问题
I'm developing tic-tac-toe game, and I need algorithm to check when game ends(and who win). In 3x3 game I would check each possible win-situation(there is 8 capabilities). But in 7x7(needed 4 signs in a row or collumn, or diagonal)is a lot of possible win patterns.
回答1:
While a very basic approach is to look at runs in all the directions from every single cell, here are an approach then only ever checks a cell in a single "line" once. A "line" is a row, column, or diagonal that can possibly win, like in a Vegas slot machine :)
- For each "line", move to start of that "line" and;
- Set counter to 0.
- For each cell in the "line" (traversing the line in order):
- If the cell is P1 and counter is >= 0, add one to counter
- If counter = 4 then P1 wins.
- If the cell is P1 and counter is negative, set counter to 0
- If the cell is P2 and counter is <= 0, subtract one from counter
- If counter = -4 then P2 wins
- If the cell is P2 and counter is positive, set counter to 0
- If the cell is P1 and counter is >= 0, add one to counter
Important Edit: If the cell contains neither P1 or P2, reset counter to 0 (doh!). I omitted this trivial but required step. Otherwise "11-11" would be counted as a win.
The "lines" can be traversed given a starting point and row/column offset per iteration (e.g. start at (0,0) and advance (1,1) for longest diagonal from NW to SE). Diagonals with lengths less than 4 can avoid being checked entirely, of course.
Happy coding.
回答2:
If you are using a bitboard for each player, you can use bit shift operations to test a board for a win.
The bitboard would have following structure:
6 14 22 30 38 46 54
5 13 21 29 37 45 53
4 12 20 28 36 44 52
3 11 19 27 35 43 51
2 10 18 26 34 42 50
1 9 17 25 33 41 49
0 8 16 24 32 40 48
If the player occupies a position in the game board, then the associated bit would be 1
otherwise 0
(notice that bits 7, 15, 23, ... are 0
). To check if the player has a winning board you could use following function:
bool haswon(int64_t board)
{
int64_t y = board & (board >> 7);
if (y & (y >> 2 * 7)) // check \ diagonal
return true;
y = board & (board >> 8);
if (y & (y >> 2 * 8)) // check horizontal -
return true;
y = board & (board >> 9);
if (y & (y >> 2 * 9)) // check / diagonal
return true;
y = board & (board >> 1);
if (y & (y >> 2)) // check vertical |
return true;
return false;
}
With the help of a example I will try to explain: The following bitboard of one player includes beside vertical and diagonal wins a winning combination in the first row.
0101010
1110111
0111011
1101110
0001000
1010101
0011110 ... four occupied positions --> winning board
The steps for the horizontal check are:
y = board & (board >> 8)
0101010 0010101 0000000 1110111 0111011 0110011 0111011 0011101 0011001 1101110 & 0110111 = 0100110 0001000 0000100 0000000 1010101 0101010 0000000 0011110 0001111 0001110
y & (y >> 2 * 8)
0000000 0000000 0000000 0110011 0001100 0000000 0011001 0000110 0000000 0100110 & 0001001 = 0000000 0000000 0000000 0000000 0000000 0000000 0000000 0001110 0000011 0000010
The horizontal check results in a board with one bit set, this means the board includes a win and the function returns true
.
I have used a similar function to check a connect four game for a win. I saw this fascinating function in the sources to The Fhourstones Benchmark from John Tromp.
回答3:
loop though all positions. For each position check the four fields down diagonal-down-right and right (always including the field itself). Put in apropriate checks to avoid blowing up you app when you are checking fields that don't exist.
回答4:
Simple. Make 4 for
loops, for all rows, columns, increasing diagonals, decreasing diagonals.
In each, test if there are 4 consecutive pieces.
来源:https://stackoverflow.com/questions/7044670/how-to-determine-game-end-in-tic-tac-toe