The article presents a method to refine Nash equilibria in games by using three axioms from decision theory. These axioms help select connected closed subsets called solutions where no player uses a weakly dominated strategy. Each solution contains a quasi-perfect equilibrium, ensuring optimal continuations from information sets. The method is robust and remains consistent even when embedding a game in a larger one. For games with two players and generic payoffs, the axioms characterize solutions as essential components of equilibria in undominated strategies, forming stable sets.