Imagine an psychological experiment that required participants to sit in front of a computer screen. Four decks of cards were shown on the screen. Within each deck were reward cards and penalty cards. Participants were given a sum of money and told to increase their capital by picking cards randomly from the decks. A reward card would see them receiving a payoff while a penalty card loses them money.

Participants had expected good and bad cards to be distributed randomly among all four decks. Unfortunately for them, the decks are rigged. Two of the decks had a higher ratio of reward cards and someone who constantly selected from these two ‘Good’ stacks would be rewarded more often. The other two decks were more punitive. A participant who selected from these would be punished with losses more often.

Initial selections were haphazard. Participants had no basis to ground their selections on …