4 ms·
I'm talking about cases where society's best outcome isn't a Nash equilibrium, but due to deception about the rewards of different actions, the socially optimal
by KMag 4y ago
I'm talking about cases where society's best outcome isn't a Nash equilibrium, but due to deception about the rewards of different actions, the socially optimal solution appears to be a Nash equilibrium. Some citizens would be better off if they defected, and society worse off, but due to deception, they don't realize they'd be better off.