Existence of Bertrand equilibrium revisited