A Lights Out board looks solvable by trial and error right up until you press a tile and watch two or three others light up instead of going dark. That is usually the moment a player gives up on guessing. There is an actual method, though: a row-by-row procedure often called the chase method, and it clears any solvable board in a fixed number of steps, with no backtracking and no luck involved.
This guide covers why the chase method works, what "only parity matters" means for your clicks, how to correct the top row once the rest of the board has been swept, and how to tell a board with no solution from one that just looks hard. For the fully worked 3x3 case with its complete correction table, see the 3x3 Lights Out solution; for the classic 5x5 board, see the 5x5 Lights Out solution.
The Rule Behind Every Click
Every board on Lights Out works the same way: click or tap a tile and it toggles โ on becomes off, off becomes on โ along with the tile directly above, below, left, and right of it. Diagonal tiles are never affected. A tile on an edge has only three neighbors instead of four, and a tile in a corner has just two, so the exact same click does less work near the border of the board than it does in the open middle.
Lights Out ships with four board sizes โ 3x3, 4x4, 5x5, and 6x6 โ and the click rule is identical on all four. Only the size of the grid, and therefore how many rows the chase has to work through, changes from one size to the next.
Why Only Parity Matters
Press a button twice and the second press undoes the first exactly, because it flips every light the first press touched right back. That means pressing any tile an even number of times has no net effect at all, and pressing it an odd number of times has exactly the same effect as pressing it once. A solution to Lights Out is therefore never really a sequence of moves โ it is a set of tiles, each either pressed once or left alone. That is what "parity" means here: for every tile, only whether the count of presses is odd or even matters, never the exact count and never the order they happened in.
Order genuinely does not matter, and that single fact is what turns the puzzle from a search problem into a row-by-row calculation. Toggling tile A and then tile B leaves the board in exactly the same state as toggling B first and then A, because both operations are just flips of the same on-or-off values, and flips commute. So the final board after any set of presses depends only on which tiles were pressed โ never on the sequence you pressed them in. Without this property, solving would mean searching an enormous tree of possible move orders; with it, the puzzle reduces to a much smaller question: which set of tiles, pressed in any order, clears the board?
The Chase Method, Step by Step
Work from the top row down. Leave the top row exactly as the puzzle gave it to you for now โ do not press anything in it yet. For every tile still lit in the top row, press the tile directly beneath it in the second row. Because the only tile that reaches up into the top row from below is the one directly underneath each lit tile, this clears the top row completely, and nothing pressed further down the board can ever re-light it.
Now look at the second row, which has just changed because of whatever you pressed to clear the first. For every tile still lit there, press the tile beneath it in the third row. Repeat this all the way down the board: clear the row above by pressing into the row below it, one row at a time, until you reach the second-to-last row.
By the time you finish, every row except the very last one is completely dark. The bottom row is the one row nothing was ever pressed into directly, so whatever is still lit there is the leftover from every choice made above it โ mostly from whatever you did, or did not, press in the top row before starting the chase. If the leftover bottom row is already dark, the board is solved. If it is not, that does not mean the board is unsolvable; it means the top row needed a different opening press, which the next section corrects.
Correcting the Top Row
The top row is the only genuinely free choice in the whole method โ every press after it is forced by whatever is still lit in the row above. Because of that, running the chase once with the top row left untouched and then reading what is left glowing in the bottom row tells you, indirectly, what the top row should have been pressed instead.
For a board of width w, there are only 2 to the power of w possible ways to press the top row โ 8 on a 3-wide board, 32 on the classic 5-wide board. Each one leaves its own fixed leftover pattern in the bottom row when tested on an otherwise empty board, so simulating every one of those combinations and matching the leftover you actually see gives a direct lookup: a specific leftover pattern in, one specific top-row correction out. That lookup is exactly what the correction tables on the 3x3 and 5x5 solution guides provide โ they were built by simulating every possible top-row press on an empty board of that size and recording exactly where the leftover lands, so the table is a verified lookup rather than a rule of thumb.
Once you know the corrected top row, start over from the board's original state, press that corrected pattern first, and repeat the exact same forced chase downward. A solvable board finishes completely dark on this second pass.
Which Boards Actually Have No Solution
Not every conceivable pattern of lit and dark tiles can be cleared, and whether a given board is solvable depends only on its size, not on skill. Testing every one of the possible starting patterns gives an exact, verified answer for each size: on a 3x3 board or a 6x6 board, literally every possible pattern has a solution โ there is no way to construct an unsolvable board at either of those two sizes. On the classic 5x5 board, exactly one in four of all possible patterns is solvable; the other three in four cannot be cleared by any sequence of presses, chase method or otherwise. A 4x4 board is even more restrictive, with only about one in sixteen of all possible patterns actually solvable.
None of that is a reason to worry about the board Lights Out hands you when you press New Game, though. The game builds every fresh board by simulating real random button presses starting from a fully dark board, not by drawing an arbitrary pattern out of thin air, and a board built that way is always inside the solvable set by construction, regardless of size. The only realistic way to end up with a genuinely impossible board is to build one tile by tile yourself, outside the game's own New Game button โ worth knowing if a friend hands you a hand-drawn pattern and dares you to solve it, but not something to worry about during normal play.
Common Mistakes That Break the Method
- Chasing individual lit tiles instead of full rows. Reacting to whichever tile is lit right now, without a row-by-row plan, usually creates more lit tiles than it removes.
- Pressing a top-row tile mid-chase. Once you start clearing the second row downward, every further press has to land directly below the row you are currently clearing โ going back up undoes the plan.
- Assuming a hand-drawn pattern is automatically solvable. A board built tile by tile rather than through New Game has a real chance of falling outside the solvable set, especially at 4x4 or 5x5.
- Pressing the same tile twice on purpose. The second press cancels the first exactly, wasting a move without changing the outcome.
Frequently Asked Questions
What does the chase method actually do?
It clears every row except the last one by pressing, for each tile still lit, the tile directly below it in the next row down, working from the top of the board to the bottom.
Why does the top row need correcting afterward?
Because the top row is the only row not cleared by a press from above โ it is the one genuinely free choice in the whole method, and the leftover pattern left in the bottom row after a first pass tells you, through a lookup table, what that opening choice should have been.
Does the order I press tiles in change the result?
No. Every press is a simple toggle, and toggles do not interfere with each other based on order โ the final board depends only on which tiles were pressed, never on the sequence.
Can a Lights Out board really have no solution?
Yes, depending on size. Every possible pattern is solvable on the 3x3 and 6x6 boards, but on the classic 5x5 board only one in four of all possible patterns is, and on 4x4 only about one in sixteen. Boards generated by the New Game button are always solvable regardless of size, since they are built from real simulated presses.
Do I need to memorize the whole board to solve it?
No. The chase method only ever asks one question at a time โ is the tile directly above lit? โ so you never have to hold the full board in memory, just work down one row at a time.
Related Guides
- 3x3 Lights Out Solution โ the complete correction table for the smallest board.
- 5x5 Lights Out Solution โ the correction table for the classic board size.
- How to Play Lights Out โ the rules and an overview of the chase method.
- Best Puzzle Games โ where Lights Out fits among the site's other logic puzzles.
- Play Lights Out โ put the chase method into practice on a real board.