Lights Out Solver

Fact one: pressing a tile twice cancels out. Whatever the first press did, the second undoes it. So no useful solution ever presses the same tile twice, and a solution is not a sequence at all. It is a set of tiles.

Fact two: order does not matter. Toggle operations commute; pressing tile A then tile B lands the board in exactly the same state as B then A. Together the two facts mean every board is really asking one question: which subset of tiles, pressed once each, turns everything off?

That is why the game feels different from every other puzzle on this site. There is no feedback loop and no partial credit. The answer exists or it does not, and once you know the set, nothing can go wrong executing it. The result is oddly calming, a stillness that holds even on a stubborn board.

Under the hood, this is the cleanest math on the site. Each light contributes one equation: "the number of presses among this light and its neighbors, counted modulo two, must equal its current state." On or off, one or zero: the entire puzzle lives in a number system with two elements.

Stack all those equations and you get a linear system, and the solver runs Gaussian elimination on it, adapted for that two-value arithmetic. The output is a provably correct press set: if it says press these tiles, pressing exactly those clears every light, on any board size, computed in milliseconds.

What makes it trustworthy, and the reason it never "fails" the way hint systems do, is that there is no heuristic anywhere in it. It is not pattern-matching against known boards or guessing promising regions. It is the same elimination you would do by hand with unlimited patience, done instantly, and provably minimal in structure.

You do not need linear algebra at the table. The classic manual strategy is called chasing, and it works on every solvable board: start at the top row, and for each light that is on, press the tile directly below it. That row is now dark. Move down a row and repeat, pushing the surviving lights downward until only the bottom row can be lit.

The bottom row is where the chase either finishes or stalls, and here is the trick: the pattern of lights remaining in that bottom row tells you exactly which tiles to press in the top row. Run the chase again with those top-row presses in place, and the whole board goes dark. The mapping from bottom-row patterns to top-row presses is fixed for each board size, and for a 5×5 the common ones are worth memorizing.

Geometry matters while you chase: a center press flips five tiles and an edge press flips four. A corner press flips three. Corners are the easiest tiles to reason about and the cheapest to fix. When you are stuck mid-chase, re-verify the corners first, because they are the tiles the eye most often skips.

Some boards have a unique press set. Some have several equally valid ones, related by what players call quiet patterns, small sets of presses that cancel out entirely, like the all-on row pattern, which you can add to any solution to get another solution with the same result. If you use the solver and get a different press set than another player did, you can both be right.

And some boards have no solution at all. On classic 5×5 grids only a fraction of configurations are solvable. The solver detects these directly: the elimination has no consistent answer, and it tells you so instead of inventing one.

No-solution detection is the single most valuable feature here. Without it, every failure looks like player error, inviting random tile presses for minutes on end. Knowing that some states are mathematically dead ends is not defeatism. It is the difference between searching and thrashing.

The solver handles everything from 3×3 minis to the classic 5×5 and larger custom layouts, because the elimination just scales: bigger board, bigger system, same two-value arithmetic, same exact answer.

What changes by hand is the feel. Small boards have few possible states and can seem random, almost scrambly; the 5×5 has enough structure for the chase to feel like a method rather than a shuffle. The bottom-row mapping differs per size, so a memorized 5×5 table is useless on a 4×4. The first few runs on any new size should go through the solver until the table is committed to memory.

If you are learning, start at 3×3 deliberately: the chase is short enough to hold in your head, and every concept (sets, quiet patterns, dead ends) shows up in miniature. The recommended path for anyone coming to the math late: small board first and math second.

Lights Out lives everywhere except the front of the shelf: as a minigame inside larger games, in puzzle collections, in speedrun categories, and in math classrooms as the friendliest possible introduction to linear algebra over finite fields. Every one of those settings is a good reason to have a solver bookmarked.

The classroom use is a strong one. Set up a board, solve it by hand with the chase, then run the solver and compare: the press set is a worked solution to a real linear system, and seeing elimination produce a set of tiles you can physically press makes the abstraction land in a way textbook exercises rarely do.

The speedrun angle is real: knowing the exact press set ahead of time turns the run into motion practice instead of problem-solving. Whether that fits the spirit of the category is a separate question. The math does not judge.