Star Battle — The No-Touch Experiment
Place stars in the grid so every row, column, and region has exactly the required number. No two stars can touch — not even diagonally. Pure spatial logic, no math.
Play Star Battle
6×6 grid, 1 star per row/col/region
Standard play. Timer runs. Hints available.
What is the Star Battle puzzle?
Star Battle is a constraint-based logic puzzle played on an N×N grid divided into N irregular regions by bold borders. Your goal is to place a fixed number of stars (usually 1 or 2) in each row, each column, and each region. The catch: stars cannot be adjacent to each other in any direction, including diagonally. Every star needs a one-cell buffer around it in all eight directions.
No numbers, no arithmetic, no given cells. The only information you get is the region layout and the star count. Everything else comes from deduction. That makes Star Battle a clean logic puzzle. You either see the constraint chain, or you don't. There is no partial answer from guessing a few numbers.
The puzzle was created by Hans Eendebak and gained popularity through puzzle competitions and the Nikoli puzzle community. If you have played the LinkedIn Stars or Queens games, you have already encountered a simplified version of the same mechanic. The original Star Battle goes deeper with larger grids, irregular region shapes, and 2-star configurations that demand more sophisticated reasoning.
Star Battle sits alongside Sudoku, KenKen, and Skyscrapers in our grid puzzle collection. Where those puzzles use number placement, Star Battle is binary: a cell either has a star or it doesn't. The adjacency constraint replaces arithmetic with spatial reasoning, making it a different kind of logic exercise.
We have 1,500 puzzles across five difficulty levels, from beginner-friendly 6×6 grids with 1 star per unit to demanding 10×10 grids requiring 2 stars per row, column, and region. Play star battle online free, or print them for offline solving. For a daily star battle puzzle challenge, check the daily experiment page.
1-star vs 2-star puzzles
The star count is the primary difficulty lever. In a 1-star puzzle, each row, column, and region needs exactly one star. With N stars total on an N×N grid, most cells are empty. Placing one star immediately blocks up to 8 neighbors, and that cascade usually solves most of the grid quickly.
In a 2-star puzzle, everything changes. Each row, column, and region has exactly two stars. The two stars within a region cannot touch, which constrains where they fit inside the region's shape. The adjacency constraint now interacts with itself: your first star in a region blocks cells that might be needed for the second. Reasoning chains get longer because you juggle twice as many placement constraints. Most players find the jump from 1-star to 2-star more difficult than simply increasing the grid size.
How to play
Click or tap any cell to place a star. Click again to remove it. Right-click or press X to mark cells where you know a star cannot go. Use arrow keys to navigate the grid. The row/column/region star counters around the grid update in real time so you always know where you stand.
When you place a star, all adjacent cells receive a subtle dimmed overlay showing they are blocked. This adjacency visualization helps beginners and is always on by default. It prevents the most common mistake: forgetting that diagonals count.
Three hints are available per puzzle in classic and timed modes. Each hint identifies a cell that can be definitively resolved through logic and tells you the strategy used. In challenge mode, hints and undo are both disabled.
For the full rules with a worked example, see the rules page.
Play modes
Classic
Timer counts up. Up to 3 hints. Undo available. The default way to play.
Timed Trial
Beat the countdown. Time limits scale with difficulty: 3 min for easy up to 25 min for einstein.
Challenge
No hints. No undo. Every star placement is permanent.
Star Battle tips and strategies
Technique by technique, from beginner elimination to advanced constraint chains.
Start with the smallest regions
Small regions have fewer possible star locations. A region with exactly S cells (where S is the star count) must have a star in every cell. A region with S+1 cells has only one cell to leave empty. Scan the grid for these first and resolve them immediately. On a 6×6 grid, regions of 2 or 3 cells almost always give you free placements.
Adjacency elimination
When you place a star, all eight surrounding cells are blocked. Mark those cells with × to keep track. Then re-examine every row, column, and region that intersects the blocked cells. If blocking those cells leaves a row or region with exactly the right number of valid positions for its remaining stars, fill them. This cascade effect does most of the work on 1-star puzzles.
Row and column counting
For each row and column, count how many unblocked cells remain and how many stars are still needed. If they match, every remaining open cell gets a star. If the count drops below what is needed, you made an error somewhere — backtrack. The star count indicators around the grid make this visual: watch for rows showing “0/1” or “1/2” with exactly that many open cells left.
Region confinement
If a region's remaining valid cells all fall within a single row or column, that region “claims” that row or column. No other region can place a star in that row or column within the confined stretch. This is the Star Battle equivalent of Sudoku's pointing pairs. It does not place a star directly, but it eliminates cells in other regions, which can trigger further deductions.
Reverse confinement
The flip side: if a row or column can only place its remaining stars in cells from one region, that region's star(s) must go in that row or column. Other cells in the region outside that row or column can be eliminated. Combine forward and reverse confinement and you get a chain of elimination that reaches across the grid.
2-star spacing within regions
In a 2-star puzzle, both stars in a region must respect the no-touch rule. Long, narrow regions force the stars to the ends. L-shaped regions constrain placement to specific arms of the L. Before doing any row/column analysis, scan each region and identify which pairs of cells could hold two non-adjacent stars. Often only a handful of configurations survive, and cross-referencing with row and column constraints narrows it to one.
These six techniques cover everything required for 1-star puzzles through 10×10. For 2-star grids at expert and einstein level, you will need to layer multiple constraint chains simultaneously. Mark cells diligently and let the adjacency visualization guide your eye.
Difficulty levels
Five levels scale the grid size and star count. The jump from 1-star to 2-star is a bigger difficulty increase than adding extra rows and columns.
| Level | Grid | Stars | Techniques | Time |
|---|---|---|---|---|
| Easy | 6×6 | 1★ | Direct elimination, small regions | 3 min |
| Medium | 8×8 | 1★ | Row/col counting, region interaction | 6 min |
| Hard | 10×10 | 1★ | Multi-region elimination chains | 9 min |
| Expert | 10×10 | 2★ | Two-star spacing, cross-constraints | 15 min |
| Einstein | 10×10 | 2★ | Full constraint propagation | 25 min |