About this app

Most other Tic-Tac-Toe games that say they use artificial intelligence really don’t. Those games use pre-programmed logic like telling the computer how to win if it has two pieces in a row already or how to block if the opponent has two in a row.

The only pre-programmed logic in “Tic-Tac-Toe Learns” on Level 0(Zero) is to pick the first open space. That is the simplest and least effective logic I could think of. This is intended to simulate playing a very young child that really doesn't understand the strategy of the game. All they know how to do is place a piece in an open space.

The more games you play the better the computer will get. Eventually it will get to where it will win or tie no matter what you do. The computer considers a tie a victory so it may or may not take the win if it already knows how to make the game end in a tie.

The objective of the game is to get the computer to learn in as few unique games as possible. Unique Games are tracked at the bottom of the playing screen.

In addition you are trying to avoid duplicate losses by the computer. Those are indicated in the "Dups" field. Any number other than zero indicates that the computer has lost a game in the exact same manner more than once. The game will continue to allow you to play after that because it might help you develop your strategy when you see what causes it to play the same losing game over and over.

This game isn't so much about the computer learning as it is about the user teaching the computer what the proper strategy is. The computer learns from its mistakes and its victories. However, if you intentionally play as hard as you can you will find that the computer will revert to making random moves because it never gets the chance to learn how to win.

You should try to let the computer win when it makes a move that is a good strategy. In addition you can force the game into a tie sometimes if you want the computer to continue trying to explore what the best option is.

If you are having trouble getting the computer to learn then you can try a different level. Level 1 is pre-programmed to take the win if it has two in a row already but it isn’t pre-programmed to block the opponent. Level 2 will always take the win or block the opponent using pre-programmed logic. All levels will use past games to improve their strategy until they will win or tie every time. Level Zero is the default level.

New features

Learning works in all levels. Level 0 uses no pre-programmed logic. It simply learns as you play. Level 1 will always try to win with pre-programmed logic but it won't block unless it learns how to do that. Level 3 will use pre-programmed logic to win or block. It learns for everything else. Default Logic level is now 0(Zero) since it learns to fast in all other levels.

Some Thinking now. If you play too fast the program can crash. Version 1.80. Ideal learning. Improved Stability.