This simple Tic-Tac-Toe will never lose, without having to do any minimax computation. It works by a fairly simple decision-making process. It is context-independent except for one patch to avoid a common trap.