Star Battle Strategies:
Techniques Every Solver Should Know

That cascade is what makes Star Battle one of the most elegant logic puzzles ever designed. No arithmetic. No word knowledge. Just spatial reasoning and the relentless power of elimination.

But between "I understand the rules" and "I can solve Expert grids" lies a gap -- not of talent, but of technique. This guide bridges it. We will work through every major solving technique, from fundamental adjacency elimination to advanced region interaction chains.

Star Battle -- which you may know as Starstruck on Netflix -- comes in several grid sizes, each divided into N colored regions:

Place exactly one star (or two, in 2-star variants) in every row, every column, and every region. Stars cannot touch each other in any direction -- horizontally, vertically, or diagonally. Each star eliminates all eight surrounding cells.

This is where every Star Battle solution begins. When you place a star, you can immediately mark all eight surrounding cells with an X -- they cannot contain a star. On a 5x5 grid, a single star in the center cell eliminates eight cells. That is nearly a third of the grid, gone in one move.

But adjacency elimination also works backward. If placing a star in a particular cell would make it impossible for an adjacent row, column, or region to get its required star, then that cell cannot hold a star. Think of each star as claiming a 3x3 block of influence, not just one cell.

If adjacency elimination is the engine, region analysis is the steering wheel. Count the available cells in each region. If a region has been reduced to exactly one cell that could hold a star, that cell is forced. Place the star, apply adjacency elimination, and watch the cascade begin.

Start every puzzle by scanning for the smallest regions. A three-cell region has at most three options for its star. After adjacency eliminations from other progress, that number drops fast.

Some regions have a special property: all their cells fall within a single row or column. These are bottleneck regions, and they are solving gold. A region whose three cells all sit in row 4 means row 4's star must be one of those cells. Any other region with cells in row 4 can safely eliminate those cells -- the row is spoken for.

Make it routine: after every wave of eliminations, scan each region and note how many viable cells remain. When a region drops to one viable cell, you have a forced placement. When it drops to zero, you have made an error -- time to backtrack.

Nothing is forced. The smallest region still has three viable cells. You feel stuck. Here is what to do: stop looking at regions and start counting rows.

Rows and columns also need exactly one star, and they provide an independent axis of deduction that cross-references beautifully with region analysis. After a round of eliminations, examine each row. If only one viable cell remains, that cell must hold a star.

The real power emerges when you combine row counting with region analysis. If a row has only two viable cells and both are in the same region, that region's star must be in that row -- eliminate all of the region's cells in other rows. The reverse also works: if a region's star must be in a specific row, no other region can use that row.

Try a quick challenge: Region C has only two viable cells left, at (row 2, col 1) and (row 2, col 4). Region D has a viable cell at (row 2, col 5). Can Region D use that cell?

No. Both of Region C's options are in row 2, so C's star must be somewhere in row 2. Since row 2 can only hold one star, it is spoken for. Region D's cell in row 2 is eliminated.

This is overlap analysis: when every possible position for a region falls within a single row, that region claims that row. No other region can place a star there. The same logic applies to columns.

The core idea here is pairing: two units that share the same two options must claim both, locking everyone else out. (Sudoku solvers will recognize this as "naked pairs" -- same logic, different puzzle.)

Picture two regions claiming two rows:

• Region E's viable cells are only in rows 1 and 4
• Region F's viable cells are only in rows 1 and 4

Between them, Region E and Region F will consume the stars for rows 1 and 4. One takes row 1, the other takes row 4. You do not know which is which -- but you know that no other region can place a star in row 1 or row 4.

Any other region with viable cells in row 1 or row 4 can have those cells eliminated. The exclusion principle generalizes beyond pairs -- three regions in three rows, four in four -- but pairs are by far the most common application in 1-star Star Battle.

The order in which you attack a puzzle matters enormously. Random scanning wastes time and misses cascades. Strategic ordering multiplies your efficiency.

Every certain placement creates ripples. Those ripples create new certainties. The strongest solvers ride the cascade -- place a star, trace every elimination, check every affected region, row, and column, and only stop when the cascade dies down.

This is the technique that unlocks Expert-level grids. It combines overlap analysis and the exclusion principle into multi-step deduction chains that propagate constraints across the entire board.

1. Region A's star must be in column 3 (overlap analysis)
2. Region B cannot use column 3, forcing it into column 7
3. Region B in column 7 eliminates a cell in Region C
4. Region C's surviving cells all fall in row 6, claiming it
5. Region D is forced out of row 6, into row 2 -- position resolved

Each link is a simple deduction. But the chain propagates a constraint across the entire grid.

Start by practicing two-link chains. Once two-link chains become automatic, longer chains emerge naturally. Experienced solvers describe this as "feeling" the constraints flow through the grid -- pattern recognition built from hundreds of puzzles.

Let me walk you through a concrete 6x6, 1-star grid. Every deduction is definite, every coordinate is specific, and every star placement is logically forced.

Our puzzle has six regions labeled A through F. Region A is a large 7-cell inverted-T shape through the top-center. Region B is a tiny 3-cell L-shape in the top-left. Region C runs down the right side. Region D is a small 3-cell L-shape in the bottom-left. Region E fills the center-left. Region F runs down the center-right.

Two regions stand out: Region B (3 cells, all in columns 1-2) and Region D (3 cells, all in columns 1-2). Both are confined entirely to columns 1 and 2.

By the exclusion principle: B and D will consume both columns between them. No other region can place a star in column 1 or column 2. This eliminates cells from Regions A and E in one stroke -- Region E drops from 8 viable cells to just 3.

Region E's three surviving cells are all in column 3. By overlap analysis, E claims column 3. This eliminates Region A's column-3 cells. Region A drops to 3 viable cells: all in columns 4-5.

Region A's surviving cells are all in columns 4-5. Region F's cells are also entirely in columns 4-5. Same pattern as B and D: two regions claim both columns. Region C's surviving cells collapse to column 6 only.

Without placing a single star, we have assigned every column:

That is the power of overlap analysis and the exclusion principle: they extract maximum information from the grid's structure before you commit to specific cells.

Region B and A both have cells in rows 1-2. Testing Scenario X (B takes row 1, A takes row 2 at (2,4)): every path leads to a diagonal adjacency contradiction. Scenario X is impossible.

In Scenario Y, B takes row 2 and A takes row 1. B's star goes to (2,1) -- the only Region B cell in row 2. This fixes B in column 1, which means D goes to (6,2).

Adjacency from D at (6,2) eliminates two of Region E's three surviving cells. E is forced to (4,3) -- the only cell remaining.

If A takes column 5 at (1,5), then F must take column 4 -- but F's only column-4 cell is in row 6, already taken by D. Contradiction. A's star must be at (1,4).

With A in column 4, F must take column 5. F's column-5 cell at (2,5) is in row 2 (taken by B), so F is forced to (3,5).

The only open row is row 5. Column 6 belongs to Region C. C's column-6 cell in row 5 is (5,6). C's star goes to (5,6). All adjacency checks clear.

The puzzle strips away everything except spatial reasoning. No numbers to calculate. No words to know. Just colored regions, empty cells, and the elegant constraint that stars cannot touch.

The grid always has enough information. The strategies above are how you read it. Now go place some stars. The first one is the hardest. After that, the constellation builds itself.