Check Tic-Tac-Toe Winner, Draw, or Game in Progress

It reports whether X or O has won, whether a filled board is a draw, or whose turn comes next while play remains possible. The checker also rejects malformed boards and impossible X-to-O turn counts, so it is useful for validating saved games, testing game logic, reviewing exercises, and powering small applications without relying on guesswork or visual inspection.

Enter the board in a clear three-line format

Provide exactly three rows, with exactly three cells in every row. Use X for the first player, O for the second player, and a hyphen for any empty square. You may place spaces between cells for readability, as in X O X, because whitespace inside each row is ignored. Keep each board row on its own line. The checker treats lowercase x and o as their uppercase equivalents, but it rejects every other cell character. This strict format makes the input easy to type, paste, store in a fixture, or generate in an application while avoiding ambiguous symbols. The first player is always X, matching standard tic-tac-toe turn order. A blank board is therefore three rows of three hyphens, and a partially played board keeps a hyphen wherever nobody has moved. If the board has too many rows, too few rows, a row with the wrong number of cells, or an unsupported mark, the request returns an input error instead of attempting to repair or reinterpret the position. That behavior keeps automated checks predictable.

Understand winner and draw detection

The checker examines all eight possible winning lines: three horizontal rows, three vertical columns, and the two diagonals. A player wins when one or more of those lines contains three of the same mark. A winning response identifies the player and returns every winning line as row-and-column coordinate pairs, using one-based coordinates so the top-left square is row 1, column 1. Returning all matching lines matters in positions where a final move completes two lines at once. If neither player has a winning line and every cell is occupied, the result is a draw. If empty cells remain, the game is reported as in progress and the response identifies the next player from the mark counts. The algorithm is deterministic and scans only the nine cells and eight known lines. It does not use randomness, network services, stored state, or assumptions about moves that are not represented on the board. The same board always produces the same structured result, making the capability suitable for unit tests and repeatable validation.

Catch positions that could not occur in legal play

A board can look visually plausible while violating the order of play. Since X moves first and players alternate, X must have either the same number of marks as O or exactly one more. The checker rejects a position when O has moved more times than X or when X leads by two or more moves. It also cross-checks a declared win against those counts: an X win requires one extra X mark, while an O win requires equal counts. A board on which both players have winning lines is rejected because play should have stopped after the first valid win. These checks are especially valuable when boards come from user input, database records, tutorials, interview exercises, or a custom game engine under development. They prevent an impossible position from being mislabeled as a valid winner or draw. The capability evaluates the supplied snapshot rather than reconstructing a complete move history, so its validation focuses on the facts a 3x3 board can prove directly. For a valid unfinished position, use next_player to continue with the correct mark.