Implementing a no-loss state in the game of Tic-Tac-Toe using a customized Decision Tree Algorithm

The game of Tic-Tac-Toe is one of the most commonly known games. This game doesn't allow one to win all the time and a significant proportion of games played results in a draw. This study is aimed at evolving of no-loss strategies in the game using decision tree algorithm and comparing them with existing methodologies, mainly focused on the implementation of the game using the minimax algorithm. The minimax algorithm does provide an optimal no-loss strategy by assuming that both players play optimally. So the question that comes out is what happens when the opponent plays un-optimally, in these cases the minimax proved to play non optimal moves, even though it wins at the next state rather than the expected state. Thus this paper provides a clear study of those trivial states and provides an optimal game play using an decision tree algorithm independent of the opponent's game strategy.