What algorithm would you use to solve a very large tick-to-play game?

Question

A small (3x3, 4x4) tick-toe-toe can be easily resolved by considering all cases. But, for example, you have 30x30 tic-tac-toe. What algorithm would you use to decide the next best move in this case?

Minimax + alpha beta cropping is one of the ways I know.

Is there any other way that is more efficient / not more efficient, but colder?

---

I know that playing will not be a very interesting game. I said 30x30 to ask that I would like the algorithms to work best in games where the number of cases that should be considered for an ideal solution is very large and therefore impossible.

Answer 1

I do not think this is probably a very fruitful problem. Cause:

• If the number of marks in the line to be won is large, the game will (as it seems to me) be drawn at any reasonable level of skill, because it is much easier to prevent a possible victory than achieve one yourself. For example, if you need 20 in a row to win on a 30x30 board, all you need to prevent a win is a mark for each row and column around the middle of the board, and a mark near the middle of each is a long diagonal.

• , , , , , . - - .

Answer 2

, , - 3030 ( , , - 3030 , , ), - .

Answer 3

. . gomoku, .

Answer 4

, . ()

Answer 5

, ? , .

Answer 6

Alpha Beta - , . Alpha beta . 1/0/-1 (//) ( ), qualty .

( tic-tac-, ) http://www.fierz.ch/strategy1.htm

Answer 7

3- 3. 3, 2, 3, 3, 2. () .

If an adversary starts, select an empty 4x4 block and start in the middle, as described above. If the opponent completes his three in front of you, you lose.

I would dare say that this is the optimal strategy for boards 4x4 and above.