Light Switch: move the problem down the board.
Learn row-by-row light chasing, then replay a worked board to see why a different top-row choice matters.
A Light Switch board can look chaotic because a single move changes several squares at once. The useful simplification is that every effect is local and predictable. A switch changes itself and its immediate horizontal and vertical neighbors. Nothing farther away is affected.
Start with a single cross
On an otherwise dark board, selecting the center lights up a cross. In the diagram below, filled circles are on and dots are off. Selecting the center a second time turns the entire cross off again.
· · · · · · · ● · · · ● ● ● · · · ● · · · · · · ·
This gives you the first rule: two selections of the same square cancel. The second rule is that switch order does not affect the final pattern. Selecting A then B has the same final effect as selecting B then A, because each selection reverses a fixed set of lights.
Clear the first row from below
Look at the first row only. For each lit square, select the square directly below it in the second row. A selection in row two affects exactly one light in row one: the square immediately above. It cannot disturb the other squares in that first row.
If the first row begins on, off, on, off, off, select row two at columns one and three. The first row is now dark. The lower rows have changed, but that is expected. You have moved the unresolved work downward.
Repeat without revisiting completed rows
After clearing row one, inspect row two and select directly underneath its lit squares in row three. Continue with rows three and four. Once a row is dark, this process does not change it again because later selections happen too far below it.
At the end, only the bottom row can remain lit. This is progress: you have reduced a two-dimensional pattern to one row of unresolved information. It is not yet a complete solver, because there is no sixth row to press underneath the bottom row.
What the remaining pattern means
A remaining bottom-row pattern means you need different choices in the top row before chasing downward. You can try a top-row pattern, run the same row-chasing process, and check the result. There are 32 possible on-or-off choices for five top-row switches, including pressing none.
For a solvable board, testing those top-row choices and chasing each one is a finite solving method. Keep each trial separate by starting from the same original board; combining partial trials makes the comparison meaningless. In the live game, Retry this board restores the current starting pattern and resets the move count. New puzzle generates a different board, so it is not the right control for comparing two attempts. Keep your own note of the top-row choices you have tried.
Replay a complete worked example
Choose Load guide example below the live board. It loads the fixed pattern below. Rows are numbered 1–5 from top to bottom, and columns 1–5 from left to right. A filled circle is on; a dot is off. This teaching pattern is made with three switches rather than the nine used for a random new puzzle.
C1 C2 C3 C4 C5 R1 ● ● ● · · R2 ● ● ● ● · R3 · ● ● · · R4 ● ● ● · · R5 · ● · · ·
For the first attempt, press nothing in row one. Chase each row using the selections below. These are switches to press, not a list of lights that stay on. The state produced by each step determines the next step.
- Press row 2, columns 1, 2, and 3 to clear row 1.
- Press row 3, columns 1 and 3 to clear row 2.
- Press row 4, columns 3 and 4 to clear row 3.
- Press row 5, column 5 to clear row 4.
You have made eight moves. Rows 1–4 are dark, but the bottom row is off, on, on, off, on. That result does not mean the puzzle is impossible: it rules out this top-row choice for this board.
Change one starting choice
Choose Retry this board to restore the example. This time, press row 1, column 1 before chasing. Only column 3 is now lit in row one, so press row 2, column 3. Row two is already dark after that move; do not press anything in row three. Row three has just column 2 lit, so press row 4, column 2. All 25 lights are now off. The successful sequence was R1C1, R2C3, R4C2: three moves.
The first switch is supplied here to demonstrate the method; chasing alone did not tell us to choose it. On a different starting board, that same choice may fail. Systematic trials test up to 32 top-row choices, always restoring the original board between trials. They find a solution for a solvable board, but do not by themselves promise the shortest solution.
A useful practice goal
Do one careful chase and predict every change in the row above your selection. If the prediction is wrong, stop and check which neighbors a switch affects. Our new puzzles are generated by pressing switches on an all-dark board, which ensures a solution exists. The lesson is to replace random presses with a controlled process, even before you can finish every board efficiently.