed groups: If a row's clue is 3 and you've identified the three filled cells (say, cells 4-6), then every other cell in that row is empty. Mark cells 1-3 and 7-10 with X. This is the most common source of X marks and the most satisfying. It closes off the row entirely.
Boundary gaps: Between two groups of filled cells that belong to different clue numbers, there must be at least one empty cell. If you've determined that cells 2-4 belong to the first group and cells 7-9 belong to the second group in a 3 3 clue, then cell 5 (or cells immediately adjacent to each group) must be empty. Mark those gaps.
Unreachable cells: Sometimes a cell can't belong to any group. Consider a row with clue 2 2 in a 7-cell row. Minimum length: 2 + 1 + 2 = 5. Leftmost placement: cells 1-2, gap, cells 4-5. Rightmost: cells 4-5, gap, cells 6-7. Wait, actually, rightmost is: 2-block at 4-5, gap at 6? No. Let me recalculate. Push right: second group at 6-7, gap at 5, first group at 3-4. So cell 5 is empty in the leftmost placement (it's a gap), and cell 5 is also empty in the rightmost (it's a gap). Cell 5 is guaranteed empty. Mark X.
Zero rows and columns: As mentioned earlier, a clue of 0 means the entire line is empty. Mark every cell with X immediately. These are free information. Always handle them first.
Pro tip: After filling any cell, immediately check its entire row and its entire column for new deductions. Many solvers focus on rows, forget to check columns, and miss easy eliminations. Discipline yourself to look both ways every time.
Color Nonograms vs Black-and-White
Standard nonograms are black-and-white: cells are either filled or empty. Color nonograms add a twist, cells can be filled with different colors, and the clue numbers are color-coded to tell you which group is which color.
The key difference: same-color groups still need a gap between them, but different-color groups don't. In a black-and-white puzzle, 2 3 always requires at least 6 cells (2 + 1 gap + 3). In a color puzzle, if the 2 is red and the 3 is blue, they only need 5 cells because the color change itself acts as the separator. A red cell followed by a blue cell is unambiguous without a gap.
This has practical consequences for solving. The minimum length calculation changes depending on which adjacent groups share colors. A clue of 2 3 1 (red, blue, red) needs 2 + 3 + 1 = 6 cells minimum, because neither boundary is same-color. But 2 3 1 (red, red, blue) needs 2 + 1 + 3 + 1 = 7 cells minimum, because the first two groups are both red and require a gap.
Color nonograms produce much more detailed pixel art, recognizable faces, landscapes, animals with shading. They're also generally easier to solve despite looking more complex, because the color information eliminates ambiguity faster. If a cell is filled with blue from a column clue, that immediately tells you which group it belongs to in the row clue.
My recommendation: start with black-and-white to build solid technique, then move to color once 15x15 black-and-white puzzles feel routine. The color versions are more visually rewarding, and the solving experience is smoother because you have more information to work with.
Nonogram App Recommendations
I've tried about a dozen nonogram apps over the past year. Some are polished, some are ad-riddled, and some contain puzzles that require guessing (a design flaw, not a feature). Here's what's actually worth your time.
| App | Platform | Price | Grid Sizes | Color? | Best For |
|---|---|---|---|---|---|
| Picross S series | Nintendo Switch | $9.99 each | 5x5 to 20x15 | B&W + Color | Polished controls, curated puzzles |
| Nonograms Katana | Android / iOS | Free (ads) / $3.99 ad-free | 5x5 to 50x50+ | B&W + Color | Huge puzzle library, user-created content |
| Hungry Cat Picross | Android / iOS | Free (ads) / $2.99 ad-free | 5x5 to 15x15 | B&W only | Charming art, gentle difficulty curve |
| Picross Luna | Android / iOS | Free (IAPs) | 5x5 to 20x20 | B&W only | Story-driven, thematic puzzle packs |
| Nonogram.com | Android / iOS / Web | Free (ads) / $6.99 ad-free | 5x5 to 20x20 | B&W + Color | Clean interface, daily challenges |
| Paint it Back | iOS / Steam | $4.99 | 5x5 to 25x20 | B&W only | Museum theme, no ads, pay once |
My pick for beginners: Nonograms Katana. The free version has thousands of puzzles sorted by difficulty, the controls are responsive, and the difficulty tags are accurate. You can filter by grid size, which means starting with 5x5 puzzles and stepping up when you're ready. The community-created puzzle section adds near-infinite replay value, though puzzle quality varies. Stick to the curated packs at first.
If you own a Switch: Picross S (any entry in the series. There are nine as of early 2026) is the gold standard. Jupiter, the developer, has been making Picross games for three decades. The controls use both touchscreen and buttons, the hint system teaches without spoiling, and every puzzle is guaranteed solvable through logic. The Mega Picross mode in later entries adds a fascinating twist where clue numbers can span two adjacent rows or columns.
If nonograms click for you, you might also enjoy the logic puzzle recommendations in our brain training apps comparison. Several of those apps include nonogram-style spatial reasoning exercises.
Common Mistakes (Guessing vs Logic)
The number one mistake beginners make is guessing. You stare at a cell, think "it's probably filled," mark it, and build further deductions on that assumption. Three minutes later, you hit a contradiction, and now you don't know which of your earlier marks was the wrong one. The entire grid becomes unreliable.
Rule: never fill a cell unless you can explain why it must be filled. "It overlaps in all possible placements" is a valid reason. "It feels right" is not. If you catch yourself guessing, stop and look for a different row or column where the logic is clearer. There's always one, in a well-made puzzle, the solution path never requires guessing at any point.
Mistake #2: Ignoring X marks. New solvers focus on filling cells and forget that marking cells as empty is equally productive. Every X shrinks the available space, which strengthens overlap calculations. When you're stuck, instead of hunting for cells to fill, ask "are there cells I can definitively mark as empty?" Often there are several.
Mistake #3: Working one row at a time. Nonograms are inherently two-dimensional. Information from columns feeds into rows and vice versa. After making any mark in a row, check the affected column immediately. Solving one row completely before touching another is possible in tiny puzzles but breaks down in anything larger than 10x10. The fastest solvers alternate constantly between rows and columns, following wherever the logic chain leads.
Mistake #4: Starting with the wrong rows. Don't begin with a row clue of 1 in a 15-cell grid. That gives you zero information until other clues narrow things down. Start with the biggest numbers relative to the grid width. A clue of 12 in a 15-cell grid gives you 9 guaranteed cells from overlap alone. Scan every row and column, and begin with the ones that yield the most overlap.
Mistake #5: Not double-checking completed rows. When you think you've placed all filled cells in a row, verify by counting the groups and comparing against the clue. A group of 4 that you accidentally extended to 5 will cascade errors throughout the grid. Fifteen seconds of verification saves fifteen minutes of backtracking.
Mistake #6: Giving up too early on big grids. A 20x20 grid looks overwhelming when it's blank. But after the first overlap pass, it typically has 40-80 cells determined. After the second cross-reference pass, another 30-50 fill in. The puzzle gets progressively easier as you solve it — the hardest moment is always the beginning. I timed myself on a batch of ten 15x15 puzzles from Nonograms Katana: the first pass took 4-5 minutes on average, and the remaining solving took 3-4 minutes. The front-loaded difficulty is part of the learning curve.
Putting It All Together: A Solving Workflow
Here's the step-by-step process I follow for every new nonogram, whether it's 5x5 or 30x30:
Step 1. Handle freebies. Mark any zero-clue rows/columns entirely with X. Fill any rows/columns where the clue's minimum length equals the grid width.
Step 2, Run overlap on every line. Go through each row and column, calculate the overlap for each clue group, and mark guaranteed cells. On a 15x15 grid this takes about 3 minutes and typically resolves 30-50% of cells.
Step 3. Cross-reference. For every cell you just marked, check its perpendicular line (if you filled a cell from a row clue, check its column). Apply overlap to the now-constrained line. Repeat until no new cells emerge.
Step 4. Edge logic. Look for filled cells near row/column boundaries. Extend groups from edges. Mark boundary gaps between identified groups.
Step 5. Elimination sweep. Identify completed groups and mark remaining cells in those rows/columns as X. Check for unreachable cells in partially solved lines.
Step 6, Repeat steps 3-5 until the grid is complete. If you get stuck, scan every row and column from scratch. You'll often spot a deduction you missed.
That's genuinely all there is to it. No memorized patterns, no lookup tables, no advanced math. Three techniques, overlap, edge logic, elimination. Applied repeatedly until the picture appears. The first few puzzles take patience. By your twentieth, the techniques fire automatically and the puzzle becomes meditative rather than effortful.
Nonograms are one of the few puzzle types where difficulty scales cleanly with grid size. A 5x5 takes under a minute. A 10x10 takes five to eight minutes. A 15x15 takes ten to fifteen. A 20x20 takes twenty to thirty. And a 30x30 or larger can occupy a satisfying hour. The techniques stay the same, the grid just gives you more to work with.
Frequently Asked Questions
Do nonograms ever require guessing, or can every puzzle be solved with pure logic?
Every properly constructed nonogram has a unique solution reachable through logic alone. If you find yourself needing to guess, either the puzzle is poorly designed (some user-created puzzles in apps like Nonograms Katana have this problem) or you've missed a logical deduction somewhere. Professional puzzle creators. Jupiter (Picross S), Nikoli, and most curated apps. Guarantee uniquely solvable puzzles. If a puzzle has multiple valid solutions, the clues are ambiguous by definition, which breaks the fundamental promise of the format. When you're stuck, walk away for a few minutes and re-scan every line fresh. I've never encountered a curated nonogram that actually required guessing.
What's the difference between Picross, Nonograms, Griddlers, and Hanjie?
They're all the same puzzle. "Nonogram" is the most common generic name, coined after Non Ishida, one of the puzzle's inventors, who published the first examples in 1987 in Japan. "Picross" is Nintendo's trademark (short for "picture crossword"), used exclusively for their game series. "Griddlers" is the name used by griddlers.net and popular in the UK. "Hanjie" comes from a Chinese adaptation. "Paint by Numbers" is sometimes used in puzzle magazines. The rules, techniques, and solving experience are identical regardless of what the puzzle is called. If an app says "Picross-style" or "Nonogram puzzle," it's the same thing.
How long does it take to get good at nonograms?
Most people can solve 5x5 puzzles comfortably after 3-5 completed grids, which takes about 30 minutes. Moving to 10x10 puzzles typically takes a few days of casual practice. Maybe 10-15 puzzles total. The jump to 15x15 is where the overlap method really needs to become automatic rather than something you consciously calculate, which takes another week or so of daily puzzles. I started with 5x5 grids in Nonograms Katana, was solving 10x10 after about 4 days, and felt comfortable with 15x15 after roughly two weeks. The progression is steady, unlike some puzzle types where you hit a skill plateau, nonograms keep opening up as your pattern recognition for overlap situations improves.
Are nonograms actually good for your brain?
Nonograms exercise spatial reasoning, working memory (holding partial information across rows and columns), and logical deduction. Whether that translates to broader cognitive benefits is debatable. The same "transfer" question applies here as with any brain training activity. What's less debatable is that nonograms require sustained focused attention for 10-30 minutes per puzzle, which is a genuine mental workout. They're not going to raise your IQ, but they're a more mentally demanding use of screen time than scrolling social media. A 2019 study in Frontiers in Aging Neuroscience found that regular puzzle-solving (crosswords and number puzzles, which share mechanisms with nonograms) correlated with better cognitive function in adults over 50, though correlation isn't causation.