The study tested different theories of learning in games by having people play a matching-pennies game. The results showed that when players could see the entire history of play, their expectations matched the predicted equilibrium. However, when players had limited information, their play did not match the equilibrium. This suggests that having complete information helps players make better decisions in games.