The Counting Trick That Cracks Star Battle Puzzles

Most people solve their first few Star Battle puzzles using one idea: place a star, then cross out the cells it blocks. That works on easy grids, but on harder ones you'll hit a wall where no single star is forced and the no-touch rule has nothing left to eliminate. The way through isn't a sharper eye — it's a different kind of reasoning entirely. It's counting, and specifically the technique called confinement, which lets you claim whole rows and columns before you know exactly where any star sits. Master this one idea and the hard grids that used to stump you start falling apart. Let's break it down. Want to practise as you read? Play a Star Battle puzzle.

Why elimination alone runs out

The no-touch rule is great at local deductions: a star here blocks its eight neighbours there. But hard Star Battle puzzles are designed so that, at some point, no star is locally forced — every region still has several options, and crossing out neighbours doesn't shrink them further.

That's the moment to stop thinking about individual cells and start thinking about counts. Star Battle is, underneath everything, a counting puzzle: each row needs an exact number of stars, each column needs an exact number, and each region needs an exact number. Those guarantees let you make deductions about groups of cells without solving any single one.

The core idea: region confinement

Here's the technique in one sentence: if every cell where a region could still place its star lies within a single row, then that region's star must be in that row — so no other region can use that row in the cells they share.

Walk through why that's airtight. Suppose region A's only open cells all sit in row 4. Region A must place a star somewhere (every region needs one), and the only places left are in row 4. So region A's star is definitely going into row 4. But row 4 needs exactly one star (in a 1-star puzzle), and region A is going to supply it. That means every other cell in row 4 — the ones belonging to other regions — can be crossed out. You've claimed the entire row for region A, and eliminated a whole line's worth of candidates, without knowing region A's exact cell yet.

The same works for columns, and the same logic works when a region is confined to a single column instead of a row.

The reverse: line confinement

Confinement runs both directions. Sometimes it's not a region trapped in a line, but a line trapped in a region.

If a row's only remaining open cells all belong to a single region, then that row's star and that region's star are the same star. That's useful because it constrains the region from the other side: the region's star is now known to be in that row, so any of the region's cells outside that row can be crossed out. One deduction, two regions' worth of progress.

Learning to scan for both kinds of confinement — region-in-a-line and line-in-a-region — is what turns a stuck grid into a solvable one.

Scaling up: counting across several lines

The real power of counting shows on expert grids, where confinement widens from one line to several. The principle: any block of L rows must contain exactly L stars (in a 1-star puzzle). So if you can find L whole regions that fit entirely within those L rows, those regions supply all L stars — and every other region poking into that block gets nothing there.

For example, take three rows that, between them, completely contain three regions. Those three regions will place their three stars inside those three rows, which is exactly the three stars the rows are allowed. So any fourth region that dips into those rows can't place a star in them — cross those cells out. It's the same confinement logic, scaled from one line to a block, and it's how the tightest puzzles get unlocked.

How to use it in practice

Fold counting into your solving like this:

1. Solve the easy stars first with normal no-touch elimination.
2. When you stall, switch to counting. Scan each region: are its open cells all in one row or one column? If so, claim that line and cross out the rest.
3. Scan each line the same way: are its open cells all in one region? If so, the line's star is that region's star — trim the region's other cells.
4. Widen to blocks of two or three lines on harder grids, looking for sets of regions that fill them exactly.
5. Re-run local elimination after each confinement — claiming a line usually unblocks a fresh batch of forced stars.

The rhythm is: eliminate until stuck, count to break the deadlock, then eliminate again. Most hard Star Battle puzzles are just that cycle repeated.

Why this is the technique that matters

Beginners place stars; strong solvers count regions and lines. Confinement is powerful precisely because it lets you make certain, provable deductions in positions where nothing looks forced — no guessing, just careful counting. It's the same trick that powers the hardest 2-star puzzles, where the counts go to two and the confinements stack up.

Next time a Star Battle grid won't budge, don't reach for a guess — reach for a count. Find a region trapped in a line, claim it, and watch the grid open up. Play a Star Battle puzzle now, or brush up on the basics in our full strategy guide.

Frequently asked questions

What is the confinement technique in Star Battle?

Confinement is a counting technique: if all the cells where a region could still place its star lie within a single row or column, that region's star must go in that line — so every other cell in that line can be eliminated. It lets you "claim" a whole row or column for a region before knowing the exact star position.

How do you solve hard Star Battle puzzles?

Hard grids reach a point where no single star is locally forced. The way forward is counting: use region confinement (a region trapped in one line claims that line), line confinement (a line whose open cells are all in one region), and multi-line counting (a block of L rows containing L whole regions uses all its stars there). Alternate counting with normal no-touch elimination.

What is line counting in Star Battle?

Line counting uses the fact that each row and column needs an exact number of stars. By identifying which regions must supply those stars — for instance, regions entirely contained within a block of rows — you can deduce that other regions can't place stars there, eliminating their cells in that block without solving any single star.

Why doesn't eliminating neighbours always solve a Star Battle?

Crossing out a star's eight neighbours is a local deduction, and hard puzzles are designed so that, at some point, no star is locally forced. To progress you need global reasoning about counts — confinement and multi-line counting — which makes certain deductions about groups of cells that simple neighbour-elimination can't reach.