Computational Complexity of Proper Equilibrium

We study the computational complexity of proper equilibrium in finite games and prove the following results. First, for two-player games in strategic form we show that the task of simply verifying the proper equilibrium conditions of a given pure Nash equilibrium is NP-complete. Next, for n -player games in strategic form we show that the task of computing an approximation of a proper equilibrium is FIXPa-complete. Finally, for n -player polymatrix games we show that the task of computing a symbolic proper equilibrium is PPAD-complete.